View PDF

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

LZS ; Xl ?
A BLS Reader
on Productivity
R eprinted from the M o n th ly Labor Review
and other sources
U S D epartm ent of Labor
Bureau of Labor Statistics
June 1983




V

A ®L§ R eader
©n ProductiwSfif
Reprinted from the Monthly Labor Review
and other sources
U.S. Department of Labor
Raymond J. Donovan, Secretary
Bureau of Labor Statistics
Janet L. Norwood, Commissioner
June 1983
Bulletin 2171

SOUTHWEST MISSOURI STATE
UNIVERSITY LIBRARY
U S. DEPOSITORY COPY

For sa le by th e S u p erin ten d en t o f D ocum ents. U .S. G overnm ent P r in tin g Office, W ash in gton , D.C. 20402







Preface

This reader presents articles from the Monthly Labor
Review and other publications, which discuss concepts
and measurement of productivity and related variables.
They analyze productivity in a number of industries and
the factors underlying productivity change over time.
They examine productivity measurement and trends in
the Federal Government; technological developments
and their impact on employment in a variety of in­
dustries; and international comparisons of productivity.
In addition, the reader contains a number of illustrative
statistical tables and charts.




Data for 1982 were not available for many of the
statistical series included in the tables at the time the
reader went to press. Such data will be furnished to in­
terested parties upon request.
The reader was prepared by the staff of the Office of
Productivity and Technology, Bureau of Labor
Statistics, U.S. Department of Labor.
Je r o m e A . M a r k

Associate Commissioner for
Productivity and Technology

JUN
iii




Contents

Page
I. Concepts and techniques of productivity measurement........................................................................................
Concepts and measures of productivity, by Jerome A. Mark ....................................................... .
Which productivity? Perspective on a current question, by Solomon F ab rican t.................... .....................
II. Productivity trends in the business e co n o m y .......................................................................................................
Multifactor productivity in the private business economy since 1948..............................................................
V'The slowdown in productivity growth: Analysis of some contributing
factors, by J. R. Norsworthy, Michael J. Harper, and Kent K unze..............................................................

30

III. Productivity trends in industries and the Federal Government............................................................................
Highlights of recent trends in output per employee h o u r..................................................................................
Measuring productivity in service industries, by Jerome A. M a rk ..................................................................

40
43
50

1
2
11
16
18

Studies of productivity in some individual industries:
Bituminous coal, by RoseN. Z eisel.............................................................................. ........................
56
Commercial banking, by Horst Brand and John D u k e ............................................................................... 58
Eating and drinking places, by Richard B. Carnes and Horst B ra n d .......................................................... 67
Farm machinery, by Arthur S. Herman and John W. Ferris........................................................................ 73
Intercity bus carriers, by Richard B. C arn es............................................................................................. ... 78
Laundry and cleaning services, by Richard B. Carnes.................................................................................. 83
Machine tools, by John Duke and Horst B ra n d ........................................................................................... 87
Non wool yarn mills, by James D. Y ork.......................................................................................................... 95
Office furniture, by J. Edwin Henneberger.................................................................................................... 99
Pumps and compressors, by Horst Brand and Clyde Huffstutler................................................................ 104
Retail fo o d stores, by John L. Carey and Phyllis Flohr O t t o ...................................................................... 112
Soaps and detergents, by Patricia S. Wilder ........................................................................................ ..
118
Measuring productivity in government, Federal, State and local, by Jerome A. M a rk .............. ................. 123
IV. International com parisons....................................................................... ............................................................. 131
International trends in productivity and labor costs, by Patricia Capdevielle,
Donato Alvarez, and Brian C o o p er.............. ...................................... ......................................................... 133
The international context, by Arnold Packer and Arthur N e e f ...................................................................... 143
V. Technology studies....................................................... .................... .................................................................... 145
Impact of new electronic technology, by Richard W. R ic h e ............................................................................ 147
Technology and labor in:
Electrical and electronic equipment, by Robert V. Critchlow.....................................................................
Electric and gas utilities, by Robert V. Critchlow..........................................................................................
Insurance, by Gustav A. Sailas.....................................................................................................................
Metalworking machinery, by A. Harvey Belitsky.........................................................................................
M otor vehicles and equipment, by Robert V. Critchlow ..........................................................................
Petroleum refining, by Rose N. Zeisel and Michael D. Dymmel ................................................................
Printing and publishing, by Robert V. Critchlow........................................................................................
Telephone communications, by Michael D. Dymmel .............................................. ...................................
Water transportation, by Robert V. Critchlow......................................................................................... ....

150
160
170
177
190
197
207
218
228

Bibliography............................................... ....................................... ............................................................................ 238




v




fpart i. <S@
oie®pts arod Ttehraiques ©f
Produetiwity M®®sur@mont

The first section of the BLS Reader on Productivity
introduces the basic concepts and ideas of productivity
measurement. Productivity refers to the relationship
between output and input.,:Generally speaking, a pro­
ductivity measure reflects the efficiency with which a
given output is produced by the resources employed.
This relationship takes a great variety of forms. The two
essays included here discuss two types of productivity
measures. Each is made up of several groups of
measures. One of the two types relates output to a single
input, such as labor, capital, or energy. The other
relates output to a combination of inputs, such as labor




and capital combined. Both types are discussed in this
Reader.
Although the productivity measures discussed here
relate output to hours of persons engaged in production
(hours of employment) or to labor and capital combin­
ed, they do not measure the specific contributions of
labor, capital, or any other single factor of production.
Rather, they reflect the joint effects of many influences,
including new technology, capital investment, the level
of output, capacity utilization, energy use, and
managerial skill, as well as the skills and efforts of the
work force.

Concepts and i¥ as tyres ©f Productivity
t]©
by <ferom@ A. i/lark*

Despite the wide attention paid to productivity over
the years, confusion prevails as to its meaning and meas­
urement. This is understandable because the concept
does lend itself to ambiguity and a wide range of produc­
tivity measures can and have been developed in response
to different analytical uses.
Productivity is loosely interpreted to be the efficiency
with which output is produced by the resources utilized.
A measure of productivity is generally defined as a ratio
relating output (goods and services) to one or more of the
inputs (labor, capital, energy, etc.) which were associated
with that output. More specifically, it is an expression of
the physical or real volume of goods and services related
to the physical or real quantities of inputs.
A variety of plausible productivity measures can be
developed, the particular form depending on the purpose
to be served. For example, output per labor input, the
most familiar measure, is useful in understanding changes
in employment or labor costs. This measure might be
based on man-hours paid or man-hours worked, with
different results. A more comprehensive measure of in­
put might be more useful in studying how the economy
is using labor and capital combined. Also, there are
various ways of adding up diverse products into a meas­
ure of output. No one measure is the right or best
measure.
Since the interpretation of these statistics depends
on the definitions and data used, an understanding of the
productivity concepts used in relation to the purpose to
be served is always essential.

that factor to production. Rather, they express the joint
effect of a number of interrelated influences on the use
of the factor in the production process—such as changes
in technology, substitution of one factor for another,
utilization of capacity, layout and flow of material, the
skill levels and the efforts of the work force, and mana­
gerial and organizational skills.
Whether for an individual establishment, an industry,
or the entire economy, the most frequently developed
and perhaps most useful productivity measure is an out­
put per unit of labor input measure of what is frequently
termed a labor productivity measure. There are several
reasons for this. Perhaps the most important is that labor
is almost universally required for carrying through all
types of production. There is a labor element of costs in
almost all endeavors; the degree varies but it is always
present. In addition, as a practical matter, it is perhaps
the most measurable input. Other factors, such as capi­
tal, are much more difficult to quantify.
There are, however, various labor productivity meas­
ures, depending on the definition of labor input. A
measure may refer to output per person or it may take
account of changes in hours of work and be based on
total hours. It may cover the hours of the entire labor
force including proprietors, unpaid family workers, and
employers; or it may be limited to selected groups
of workers.
Another set of productivity measures relating output
to a single input is output per unit of capital. These
measures are particularly useful in understanding move­
ments in unit nonlabor costs by relating the measures to
corresponding measures of returns to capital. As in the
case of other single factor productivity measures they
indicate the changes in the use of capital per unit of out­
put not the contribution of capital alone. The measures

CONCEPTS OF PRODUCTIVITY

There are two broad classes into which productivity
concepts and in turn measures can be grouped. One in­
cludes those measures which relate output of a produc­
ing enterprise, industry, or economy to one type of in­
put such as labor, capital, energy, etc.; the other in­
cludes those which relate output to a combination of
inputs extending to a weighted aggregate of all associate
inputs.1
Although the former measures relate output to one
input, they do not measure the specific contribution of

0 Assistant Commissioner for Productivity and Technology,
Bureau of Labor Statistics, U.S. Department of Labor.
1 Even at the level of an individual craftsman where the
output-input relationships are limited, these two concepts are
present. Productivity can refer to the volume of work the in­
dividual is able to accomplish within a given time span— i.e.,
output per man-hour. It can also refer to the volume of work
completed per unit of his time, Ms tools, and his materials—
i.e., an output per total factor input.

Reprinted from BLS Bulletin 1714 (1971),
The M eaning and M easurem ent o f Productivity.




2

impact on man-hour requirements depending on whether
output is measured by the yard or by the pound.
For the more usual case of a plant or an industry
producing many heterogeneous products, the different
units must be expressed on some common basis. They
can also be combined in terms of their man-hour re­
quirements. The advantage of the latter method for
measures of output per man-hour is that the change in
the productivity of the entire plant or industry is then
a simple arithmetic average of the changes in the pro­
ductivity of the individual components.
When the components are combined with value or
price weights, that is, on the basis of their dollar value,
then the output per man-hour measure for the total
reflects not only changes in the productivity of the
components but shifts in the importance of the com­
ponents.
Physical quantity data are often not readily available,
so deflation of dollar value is used. That is, total value of
production is adjusted for change in price by use of a
price index. This type of index is usually referred to as
constant dollar output or deflated value of output. Such
indexes are conceptually equivalent to indexes which
use physical quantities combined with price weights.
The contribution of a producing unit lies in the value
added, by its own labor and capital, to the materials and
services purchased from other producing units— i.e., its
net addition. Net output, therefore, is the constant
dollar value of production minus the constant dollar
value of purchased goods and services. In measuring
productivity, the net measure would then be related to
the particular input or all associated inputs except the
material inputs. Relating a net output measure to a
single input, when the various commodities produced
and purchased are combined with value or price weights,
will result in a single factor productivity measure that
reflects not only the changes in productivity of the
components and shifts in the importance of these
components but also savings in material consumption.

have been limited to reproduction of what has been
termed “tangible” capital.
Other single factor measures such as output per
energy input or output per material input are relevant
for plant and industry study where these inputs are of
considerable importance in the production process or
represent relatively scarce resources. For example, in the
aluminum industry where electrical power is an import­
ant element in processing bauxite, output per KWH-is
very useful as an indication of the efficiency with which
electricity is being utilized.
As mentioned earlier, all single factor productivity
measures reflect the joint effect of a variety of factors
including the substitution of one factor for another. For
some purposes, to develop a measure which eliminates
the effect of that substitution is useful. This type of
measure relates output to a combination of inputs. Thus,
a productivity index of output per labor and capital
combined elininates the effects of changes in amounts
of capital per workers These measures have been termed
multifactor or “total factor” or simply “total” produc­
tivity measures. For both conceptual and statistical
reasons they have generally been limited to labor and
“tangible” capital inputs and have not included as in­
puts activities such as research and education which can
be viewed as intangible capital.2
Output

For all productivity measures, output is measured in
physical or real terms. The concept is one of work done
or the amount of product added in the various enter­
prises, industries, sectors, or economy. It refers not to
activity as such but to the results of activity.
In this sense, at the plant level, production and hence
productivity measurement differs from work measure­
ment. Work measurement generally refers to the analysis
of the stages of activity and the requirements at each of
these stages. Productivity refers to the finished product
(the result of activity) and its relationship to input.
In the case of a producing unit making one homo­
geneous commodity, production in physical terms would
merely be a count of units produced. For a commodity
to be regarded as homogeneous, certain conditions
should be fulfilled. The product should be of a specified
quality (e.g., carbon steel) and it must conform to pre­
cise standards of size, volume, unit, etc. Even though
the measure of production in this case is a single count,
the way of defining the unit of product can have different
implications for productivity measurement. For example,
carpeting can be measured in pounds or square yards.
A change in the density of the carpeting would affect
the weight per yard and, therefore, have a different



Labor input

For all productivity measures where labor is rele­
vant, labor input is measured in physical terms. The
measure can refer either to the total number of individ­
uals engaged in production or to only part of the work
force, or it can refer to the man-hours of workers.
It is usually preferable to include the entire employed
work force in the labor input measure— blue-collar and
2
Denison in his w ork on the sources of econom ic growth
has made estim ates of the contributions of intangible factor
input such as research, education, organization, etc., to total
o u tput. See Edward Denison, The Sources o f E conom ic Growth
in the United States and the A lternatives B efore Us (New Y ork,
1962) and Why G rowth R ates D iffer (W ashington, D.C., The
Brookings Institution, 1967).

3

white-collar workers, corporate officers, and the selfemployed. The assessment of manpower needs must take
all labor input requirements into account. But of course,
there are times when analysis of labor requirements and
the analysis of cost components suggests the use of mea­
sures which include only a component of the work force.
To analyze the productive capacity of labor and the
effects of changes in working hours, or in use for pro­
jections of manpower needs, an output per man-hour
measure is most relevant. The most suitable unit of
measure is man-hours worked. There are some ambi­
guities or differences of opinion on what to include, for
example, standby time, coffee breaks, etc. In general,
“hours worked” refers to the time spent at the place of
employment, and therefore excludes hours paid for but
used on leave for vacation, holiday, illness, accident, etc.
In some cases, total hours paid are utilized in the pro­
ductivity measures because data on hours worked are
not available.
In developing a labor input measure, in many cases
man-hours are treated as homogeneous and additive.
These measures are particularly relevant to problems
of estimating total man-hour requirements. But merely
adding up the number of hours ignores the qualitative
aspect of an hour worked by different individuals.
Therefore, a productivity measure which is based on
the sum of undifferentiated man-hours will reflect
changes in the composition of the work force with dif­
ferent qualitative characteristics.
For some purposes, it may be desirable to develop a
productivity measure which takes into account the dif­
ferences in the “quality” of an hour of labor. That is,
an hour of high quality labor is counted as propor­
tionately more than an hour of low quality labor. To
do this some methods have to be introduced to differ­
entiate these hours. One way which has been utilized
is to combine the man-hours of various employees in
terms of pay differentials. The man-hours of higher
paid workers are given more weight than lower paid.
This assumes that differences in earnings reflect dif­
ferences in education, experience, skill, and their con­
tribution to output. 3 Another method is to adjust the
data to take into account changes in vocational training,
length of schooling, or type of education, etc., of the
work force, assuming there is a close relationship be­
tween qualifications and quality. When adjustments are
made for changes in the quality of labor input the re­
sultant productivity measure will not reflect changes
in the composition of the work force as a productivity
change but rather as a change in factor input.

Capital input

Capital stock estimates include the constant dollar
value of structures, plants, and equipment current avail­
able for production. These estimates may also take into
account the value of land, inventories, and working
capital.
Generally capital stock measures are derived by ad­
justing the value of existing plant and equipment for
new investment and the retirement of old assets. There
are different ways of measuring the stock of capital; for
example, they may be gross or net. Net stock estimates are
derived by depreciating assets (and there are various
methods of depreciation). Gross stock estimates are
derived by retaining assets at their full value until they
are retired from use. Since these are physical measures,
the value of capital stock must be adjusted for price
changes.
For productivity analysis, however, the flow of
capital services rather than the stock is the preferred
measure. A capital stock measure does not account for
differences in the intensity of use over time. Equipment,
for example, may be used for several shifts during a
business expansion or may be idle during a contraction.
Then, too, a large part of existing capital capacity may
be standby and employed only during periods when the
economy is operating at very high rates. There is also
a loss of efficiency of assets as they grow older. A
flow measure reflects differences in usage and efficiency
and how they affect varying levels of output, which is
the basis of productivity estimation. Ideally flow meas­
ures should indicate the amount of capital employed to
produce current output.
To derive this capital flow measure, an aggregate of
the capital hours used weighted by the rental value of
each type of structure and piece of equipment is needed.
The data for this measure are often not available in the
detail necessary for a capital flow measure.
A commonly used flow of capital service measure is
depreciation. However, this is based on accounting prin­
ciples which often reflect current income tax regulations
rather than the actual amount of capital used for cur­
rent production. Because of the difficulty of estimating
a capital flow measure, however, most analyses of capital
and production use capital stock estimates.
PROBLEMS OF MEASUREMENT
OF PRODUCTIVITY

The measurement of productivity trends involves two
fundamental problems which are applicable to both out­
put and input data. First, because of difficulties in ob­
taining direct quantity measures of output and input,
substitute measures or approximations must be used in
many cases. Second, since most data are collected for

3
Except to the extent that regional or similar wage dif­
ferentials affect average hourly earnings.




4

For the bulk of service activities, however, the de­
flation approach is used and its validity for the resultant
output measure rests on the adequacy of the price in­
dexes. Most of the price indexes used are components of
the Consumer Price Index, which in turn have different
degrees of reliability. The indexes for medical services,
for example, do not adequately take into account
changes in the quality of medical services performed.
As mentioned earlier, the real product measure is
conceptually a net output measure but in many of the
service activities data are not available on the real value
of the material inputs. In such cases estimates of real
product are made on the basis of changes in the total
volume of output. This does not present a serious prob­
lem, however, since in most service industries inter­
mediate purchases constitute a relatively small pro­
portion of total value.
The other major activity in the national accounts
where the output measure has severe limitations for pro­
ductivity measurement is the construction industry. The
constant dollar output measure is obtained by deflation,
but the price index used is really a cost index. For the
most part, these are measures of the change in costs of
materials and labor weighted in terms of their base
period importance. These indexes do not take into ac­
count any savings in the utilization of materials or labor,
and, as a result, there is an overstatement of price in­
creases. Consequently, there is an understatement of
gains in real output and hence productivity.
Productivity indexes based on real product for con­
struction show an average annual decline of 0.2 percent.
This is somewhat inconsistent with studies which the
Bureau of Labor Statistics has conducted of labor re­
quirements for various specific types of construction
during this period. These studies indicate for schools
there was approximately a 2% percent per year gain
over much “ this period, for highways a 3-percent in­
of
crease per year, and for hospitals not mcuh change.
With regard to lack of comparability of coverage be­
tween output and labor, perhaps the largest evidence of
this occurs in the real estate activity. For national in­
come accounting purposes, an inputed rent for home
ownership is added to the output of the real estate
industry. There is, however, no corresponding labor in-

purposes other than productivity measurement, defi­
nitions already established and procedures for reporting
information on production and factor inputs must be
used; these may or may not be consistent with concepts
appropriate for productivity measurement.
Output
Economy and sector level. The problems of using the
gross national product data for productivity measure­
ment involve primarily the inadequacies of the measures
of real output for some components and the lack of
comparability of coverage between output and labor
input measures.
Limitations in measuring output affect the reliability
of productivity statistics in some sectors more than
others. Since the implications for productivity move­
ments can be offsetting among the sectors, the effect on
measures for the overall economy is not as large as it is in
each of the sectors.
The three areas where the real product measures as
derived from the national accounts are particularly weak
for productivity measurement are government, construc­
tion, and services (including business and personal serv­
ices, and finance, insurance, and real estate).
In the absence of market valuation of the services of
general government agencies, the practice in national in­
come accounting is to value government output in terms
of the wages and salaries of government employees. The
deflated, or constant dollar, measure is derived from
changes in employment. Such an output measure, when
related to a labor input measure, results in no statistical
change in productivity. This measure of government out­
put may be increasingly difficult to continue in view of
the reported increases in output per man-hour in certain
government operations which are subject to measure­
ment. 4 Based on these data the trend of output per
man-hour for the national economy would be biased
downward. As a consequence, the available measures of
productivity are limited to the private economy.
Measuring output in the service activities is difficult
because of the absence of a directly quantifiable entity
which describes a unit of service. Consequently, various
substitute indicators are utilized in the national accounts.
These usually involve the use of some “price” index for
deflating the value of the service activities or the use of
an employment index to develop trends in producers of
services.
As in the case of government, the use of the employ­
ment movements as an indicator of the change in real
output implies a constant labor productivity. This ap­
proach is utilized for such activities as security and
commodity brokers, 5 insurance agents, and miscella­
neous business and repair services.




4

_
_
Nestor E. Terleckyj “ Recent Trends in O utp u t and Input
of the Federal G overnm ent,” in Proceedings o f the Business and
Economics Section, Am erican Statistical A ssociation, 1964,
pp. 76-94.
5 In view o f the rapid spread in recent years o f electronic
data processing in this industry this measure m ust be very m uch
understated since the productivity gains undoubtedly were
large, John Kendrick has suggested th a t data on shares o f stocks
and bonds sold appropriately weighted w ould be a better
measure. See “ Production and Productivity in the Service
Industries,” V ictor Fuchs, Educational Studies in Incom e and
Wealth No. 39, N ational Bureau o f Econom ic Research, 1969.

5

put associated with that output. Rough estimates indi­
cate that the removal of this activity from the output
account would reduce the productivity trend for the
private economy about two-tenths of a percent per year
over the last two decades.
Industry level. The effects of certain measurement
problems are greater at the individual industry level than
at the national level where there is a tendency for errors
and biases to offset each other. On the other hand, at
the industry level, more flexibility is possible because
the output measures are not part of an overall frame­
work (such as the national income and product accounts)
which requires certain definitions and measures not nec­
essarily consistent with the desired productivity measures.
Three major problems are encountered in developing
measures of output from available data for industry
productivity indexes. First, for many industries, the
appropriate detailed product data are not available.
Second, there is the well known quality change problem
which results from the development of new products
and the changing specifications of existing products.
Third, appropriate weights are often not available for
deriving the desired industry measure.
Some of the presently available industry indexes are
based on unit man-hour weights; others are based on unit
value or price. The use of unit value or price weights
is not a serious problem among commodities where labor
consts or inputs are a high proportion of price.
Labor Input
With regard to labor input measures there are several
data gaps in presently available measures. They relate
to changes in the composition of labor (the quality),
groups of the work force for which data are lacking or
incomplete, the relationship of output to the time of
research development and other workers whose activities
are not directed to current production, and finally, the
absence of adequate hours worked data on a comprehen­
sive basis.
1.Quality. As mentioned earlier, changes in the com­
position of labor input are adjusted in some measures by
weighting industry man-hours with the average hourly
earnings of workers in the industry. Insofar as earnings
differentials reflect productivity differences among work­
ers, this measure captures changes in the quality of
workers of different industries. However, this approach
has severe limitations. Pay differentials between indus­
tries reflect many factors unrelated to productivity dif­
ferences, such as the degree of unionization or regional
and geographical differentials. Moreover, the industry
hourly earnings differential does not take into account
occupational changes which occut within an industry.




Estimates of the effects of shifts among major
sectors— farming, manufacturing, mining, etc., show
that shifts contributed about 0.3 percent per year of the
output per man-hour growth over the last two decades.
The shift in composition of the work force within man­
ufacturing between production and nonproduction work­
ers contributed 0.1 percent per year to the rate of in­
crease in private output per man-hour over the last two
decades. In recent years this has been reduced consid­
erably to less than 0.05 percentage points.
In view of the limited information on occupational
detail, another approach (followed by Denison) to as­
certain the impact of shifts and changes in the work
force has been to utilize information on changes in age,
sex, and education. He estimates, for example, that the
increase in education of the work force contributed 0.7
percentage points to the trend rate in output per man­
hour from 1950-62. For a longer period, 1929-57, he
estimates the effect to be 0.9 percentage points per
year. Another estimate of the contribution of education
to the growth rate by Schwartzman6 provides a much
lower figure— three-tenths of a percent per year for a
roughly comparable period, 1929-63. The magnitude of
these differences in this critical area suggests that there
is need for further exploration of the interrelationship
between education skills, training, earnings, and pro­
ductivity.
2. Gaps in coverage. Payroll data on employment and
average weekly hours, which are the primary source of
man-hours estimates, do not include the entire economy,
but are limited to nonfarm wage and salary workers.
These data do not cover farm workers, proprietors, un­
paid family workers, and domestics. Estimates for these
sectors, for the most part, are taken from the labor force
series (based on household surveys) which is not strictly
comparable to the payroll series. Employment is a count
of persons rather than jobs as in the payroll data, so
that appropriate adjustments must be made.
Average hours for supervisory worker? in nonmanu­
facturing industries are not available. The assumption is
made that the average workweek for these workers is the
same as for the nonsupervisory workers in each industry.
Since 85 percent of all employees in nonmanufacturing
industries are nonsupervisory workers, however, the
effect of this imputation may be minimal.
Sampling procedures also affect the man-hours esti­
mates. One week of each month is used to represent the
entire month. If anything unusual, such as an unpaid
holiday, strike, or bad weather, occurs during this period,
the estimates will reflect these aberrations for the entire
6 David Schwartzm an, “ Education and the Quality of Labor,
1929-63,” American Economic Review, June 1968.

quality of new capital should be incorporated in the
capital stock measures or treated as a productivity in­
crease, Both interpretations have been used in produc­
tivity analysis.
Total factor measures as currently presented are not
consistent with their treatment of capital and labor.
In general, labor refers to actual man-hours whereas
capital refers to available stock not taking into account
varying levels of utilization.

month. On the other hand, fluctuations in employment
between survey periods may not be reflected in the
sample estimates. For example, short-term layoffs and
plant shutdowns of 1 to 2 weeks between survey periods
would not be reflected in the man-hour estimates for
the month— leading to an overestimate of man-hours
and an underestimate of productivity.
3. Hours paid versus hours worked. Because of lack
of data, productivity measures for the most part refer
to hours paid rather than the more desirable measure of
hours worked.
Surveys now are being conducted biannually for the
nonfarm economy where information on leave hours
and hours worked will soon provide a body of data
which will fill some gaps in this area. Estimates of the
effects of the difference between hours paid for but not
worked on output per man-hour measures developed by
the Bureau of labor Statistics indicate that the effect
over the last 15 years has been about 0.1 percent per
year for the nonfarm economy.
The effects of course can vary substantially by sector
and at the industry level. Within manufacturing, the
annual surveys and censuses of manufactures do provide
measures of what could be termed plant man-hours,
and these are used in many industry productivity
measures.

AVAILABLE MEASURES OF PRODUCTIVITY

National measures. Each quarter, the BLS prepares
and publishes indexes of output per man-hour for the
private economy and for the farm, nonfarm, and man­
ufacturing sectors. 7 For these measures, output per
man-hour refers to the constant dollar value of goods
and sendees produced in relation to the man-hours of
all persons employed (including proprietors and unpaid
family workers). Corresponding and comparable indexes
of hourly compensation and unit labor costs are also
developed. •
The output measure for these productivity indexes
is real gross national product originating in the private
economy or the individual sectors. It comprises the pur­
chase of goods and services by consumers, gross priyate
domestic investment (including the change in business
inventors), net foreign investment, and government all
deflated separately for changes in prices.
Final goods and services are differentiated from inter­
mediate products in that they are usually not purchased
for further fabrication or resale. In addition to purchases
in the market, final goods and services also include some
items provided but not actually purchased such as food
furnished to employees, food produced and consumed
on farms, and the rental value of owner occupied homes.
Measures for the farm, nonfarm and manufacturing
sectors are derived by subtracting the value of goods and
services purchased by the sector from the constant
dollar value of products and services leaving the sector.8
The labor input measures for these series are based
largely on a monthly survey of establishment payroll
records. Since this survey does not cover total employ­
ment in the private economy and because there are gaps
in the hours information, it is necessary to use some sup­
plementary data to derive man-hours estimates for all
persons engaged in producing the output of the private

Total fa©t@ measures
r
Total factor productivity measures relate output to
the weighted sum of labor and capital and are therefore
subject to the limitations of each of these data series.
The problems of measuring output and labor input have
been discussed. Capital measures, however, are probably
the most difficult and complex measures to derive. They
contain highly differentiated elements, and to express
this differentiated stock in physical terms requires ad­
justing dollar values of assets for price change.
For the most part, available data for prices of struc­
tures are based primarily on cost information. As men­
tioned earlier, they do not take into account savings
in utilization of materials and other inputs. Furthermore,
the problems of obtaining representative prices for
equipment which is highly differentiated severely affects
the adequacy of the price measures used in capital
measurement.
In addition, technical advances are often built into
capital so that a piece of equipment produced in an
earlier period may not be as efficient as one currently
produced. In constructing price indexes, some of these
technical improvements may be incorporated as quality
changes, but adjusting for quality is often difficult. There
is some question as to whether improvements in the




7 Productivity, Wages and Prices, quarterly release issued by the Bureau
of Labor Statistics, U.S. Department of Labor.
8 This actual measures are developed according to a variety of approaches
because of data limitations. However, all are attempts to approximate this
concept.

7

economy. Various sources are utilized and data from
them are adjusted for consistency with the establishment
man-hours.
The establishment man-hours are based on an hours
paid rather than an hours worked concept. That is,
the estimates include paid holidays, vacations, sick
leave and other time off paid for by the employer
in addition to actual hours worked.
Another set of labor productivity indexes is developed
based on man-hours obtained from a monthly survey
of the noninstitutional civilian population. This survey
of households provides information on the labor force,
employment, unemployment as well as man-hours. The
man-hours estimates for the labor force series are based
on an hours worked concept, i.e., hours spent at the
establishment, thus excluding vacation and sick leave
but including such things as rest periods and standby
time.
Since compensation data are derived primarily from
establishment payroll records, when relating labor pro­
ductivity measures to hourly compensation, the appro­
priate series is the one based on establishment man-hours.
On the other hand, when examining the relationship
between productivity changes and displacement of work­
ers, since the employment and unemployment measures
are based on the household survey, the more consistent
output per man-hour measure is the one based on labor
force data.
In addition to the current indexes of output per man­
hour published by .the BLS, John Kendrick of the
National Bureau of Economic Research has published
indexes of output per man-hour for the private economy
and major sectors (as well as total factor productivity)
which include a series that makes adjustments for
changes in the composition of man-hours.9 Using average
hourly earnings for weighting man-hours at the industry
level he derives an index of output per weighted
man-hour. The basic man-hour data for this series are
generally the same as those for the establishment series
and the weights are also derived from BLS average
hourly earnings data. These indexes presently cover
the period 1887 to 1966.
Edward Denison of the Brookings Institution has
published measures of output per labor input in the
form of growth rates for selected periods, the most recent
being 1950-62.10 These measures also take into account
changes in the quality of labor; however, the procedure
differs from Kendrick’s. Adjustment based on age, sex,
education, and other changes in the labor force are
applied to basic employment and man-hours measures
to derive labor input reflecting changes in quality.
These measures are available only at the national level.




Industry measures. In addition to the indexes for the
private economy and major sectors, the BLS publishes
annually indexes of output per man-hour for selected
industries.11 At the present time, measures for about 40
manufacturing and nonmanufacturing industries such
as steel, motor vehicles, railroad transportation, coal,
etc., are prepared.
The output measures for these indexes are developed
by combining the data on quantities of commodities
or services within the industry with fixed period weights.
As mentioned earlier, man-hour weights would be pre­
ferred for developing these measures and insofar as pos­
sible these are used. However, where such information
is not available, other weights such as unit labor costs,
or unit value (price) are used. These substitutions are
introduced on the assumption that unit values are good
commodities in an industry.
In addition, for some industries where it is not pos­
sible to obtain any quantity information, indexes of
deflated value of output are developed. For these
industry measures current dollar value estimates are
divided by indexes of price change for the industry to
derive a real output measuer. The adequacy of these
measures is dependent on the quality of the price
measure.
The labor input data for these measures are estab­
lishment man-hours. As in the aggregate measures, they
are derived from payroll records and for the most part
are based on an hours paid concept. For manufacturing
industries, however, additional man-hours information
is available in terms of hours at the plant. These data
which theoretically exclude vacation, holdiays, and such
leave hours are closer in concept to an hours worked
measure. Unfortunately, the information on plant man­
hours is usually not as current as that from other
sources on establishment man-hours.
Capital productivity

Measures of output per unit of capital are not avail­
able on a current basis. Historical measures have been
developed by a number of researchers.
Separate estimates of capital stock and hence meas­
ures of capital productivity are available for the private
economy from sources such as the National Industrial
Conference Board, National Bureau of Economic Re­
search, and many economists doing research in pro­
duction analysis, in addition, the Office of Business
Economics prepares 12 different capital series depending
on alternative options for service lives, depreciation, and
9 John W. Kendrick, P roductivity Trends in the United
States, Princeton University Press, 1961.
10 Edward F. Denison, op. cit.
1 Indexes o f O utput per Man-Hour, Selected Industries,
1
1939 and 1947-70 (BLS Bulletin 1692).

ices and for selected industries within these major
groups.
Labor input is measured in man-hours and adjusted
for quality change using industrial hourly earnings.
Capital includes the net stock of structure, plant equip­
ment, inventories, working capital and land. The capital
measures do not include quality inprovements which
can occur because of technical advance. The capital and
labor input are added together with factor prices as
weights. The base period for the weights was changed
periodically to reflect economic conditions of the vari­
ous subperiods under analysis.
The combined factor measure developed by Edward
Denison relates net domestic product (excluding depre­
ciation) to weighted sum of capital and labor. The
weights are the base period share of dollar output of
each of these inputs. He also periodically shifted the base
period for the factor shares to reflect current economic
conditions. Denison’s analysis covers the total economy
for the 1919-62 period, and he is currently updating
his work.
His labor input is employment adjusted for quality
change using relative earnings for selected age-sexeducation groups. He also adjusts the labor input for
intensity of effort as reflected in varying lengths of the
workweek. His assumption is that as the workweek de­
clines productivity improves because the worker is less
fatigued and can work more diligently.

adjustment for price change. These estimates are con­
structed within the framework of the national accounts
and are consistent with output measures used for labor
productivity in the private economy. However, the vari­
ation among these series gives different results for the
growth of capital productivity.
The OBE estimates are prepared annually for the
private economy, farm, manufacturing, nonfarm, and
nonmanufacturing. They include equipment and nonresidential structures, but exclude such items as housing,
motels, and hotels.
These capital stock series are developed using the
perpetual inventory method. That is, each new piece of
equipment or structure is added into the stock estimates
and remains there until retired from use. Retirements of
assets are based on mean useful service lives published
in the Internal Revenue Services Bulletin of Service Lives
of Assets. Because the latter are believed to over­
state asset lives, OBE prepares estimates of capital
series either based on Bulletin F or 85 percent of
Bulletin F service lives.
Net capital measures are derived using either straight
line or double declining balance depreciation.
Most other capital series are developed in a manner
similar to the OBE measures with other variations on
mean service lives and methods of depreciating assets.
C@ feim
m i©d factor input productivity

The second group of productivity measures— those
which relate output to several factors— involve the
weighting together of the quantities of the separate
factors. For the most part, these measures have been
limited to output per unit of capital and labor combined.
Just as the separate components of an output index
must be combined with appropriate weights, the separate
components of the input measure also must be appro­
priately weighted together. Capital and labor can be
aggregated using their unit costs (e.g., wages, rate of re­
turns of capital) in a base year as weights. These weights
can also be viewed as the proportion of current dollar
output earned by each input (factor share) in a base
period.
Two sources of combined factor input or total factor
productivity measures exist— the work done by John
Kendrick for the National Bureau of Economic Research
and that by Edward Denison in his work on sources of
growth.
Kendrick provided annual measures for the period
1889 to 1957, and more recent measures, 1957-69, will
be available early in 1972. Estimates cover the private
economy and are based on GNP output measures. Sep­
arate measures were made for farm, manufacturing,
trade, finance, transportation, public utilities, and serv­




RECOMMENDATIONS

Within the general constraints imposed on all users of
economic data, productivity measures for the total priv­
ate and private nonfarm economics and for selected
major sectors (manufacturing, mining, trade, transporta­
tion, communication, and public utilities) are reliable
and useful for economic analysis. Conversely, produc­
tivity measures for construction and service-type indus­
tries are not reliable measures for identifying either the
magnitude or direction of change in productivity for the
reasons outlined above. To improve these measures, ad­
ditional information must be developed in two areas—
the data base from which output, input, and price
statistics are compiled, and the conceptual base upon
which the output and price data are developed.
Additional pries oraformati©?!

More and better price information in the service and
construction industries is of prime urgency. In construc­
tion, work is currently underway by the Census Bureau
to develop price measures in order to improve the meas­
ures of the real volume of residential construction put in
place. Additional research is also necessary for nonresidential construction. In the service sector, more adequate

9

and extensive price information on personal services is
needed as well as the expansion o f wholesale prices to
include business services. The BLS at the present time is
trying to develop an index of the general price level. The
development o f this index will materially assist produc­
tivity measurement because it will require the collection
of a wider range o f service prices.
More work is also needed in collecting data on durable
equipment (such as heavy machinery and aircraft) which
is highly differentiated and often custom-made.
Another recommendation is to have timely (quarterly
and annually) estimates on the imputed rental value of
owner occupied housing so that it may be excluded from
the output estimate and not bias productivity meas­
urement.

and man-hours. Some research is being carried out using
occupational and wage data to account for some changes
in labor quality, but it is necessary to develop a more
integrated system for collection than in currently in
effect.
Capital information is also needed to make a more com­
plete analysis o f factors affecting productivity growth.
Of paramount importance are better data on changes in
the quality of capital equipment and the length of
service lives.
The concepts used to define output need to be changed
to conform to a productivity framework. This is particu­
larly true of government and households and institutions
where actual output definitions are needed rather than
merely relying on an employment measure.
Financial intermediaries also present definitional prob­
lems. For example, banking output is currently measured
as liquidity. If the output reflected changes in the num­
ber o f transactions weighted by some value measures, it
would be more compatible with the inputs and provide
a means for making a better productivity measure.
These recommendations will not solve all o f the prob­
lems of productivity measurement. However, they will
certainly improve the output and related input measures
and thereby make productivity a more viable tool for
economic analysis.

Better man-hour and capital data
Input data also need to be improved. Most important
to productivity measurement would be better estimates
for certain components of labor input. This would entail
estimating supervisory hours in nonmanufacturing and
expanding employment sampling coverage so that re­
sultant data refer to the entire month rather than
1 week. Adjustment for changes in labor quality calls
for more detailed information on occupational wages




10

is the variety of productivity concepts and meas­
urem ents and thus the need to be explicit. The
other is the dependence of concept on purpose:
The concept m ust be appropriate to the use to
which the productivity m easurement is put.

Which Productivity?
Perspective ® a
m
Current Question

Variety off Concepts and M easurem ents
To illustrate the variety of concepts and meas­
urem ents attached to the word “productivity,”
let us ask an apparently specific question and see
some of the answers th a t can be given to it. The
question: W hat has been the average rate of
growth in the N ation’s productivity over the
period 1947-57? The answers, taken from a
study by the National Bureau of Economic Re­
search,* are as follows: (1) 3.4 percent per annum,
2
1
3
if we measure the rate of growth by the average
annual increase in real gross national product
per unweighted man-hour, in the private economy;
(2) 3.1 percent per annum, if we shift from product
per unweighted to product per weighted m an­
hour, otherwise keeping everything else the same;
(3) 2.3 percent, if we shift further to product per
weighted unit of man-hours and tangible capital
combined, still keeping everything else the same;
and (4) about 2.0 percent, if we measure the rate
for the entire economy, including government—•
which requires even bolder estimation than do the
preceding measures.
This is not an exhaustive list of all the “answers”
even from this single source; and, of course, in­
clusion of the concepts and calculations of other
statisticians would add to the variety. B ut
enough has been cited to m ake the point: W ithout
specification of its content, the word “produc­
tiv ity ”—even the phrase “national productivity”—
can mean more than one thing.
In concept, productivity is always a ratio of
output to input, and a productivity index is
always the ratio for one period (or place) relative
to the corresponding ratio for another period (or

S o lo m o n F a b r ic a n t *

“ W h e n I u s e a w o r d ,” H um pty D um pty could
say to Alice in the W onderland spun out of Lewis
Carroll’s dreams, “it means just w hat I choose it
to m ean—neither more nor less.” In the every­
day world in which we have to live, we cannot
afford this individual freedom of choice. Yet
everyone seems to interpret the word “produc­
tiv ity ” in the light of his own experience and in­
terests, and because these vary, productivity
means different things to different people and they
use the word differently. Confusion results.
Only by discussion and explicit definition can the
confusion be cleared up or avoided.
The same objective—to discuss and clarify the
meaning of productivity—brought people together
in 1946 a t the first productivity conference spon­
sored by the Bureau of Labor Statistics and the
Bureau of the B udget.1 Since 1946, much has
been done by the BLS and others to deepen the
public’s understanding of productivity and pro­
ductivity change. B ut education is a perennial
task. The rapid growth of Russia and other
countries, widespread concern over inflation and
the international balance of payments, the use of
improvement factors in wage negotiations and
contracts—these events since 1946 have stim ulated
greater interest in the rate and process of produc­
tivity change and thus in the meaning and meas­
urem ent of productivity.
The subject of productivity is large; therefore,
I shall direct attention to two main points. One

•Director of Research, National Bureau of Economic Research, Inc., and
Professor of Economics, N ew York University.
T his article is based on a paper presented b y Dr. Fabricant at discussions
on productivity held by the Bureau of Labor Statistics in 1960.
1Summary of Proceedings of Conference on Productivity, October S8-S9,1946
(BLS Bulletin 913, 1947).
3 John W. Kendrick, Productivity Trends in the United Stales (Princeton,
N .J., Princeton University Press for the National Bureau of Economic
Research, Inc., 1961), ch. 3 and app. A.

Reprinted from the M onthly Labor Review,
June 1962.




1!

place). B ut output can be defined broadly to
cover the N ation as a whole, for example, or
narrowly to cover a single industry. In p u t can
refer to the simple sum of all man-hours worked,
or to the weighted sum of man-hours (in which
an hour put in by a skilled worker is counted as
more than the hour put in by an unskilled worker),
or to the weighted sum of the hours served by
both labor and capital goods.
B ut modifying the noun “productivity” by
appropriate adjectives and defining our terms, at
least at the outset of each discussion, is only
p art of w hat needs to be done. Since produc­
tiv ity concepts and measurements vary, we m ust
also choose among them before starting discussion.
The choice depends on the purpose, for not every
productivity concept is appropriate to every
purpose.
Let us consider two uses of productivity con­
cepts and measurements—in the analysis of output
change and in wage determ ination or analysis.
These are distinct purposes and require quite
different concepts of productivity.
Analysis off Output Change
As a step in its analysis, output change m ay be
viewed as resulting from change in each of two
components. One is change in input, th a t is,
the total quantity of resources used to turn out
the product. The other is change in productivity,
in the sense of change in the efficiency w ith which
the resources are used to turn out the product—■
a change which results from change in technology,
economic organization, and the other deter­
m inants of economic efficiency. M easurement of
these components provides information on the
quantitative contribution—the im portance—of
each. The combination of both kinds of change—
in input and in productivity—equals the change in
output.
In this analysis of output, the appropriate input
is not the input of some of the resources used, b u t
of all resources used. I t is total input. The
appropriate productivity is output per unit of
to tal input. In effect, productivity is defined and
m easured in this way in order to make—
O u tp u t= T o ta l I n p u tX To^ " ~




12

Each term stands for an index number, in relative
form, and all three index numbers are on the sam e
base.
Change in efficiency in the use of resources
cannot be determined by comparing change in
output with change in a limited class of inputs.
Therefore, output relative to a particular class of
inputs—for example, labor alone or capital alone—
would be inappropriate for the present purpose.
O utput per unit of plant and equipment m ay fall,
for example, yet efficiency m ay rise if savings of
labor’s services per unit of product exceed the
increase in the services of capital per unit of
product. Similarly, an increase in output per
man-hour m ay overstate the rise in efficiency if
some of the laborsavings result from substitution
of labor by capital. I t is informative, of course,
to see how each class of inputs is changing in rela­
tion of output. The index of output per unit of
total input can, therefore, profitably be supple­
m ented with the “partial” indexes of output per
unit of labor input and of output per unit of capital
input. B ut neither of these partial indexes is a
very good substitute, in measuring efficiency, for
the index th a t combines them.
If it is the N ation’s total output th a t is being
analyzed, the productivity index refers to national
productivity—th a t is, real national product per
unit of labor and capital used in the entire
economy. If it is an industry’s output th a t is
being analyzed, the productivity index refers to
the industry’s productivity—its physical product
relative to the labor and capital used by the
industry.3
Wage Beterminatiom or Analysis
If the use of the productivity measure is in
wage analysis, negotiation, or determination, the
appropriate productivity concept is not ou tp u t
8 The Industry’s volume of production, it is assumed, is measured by its
“net output.” (N et output is roughly equivalent to deflated value added,
with value added as defined in the Census of Manufactures.) The industry’s
production may also be measured by “gross output,” and most frequently is
so measured. (Gross output is equivalent to deflated value of product, with
value of product equal to value added plus materials consumed.) In the
latter case, the appropriate total input is labor plus capital plus materials,
and efficiency is measured by gross output per unit of labor, capital, and
materials combined.
The two productivity measures w ill seldom be equal. The second measure
has some advantages over the first for measuring an industry’s overall effi­
ciency. The important point, however, is the need to be aware of the differ­
ence between the two and to avoid treating them as comparable.

per unit of total input. Instead, it is output per
m an-hour and, particularly, output per weighted
man-hour. (The wage index is presumed to be
adjusted to exclude the effects of interindustry
and intraindustry shifts in the relative importance
of different classes of labor. If, as is usually the
case, wages are measured in such a way as to reflect
these shifts, the appropriate productivity concept
is o u tput per unweighted man-hour.) F urther,
the appropriate concept is output per weighted
man-hour in the economy a t large, whatever sector
or industry happens to be under consideration.
The reason for choosing output per m an-hour
rather than output per unit of total resources is
not th a t the economy’s efficiency is unim portant
in the determination of wages. I t is very impor­
tant. B u t it is also insufficient. Wage levels are
also determined by other factors. O utstanding
among these is the scarcity of labor relative to
capital in the economy a t large. O utput per
weighted m an-hour combines both im portant
factors, efficiency and labor scarcity. T h at is, the
efficiency of the economy is measured by national
product divided by weighted m an-hours plus
capital. The relative scarcity of labor is measured
by weighted capital divided by weighted m an­
hours, or (more conveniently in the present
context) by weighted man-hours plus capital
divided by weighted man-hours. A combination
of the two factors, obtained by m ultiplying them,
yields”national product per weighted m an-hour:4
*

they are far less relevant than national efficiency
and capital per worker in the economy as a whole.
They belong among the “qualifications” th a t m ust
always be attached to any short list of wage
factors, no m atter how im portant, in order to ex­
plain, or perhaps justify, a discrepancy between
change in wages and change in national output
per man-hour.
Before turning to these qualifications, a bit
more needs to be said about the rationale of the
connection between national o utput per m an-hour
and wages.
Tie Between National Output Per M an-Hour and
Wages. Technological advance and the other
sources of rising national efficiency tend to cause
all incomes, labor and property alike, to rise. An
increase in national capital per worker tends to
cause wages generally to rise more rapidly than
returns per unit of capital. B ut it does not
necessarily follow th a t the rise in wages m ust
exactly equal, and m ay not exceed or fall short of,
the combined rise in efficiency and labor scarcity.
N or does it necessarily follow, as parallelism
between wages and output per m an-hour implies,
th at the share of aggregate wages in national
income (ignoring some differences between nation­
al product and national income) is constant, or—
it is essentially the same thing—th a t wage cost
per unit of product in the economy as a whole is
constant. For these to follow, certain conditions
m ust hold.
These conditions are: (1) Com petition prevails
in the m arkets for commodities, labor, and capital,
so th a t increased efficiency and greater capital
investm ent are free to bring appropriate increases
in wages.6 (2) The technical possibilities of sub­
stituting capital for labor are such th a t a change in
capital per m an-hour results in an equal change in
the ratio of wages to the rate of return on capital,
or—in terms of the measures we are using—a
given percentage change in the ratio of total input
to weighted man-hours brings an exactly equal
percentage change in wages. (3) The two fac­
tors—efficiency and labor scarcity—are inde­
pendent of one another, so th a t m ultiplying their
indexes together gives an economically (not
merely m athem atically) correct combination of
them . For example, technological change does
not alter the relative scarcities and thus the

Weighted M an-H ours
N ational Product . ._____ + Capital_____
W eighted M a n -H o u rs* Weighted M an-H ours
+ Capital
N ational Product
W eighted M an-H ours
The reason for choosing national o utput per
m an-hour even when the wage in a sector of the
economy is under consideration is not because the
efficiency and the capital per worker of the sector
are irrelevant. They are relevant. In the long
run, however, and as a rule even in the short run,
4If the wage index in question is not adjusted to exclude interindustry and
intraindustry shifts, there is need to include a third factor—the quality of
labor. This factor is measured by weighted man-hours divided by un­
weighted man-hours. The combination of all three factors is national prod­
uct per unweighted man-hour.
4This means that national product per man-hour provides the competitive
norm by which to judge actual wage behavior or wage proposals in non­
competitive situations.




13

tions th a t would yield the same outcome) are,
on the whole, not too wide of the mark, it also
supports the case for dwelling on the qualifications.
Indeed, even the hastiest backward glance re­
minds us th at up to this point our attention has
been confined to real w;ages. In a regime of rising
(or falling) prices, another m ajor factor affecting
wages is the price trend—and when the price
level is moving rapidly,, this can be the predomi­
nant factor.8 Explicit account can be taken of
it by combining the price index with the other
factors considered. This would yield an index of
value (not volume) of national output per m an­
hour, and this also m ight be called a productivity
index. B ut there seems to be little advantage in
doing so.

relative prices of labor and capital. These con­
ditions—or assumptions, for th a t is w hat they
are—need to be brought out into the open if the
reasoning back of the tie between output per
m an-hour and wages is to be understood.6
Since m any people have become accustomed to
accepting output per m an-hour as the productivity
measure appropriate to wage discussion, the path
we have followed in reaching it m ay strike them
as rather tortuous. W hy is it not sufficient to say,
as has often been said, th a t output per m an-hour
is appropriate because it keeps constant the labor
share in national income and the cost of labor per
unit of product? I t is insufficient if we are not
told w hat economic forces and w hat techno­
logical and other characteristics of the economy
keep, or tend to keep, the labor share and unit
labor costs constant. I t is necessary to say at
least w hat has been said above.
Laying bare some of the assumptions underlying
the use of output per m an-hour in wage discussion
raises questions about their validity and the
reservations th a t need to be attached to this use.
So does the historical record.

Necessary Qualifications. History, and economic
theory as well, remind us also of other things th a t
m ay not be overlooked. I can discuss these here
only briefly.
First, changes in technology, in tangible capital
per worker, in the quality of labor, and even in
the cost of living take time to work out their
effects in the economy and make their im pact oil
wages. This is true even in a competitive economy
and even with such allowance as needs to be m ade
for the influence of expectations on the lags. N ot
each industry in the economy, nor the economy

Historical Record. The long-term record does
suggest something close to parallelism between
wages in various industries and national output
per man-hour, and, therefore, a greater rise in
wages generally than in national output per unit
of total input.7 However, relative trends over
shorter periods are not quite the same as over the
long period for which the record is available.
Further, while wages in different industries have
moved up together, there have been differences in
the rates of increase. From one point of view the
similarities are more striking than the differences,
b u t the latter are not negligible. This means
th a t the wages of some industries have risen
significantly more, and of others, less rapidly than
national output per man-hour, in the longer as well
as the shorter period. Also, the differences do
not seem closely correlated w ith such clues as there
are to the presence of monopoly, Even if the
conditions specified above hold for the "average
in dustry” and the "average period,” there is little
reason to suppose th a t they hold for every industry
in every period.
While the historical record suggests th a t the
assumptions noted (or alternative sets of assump­




8 In the technical language of economics, the conditions are: (1) competi­
tion, so that the wage is equal to the marginal product of labor; (2) a produc­
tion function of such shape that the elasticity of substitution between labor
and capital is unity, so that a given percentage rise in the ratio of capital to
labor is accompanied by an exactly equal percentage decline in the marginal
rate of substitution of capital for labor; (3) neutrality of change in efficiency.
(In most formal mathematical presentations of these conditions, use is made
of the weighted geometric mean of inputs, rather than the weighted arith­
metic mean that is implied in the text. According to available calculations,
the difference between the two means is slight.)
7 Over the period 1889-1957, and for the private domestic economy as a
whole, the average annual percentage rates of change are as follows:
Real gross national product per weighted unit of man-hours and
tangible capital_____________ ______ - ............ ...................................
1.7
Weighted man-hours and real tangible capital (total input) per
weighted man-hour------- ------------------------------------------ . ----------------3
Real gross national product per weighted man-hour........................ .. 2.0
Weighted man-hours per unweighted man-hour.......................................... 3
Real gross national product per unweighted man-hour..................
2.3
These figures may be compared with:
Real hourly earnings, all workers (including proprietors and family
workers)----- ------------ ---------------------------------------------------------2.4
Real hourly earnings, wage earners in manufacturing....................
2.3
Because the hourly earnings indexes are unadjusted for the shifts noted above,
the productivity index comparable with the wage indexes is real gross
national product per unweighted man-hour.
The approximate character of the estimates must be kept in mind; see m y
Basic Facts on Productivity Change (N ew York, National Bureau of Economic
Research, Inc., 1959), Occasional Paper No. 63, pp. 29-37.
8 Whether and what the “ wage-push” contributes to inflation cannot be
considered here.

14

as a whole, is ever in full equilibrium either in the
present or in the base period with which the present
is compared. Consider, for example, an industry
in which output per man-hour has begun to rise
more rapidly than in the Nation at large. More
often than not, its output will respond by rising
more rapidly than national product, and its prices,
by falling relative to the general price level. If
the rise in its output is at a rate sufficient to cause
its demand for labor to grow relatively, wages in
the industry will tend to be higher than in other
industries, th a t is, its wages will tend to rise
more rapidly than national output per man-hour.
If its output does not expand rapidly enough, labor
demand will decline relatively, and the industry’s
wages will tend to be lower than in other indus­
tries, th a t is, its wages will tend to rise less rapidly
than national output per man-hour. W ith appro­
priate changes, much the same qualification m ust
be made for industries in which capital per worker
is rising more or less rapidly than in the economy
a t large.
Second, while the quality of labor has generally
risen as skills have improved and education has
spread, the rate of increase has varied among
industries and occupations. This, too, will affect
the relation between a particular wage and national
output per man-hour. So, also, will relative
changes in the “noneconomic” advantages and
disadvantages attached to a job and changes in
the values p u t on them.
Third, with governm ent—and the taxes govern­
m ent levies, the transfers it makes, and the services
it renders—economically so much more im portant
now than in earlier periods, changes in the struc­
ture and level of taxes and of government expendi­
tures also significantly affect, at least in the long
run, the relation between wages and national out­
p u t per man-hour. In addition, the difficulties of
m easuring wages and productivity are aggravated.
This explains the prevailing custom of using
indexes for the “private economy” to represent
national productivity.
And fourth, even for a given concept, produc­
tivity m easurem ents m ay differ because sources,
of data differ and because of differences in the
ways in which gaps in the data are filled. The
difference between Census labor force data and




15

BLS establishment data on m an-hours provides
an example. Further, the trend in productivity,
even with the concept and source of data speci­
fied, depends on the particular period covered,
for productivity does not advance evenly from
one year to another. For example, if the period
chosen begins with a trough and ends with a peak
in business, the rate of increase in productivity
will almost always be higher than if the period
begins and ends in the same business cycle phase.
I m ust entirely pass over other troublesome
points th a t arise when we ask the question, “Which
productivity?” These include the extent to
which labor and tangible capital adequately ac­
count for total input; the m easurem ent of tangible
capital input and of labor quality; the treatm ent
of unemployed resources; the term s of foreign
trade, which affect the N ation’s productivity as
well as its productivity m easures; and the relative
merits of different formulas for particular purposes,
in which the choice between “base” and “given”
years in selecting weights is only one element.
M ost of these and the other points raised deserve
more intensive study than has yet been given to
them.
I t should be evident by now, however, th a t the
assumptions involved in applying national output
per m an-hour in wage negotiation, determ ination,
or study are m any, and th a t they do not hold
fully in every period and in every situation.
Summary. The index of national ou tp u t per
m an-hour is highly useful because it summarizes
the m ost im portant general factors th a t determine
changes in wages in a competitive society. I t does
not cover all the relevant factors, it is tru e ; there­
fore, deviations of wages from national output
per m an-hour m ay be expected to crop up every­
where. B u t because it does cover the m ost im ­
portant general factors, in the absence of monopoly
large deviations should be infrequent. The bigger
the deviation, the more reason there m ust be to
question and inquire into it. However, to under­
stand the difference between a particular wage
movement or proposal and the current trend of
national output per m an-hour, and to judge it,
requires more than the index of output per m anhour.

Fart II. Produetiwity Tr®nds ° the
m
iy s in s s s Economy

Indexes of compensation per hour measure the hourly
cost to employers of wages and salaries, as well as sup­
plemental payments, which include employers’ con­
tributions to social security, unemployment insurance
taxes, and payments for private health insurance and
pension plans. Measures of real compensation per hour
reflect the adjustment of hourly compensation for
changes in the Consumer Price Index for All Urban
Consumers (CPI-U).
Unit labor cost measures the cost of labor input re­
quired to produce one unit of output and is derived by
dividing compensation in current dollars by output in
constant dollars. Unit nonlabor payments measure the
cost of nonlabor items such as depreciation, rent, in­
terest, and indirect business taxes, in addition to cor­
porate profit and profit-type income of proprietorships
and partnerships.

This section examines the b l s measures of output per
hour of all persons engaged in production and of
multifactor productivity for the business economy.
Developments in output per hour and in multifactor
productivity in the business sector over the last three
decades are outlined. The slowdown in the rate of pro­
ductivity improvement during the 1970’s is also ex­
plored.
Description of Measures for the
Business Economy1
Productivity and related measures are prepared for
the following sectors of the U.S. economy:
Quarterly and annual measures
Business sector
Nonfarm business sector
Nonfinancial corporations
Manufacturing, total, durable, nondurable

Earlier this year, the Bureau began a program of
multifactor productivity measurement to supplement
the labor productivity measures and to provide addi­
tional insights into productivity growth and economic
changes. This program is an outgrowth of analytic
studies undertaken by the Bureau investigating some of
the factors contributing to productivity growth.3 The
multifactor productivity measures for the business and
nonfarm business sectors are based upon capital and
labor inputs.
In aggregate sectors, productivity changes through
time reflect movements within the various component
industries as well as shifts in the relative importance of
each of the industries. For example, changes in labor
productivity and multifactor productivity are influenced
by the relative shift of inputs (labor and capital) from
low- to high-productivity industries and by productivity
changes in the component sector. Within industries,
other shifts occur which are not accounted for ade-

Annual measures only
Agriculture
Mining
Transportation
Communications
Utilities
Wholesale and retail trade
Finance, insurance, and real estate
Government enterprises
The Bureau’s output per hour measures are constructed as
the ratio between gross domestic product— GDP—
originating in the private business economy and its
subsectors, and the corresponding hours of all persons
engaged in each sector.2 The changes through time in
these major indexes reflect efficiency in the use of labor,
and indirectly, the effect of other input factors in the
domestic production of goods and services. The changes
in the productivity and related measures through the
business cycle typically show patterns which differ
substantially from those found in long-term move­
ments,- and, therefore, are the objects of special analytic
studies.
Labor input measures are based primarily on BLS
establishment payroll data on employment and hours
and reflect hours at work and paid time off for vaca­
tions, holidays, and sick leave as well. A survey has been
introduced to develop a set of labor input measures bas­
ed on hours at work and will be used to extend the pres­
ent series.



1 A more detailed discussion of bls measures of productivity and
related variables, such as compensation per hour, unit labor costs,
output, etc., may be found in the b l s H an d b o o k o f M ethods, Bulletin
2134-1, December 1982.
2 Gross domestic product is gross national product less the net
return on foreign investments. Net return on foreign investments is
considered as originating in the “ rest-of-world” sector.
3 J.R. Norsworthy, Michael Harper, and Kent Kunze, “The
Slowdown in Productivity Growth: Analysis of Some Contributing
Factors.” See p. 30 of this reader.

16

series o f b l s news releases: “ Productivity and Costs in
the Business Sector,” and “ Productivity and Costs in
Nonfinancial C orporations.” In addition, quarterly and
annual analyses are published regularly in the Monthly
Labor Review. Historical indexes o f these and related
data are available on request, as are detailed descrip­
tions o f data sources and computational procedures.
Indexes o f output per hour and related cost data are
published monthly in Employment and Earnings and
the M onthly Labor Review, and in each edition o f the

quately— changes in income and tastes, for example,
may contribute to shifts in consumption patterns to
higher quality goods, or to services rather than goods.
Short-term movements in productivity and unit labor
costs often result from cyclical variation in output; this
tends to distort the long-term relationship between out­
put and labor input, as noted below, or output and
multifactor input. A number o f studies are being con­
ducted to separate cyclical from long-term productivity
movements.
Indexes o f output per hour, compensation per hour,
and related cost data are published quarterly in two




Handbook o f Labor Statistics.

17

Mulfifactor Productivity in the Private Business Economy Since 1948

P r o d u c t i v i t y , a s m e a s u r e d b y o u t p u t p e r u n i t o f c o m b in e d l a b o r an d c a p i t a l
i n p u t s — m u l t i f a c t o r p r o d u c t i v i t y — r o s e a n a v e r a g e o f 1 . 5 p e r c e n t p e r y e a r fr o m
1 9 4 8 t o 1 9 8 1 i n t h e p r i v a t e b u s i n e s s s e c t o r , a c c o r d i n g t o a new m e a s u r e
i n t r o d u c e d t o d a y b y t h e B u r e a u o f L a b o r S t a t i s t i c s o f t h e U .S . D e p a r tm e n t o f
L abor.
T h i s new s e r i e s sh o w s t h e c h a n g e s i n t h e am ount o f l a b o r an d c a p i t a l u s e d i n
p r o d u c tio n ( t a b l e A ) .
As s u c h i t r e f l e c t s t h e j o i n t e f f e c t o f many i n f l u e n c e s ,
in c lu d in g ch a n g es in te c h n o lo g y , th e l e v e l o f o u tp u t, u t i l i z a t i o n o f c a p a c it y ,
t h e o r g a n iz a t io n o f p r o d u c t io n , m a n a g e r ia l s k i l l s , a s w e l l a s c h a n g e s i n th e
c h a r a c t e r i s t i c s and e f f o r t s o f th e w o r k fo r c e .
T he t r a d i t i o n a l p r o d u c t i v i t y s e r i e s — o u t p u t p e r h o u r o f a l l p e r s o n s — r e f l e c t s
t h e s e i n f l u e n c e s a n d a l s o t h e im p a c t o f c h a n g e s i n c a p i t a l p e r u n i t o f l a b o r
in p u t.
T he n ew m e a s u r e , t h e r e f o r e , s u p p le m e n t s t h e e x i s t i n g m e a s u r e b y
p r o v i d i n g a b a s i s f o r m e a s u r in g t h a t i m p a c t .
O ver t h e 1 9 4 8 - 8 1 p e r i o d , w h en m u l t i f a c t o r p r o d u c t i v i t y i n c r e a s e d 1 .5 p e r c e n t p e r
y e a r , th e t r a d i t i o n a l p r o d u c t iv it y m easu re o f o u tp u t p er h our r o s e 2 .4 p e r c e n t
per y e a r .
T h e r e f o r e , th e g r o w th i n c a p i t a l p e r h ou r c o n t r ib u t e d 0 .9 p e r c e n ta g e
p o in t t o t h e g r o w th i n o u tp u t p e r h ou r ( t a b l e B ) .
The g r o w t h i n m u l t i f a c t o r p r o d u c t i v i t y sh o w ed tw o d i s t i n c t p a t t e r n s : 2 . 0 p e r c e n t
p e r y e a r fr o m 1 9 4 8 t o 1 9 7 3 , b u t o n l y 0 . 1 p e r c e n t p e r y e a r fr o m 1 9 7 3 t o 1 9 8 1 .
T h i s slo w d o w n r e f l e c t e d a f a l l o f f i n o u t p u t g r o w t h , c o u p le d w i t h a f a s t e r g r o w th
o f c o m b in e d i n p u t s o f l a b o r a n d c a p i t a l .
T he a c c e l e r a t e d i n c r e a s e i n i n p u t s o f
la b o r and c a p i t a l a f t e r 1 973 w as du e t o t h e f a s t e r in c r e a s e i n th e h o u r s o f a l l
p e r s o n s s i n c e t h e a n n u a l r a t e o f g r o w th o f c a p i t a l w as s lo w e r a f t e r 1 9 7 3 .

Reprinted from BLS news release USDL 83-153,
April 6, 1983.



18

T a b le A .

A v e r a g e a n n u a l r a t e s o f g r o w th i n p r o d u c t i v i t y i n d e x e s and r e l a t e d
m e a s u r e s b y m a jo r s e c t o r , 1948 t o 1981 1 /

P r iv a te
b u s in e s s

M ea su r e

P r iv a te
n o n fa r m b u s i n e s s

2/

2/

19481981
P r o d u c tiv ity in d e x e s :
O u tp u t p e r h o u r o f
a l l p e r s o n s ^ 0 • Q. « • • .
O u tp u t p e r u n i t o f
c a p ita l s e r v ic e s . , 0 0
M u ltifa c to r
p r o d u c tiv ity 3 / • • • . •
O u tp u t

O o o e o 0 o o G o e

o o o e e e

e

In p u ts s
H o u rs o f a l l p e r s o n s 00
C a p ita l s e r v i c e s , o o 0e
C om bined l a b o r and
c a p it a l in p u ts 4 / . . *
g

1/
2/
3/
4/

1948“
1973

19731981

19481981

19481973

19731981

2 04

3.0

0.8

2.0

2.5

0.6

-0.1

0.2

-1.0

-0.1

0.2

1.5

2.0

0 . 1

1.3

3 .3

3 .7

2 .2

0o9
3 .5

0 .7
3 .6

1.8

1.7

M a n u fa c tu r in g

19481981

194819 7 3

19731981

2.6

2.9

1.5

“ 1.1

-0.2

0.6

-2.6

1.7

0.0

1.8

2.2

0.4

3 .4

3 .9

2 .1

3 .3

4 .0

1 .2

1 .4
3 .2

1 .4
3 .6

1 .3
3 .6

1 .5
3 .3

0 .7
3 .6

1 .1
3 .5

-0.2
4 .0

2.0

2.1

2.1

2.1

1.6

1.8

0.9

A v e r a g e a n n u a l r a t e s b a s e d o n com pound r a t e fo r m u la e
E x c lu d e s g o v e r n m e n t e n t e r p r i s e s .
O u tp u t p e r u n i t o f c o m b in e d l a b o r and c a p i t a l i n p u t e
H o u rs o f a l l p e r s o n s c o m b in e d w i t h c a p i t a l s e r v i c e i n p u t s i n d e x w e ig h t e d
by la b o r and c a p i t a l s h a r e s .




19

T a b le B .

A v era g e a n n u a l r a t e s o f g r o w th i n o u tp u t p e r h ou r o f a l l p e r s o n s , th e
c o n t r i b u t i o n o f c a p i t a l s e r v i c e s p e r h o u r , and m u l t i f a c t o r
p r o d u c t i v i t y , b y m a jo r s e c t o r , 1948 t o 1981 _1/

19481981
(1 )

19481973
(2 )

19731981
(3 )

2 .4

3 .0

0 .8

-2.2

0 .9

1 .0

0 .7

-0.3

1 .5

2 .0

0 .1

-1.9

2 .0

2 .5

0 .6

-1.9

0 .7

0 .8

0 .6

-0.2

1 .3

1 .7

0 .0

-1.7

2 .6

M e a su r e

2 .9

1 .5

-1.4

0 .8

0 .7

1 .1

0 .4

1 .8

2 .2

0 .4

-1.8

S lo w ­
down
(4 )
(Col.2
- C o l . 3)

P r iv a te b u s in e s s 2 /
O u tp u t p e r h o u r o f a l l p e r s o n s
M in u s :

E q u a ls :

C o n tr ib u tio n o f c a p i t a l s e r v ic e s
per hour 3 /
M u ltifa c to r p r o d u c tiv ity

P r i v a t e n o n fa r m b u s i n e s s

4/

2/

O u tp u t p e r h o u r o f a l l p e r s o n s
M in u s:

E q u a ls :

C o n tr ib u tio n o f c a p i t a l s e r v ic e s
p er hour 3 /
M u ltifa c to r p r o d u c tiv ity

4/

M a n u f a c t u r in g
O u tp u t p e r h o u r o f a l l p e r s o n s
M in u s :

E q u a ls :

C o n tr ib u tio n o f c a p i t a l s e r v ic e s
per hour 3 /
M u ltifa c to r p r o d u c tiv ity

] A verage a n n u al r a t e s b ased
\J
2 / E x c lu d e s g o v ern m en t e n t e r p
3 / C h an ge i n c a p i t a l p e r u n i t
o u tp u t.
4 / O uput p e r u n i t o f c o m b in e d




4/

o n com pound r a t e f o r m u la ,
r is e s .
o f la b o r w e ig h te d b y c a p i t a l ’ s s h a r e o f t o t a l
l a b o r an d c a p i t a l i n p u t .

20

The following note briefly describes the major data sources and the procedures
used in deriving the new BLS multifactor productivity indexes* More detailed
information on the methods, limitations, and data sources are available on
request from the Bureau of Labor Statistics*
Tables 1-6 include all the productivity and related indexes for each year*

Summary of Methods

The multifactor productivity indexes are derived by dividing an output index by
an input index which is a weighted average of the hours of all persons and of
capital services* The output indexes are computed from measures of constant
dollar gross domestic product, derived from the national income and product
accounts developed by the Bureau of Economic Analysis of the U,S* Department of
Commerce *
The labor component of the input indexes is developed from measures of
employment and average hours, drawn mainly from the BLS Current Employment
Statistics program (the "establishment" survey) and the Current Population
Survey (the "household” survey)* The establishment survey provides information
about employees on nonagricultural payrolls; the household survey about the
self-employed, unpaid family workers, and those engaged in agriculture* The BLS
has done considerable research on the effects on productivity growth of
workforce composition (changes in the age, sex, and educational structure of the
workforce)* This work is not included in the measures published today because
more research is required*
The capital services component of the combined input indexes is developed from
measures of the stock of physical assets— equipment, structures, land, and
inventories— and rental prices for each type of stock. The stock measures, in
turn, are derived from data in the national accounts and other sources on
investment, service lives, and capital deterioration functions* The rental
prices are derived from data on depreciation costs and estimates of rates of
return on the capital assets*
The labor and capital components of the input indexes are combined with weights
which represent each component's share of total output* The index uses changing
weights where the share in each year is averaged with the preceding year’s
value *
Data are presented for the private business, private nonfarm business, and
manufacturing sectors* The private business sector, which accounts for about 80
percent at the gross national product includes all activities in the economy
with the exception of general government, government enterprises, the "rest of
world" sector, owner-occupied housing, nonprofit institutions, and private
household employees* The private nonfarm business sector also excludes
agriculture but includes agricultural services*




21

The traditional productivity measure of output per hour slowed-— dropping from a
growth rate of 3.0 percent during the 1948-73 period to 0.8 percent from 1973 to
1981. Of this 2.2 percentage point falloff, 0.3 percentage point was the result
of the slowdown in the growth of capital per unit of labor input. The
balance— that of multifactor productivity growth— -reflected the remaining
influences.
Output per unit of capital services, another productivity measure introduced
today, fluctuated between 1948 and 1981 but did not register a significant trend
over the period as a whole (chart A).
Private nonfarm business. From 1948 to 1981, multifactor productivity growth in
this sector averaged 1.3 percent annually as output rose 3.4 percent per year
and combined labor and capital inputs increased 2.1 percent per year (table A).
As was.the case for the private business sector as a whole, after 1973 there was
a marked change in the trend. Multifactor productivity grew 1.7 percent
annually from 1948 to 1973, but did not grow at all after 1973 (table B).
Hence, all of the increase in output during the later period came from increased
inputs of capital and labor.
The traditional productivity measure for the private nonfarm business sector,
output per hour of all persons, rose 2.5 percent per year from 1948 to 1973,
compared with 0.6 percent per year between 1973 and 1981. The 1.9 percentage
point slowdown in output per hour in this sector partly reflects a 0.2
percentage point decline in the contribution of capital services per unit of
labor input. Most of the slowdown, however, was due to the 1.7 percentage point
falloff in multifactor productivity growth, which in turn reflected the impact
of other influences.
As in private business, output per unit of capital input in the private nonfarm
business sector fluctuated from year to year, but there was no evident trend
between 1948 and 1981 (chart B ) .
Manufacturing. From 1948 to 1981, multifactor productivity grew faster in
manufacturing than in the more comprehensive business sectors. There was a 1.8
percent annual gain, reflecting a 3.3 percent average rise in output coupled
with a 1.6 percent annual increase in combined labor and capital inputs.
The falloff in the multifactor growth rate also occurred in manufacturing after
1973— from a rate of 2.2 percent during the 1948-73 period to 0.4 percent from
1973 to 1981 (table B).
The slowdown in the traditional output per hour indexes for manufacturing after
1973 was 1.4 percentage points, less severe than for the more comprehensive
business sectors. Moreover, the growth of capital per hour in manufacturing
accelerated after 1973. From 1948 to 1973 the growth in capital services per
unit of labor contributed 0.7 percent per year to the growth in output per hour
in manufacturing and, after 1973, 1.1 percent per year.




22

Table 1.

Private business sector: Productivity and related measures, 1948-81.

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

\J

Inputs
Multifactor
Productivity
2/

Output
v

Hours of
All Persons
47

Capital
5/

Combined Units
of Labor and
Capital Inputs, 6/

Capital per
Hour of
All Persons

Indexes 1977=100
1948
1949

45.3
46.0

99.2
93.6

60.1
59.4

36.8
36.1

81.3
78.6

37.1
38.6

61.3
60.8

45.6
49.1

1950
1951
1952
1953
1954

49.7
51.2
52.9
54.6
55.6

98.7
100.2
99.4
100.7
96.3

63.6
65.1
66.3
68.0
67.8

39.5
41.8
43.2
45.1
44.4

79.5
81.8
81.8
82.6
79.8

40.0
41.8
43.5
44.9
46.1

62.1
64.3
65.2
66.4
65.5

50.4
51.1
53.2
54.3
57.7

1955
1956
1957
1958
1959

57.8
58.5
60.0
61.8
63.9

100.9
100.0
97.9
94.3
99.3

70.7
71.0
71.6
72.0
74.9

47.9
49.2
49.7
48.9
52.5

82.9
84.2
82.9
79.0
82.1

47.5
49.2
50.7
51.9
52.9

67.8
69. 3
69.4
67.8
70.0

57.3
58.5
61.2
65.6
64.4

1960
1961
1962
1963
1964

64.8
67.0
69.6
72.3
75.4

98.4
98.0
101.2
102.6
105.2

75.4
76.9
79.7
82.0
84.9

53.3
54.2
57.2
59.7
63.3

82.2
80.9
82.2
82.7
84.0

54.1
55.3
56.6
58.2
60.2

70.7
70.5
71.8
72.9
74.6

65.8
68.4
68.8
70.4
71.6

1965
1966
1967
1968
1969

78.1
80.4
82.3
85.1
85.3

107.8
108.0
104.9
105.5
103.7

87.6
89,3
89.6
91.7
91.3

67.6
71.3
72.9
76.7
78.9

86.7
88.7
88.6
90.1
92.5

62.8
66.1
69.6
72.7
76.1

77.2
79.9
81.4
83.7
86.5

72.4
74.5
78.5
80.7
82.3

1970
1971
1972
1973
1974

86.1
89.2
92.4
94.7
92.4

98.6
98.1
101.0
103.0
96.5

90.2
92.2
95.2
97.5
93.8

78.3
80.6
86.0
91.8
89.9

90.9
90.4
93.2
96.9
97.2

79.4
82.2
85.2
89.1
93.1

86.8
87.5
90.4
94.1
95.8

87.4
91.0
91.5
92.0
95.8

1975
1976
1977
1978
1979

94.5
97.6
100.0
100.6
99.6

91.9
96.1
100.0
101.8
100.3

93.6
97.1
100.0
101.0
99.9

88.0
93.7
100.0
105.5
107.8

93.1
95.9
100.0
104.9
108.3

95.7
97.5
100.0
103.6
107.5

94.0
96.5
100.0
104.4
108.0

102.8
101.6
100.0
98.8
99.3

1980
1981

98.8
100.6

95.3
95.0

97.6
98.6

106.2
108.8

107.4
108.2

111.3
114.5

108.8
110.3

103.6
105.8

Average annual percent change 7/
1948-73
1973-81

3.0
0.8

-

1.0

2.0
0.1

3.7
2.2

0.7
1.4

3.6
3.2

1.7
2.0

2.8
1.8

1948-81

2.4

-0.1

1.5

3.3

0.9

3.5

1.8

2.6

0.2

See footnotes following table 6.




23

Table 2.

Private nonfarm business sector: Productivity and related measures, 1948-81. 1/

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

Inputs
Multifactor
Productivity

2/

Output

V

Hours of
All Persons
4/

Capital
5/

Combined Units
of Labor and
Capital Inputs 6/

Capital per
Hour of
All Persons

Indexes 1977=100
1948
1949

51.2
52.3

98.1
92.8

64.6
64.2

35.6
34.9

69.6
66.8

36.3
37.7

55.1
54.4

52.2
56.3

1950
1951
1952
1953
1954

55.6
56. 6
58.0
59.0
59.9

98.4
100.6
99.7
100.9
96.2

68.2
69.5
70.4
71.5
71.0

38.3
40.9
42.2
44.1
43.2

69.0
72.2
72.8
74.7
72.1

39.0
40.6
42.4
43.7
44.9

56.2
58.8
60.0
61.7
60.8

56.5
56.3
58.2
58. 5
62.3

1955
1956
1957
1958
1959

62.3
62.5
63.6
65.1
67.4

100.9
100. 1
98.0
94.0
99.5

74.1
74.0
74.3
74.3
77.5

46.8
48.2
48.7
47.8
51.6

75.1
77.0
76.6
73.4
76.6

46.4
48.1
49.7
50.8
51.9

63.2
65.1
65.6
64.3
66.6

61.7
62.5
64.9
69.3
67.7

1960
1961
1962
1963
1964

67.9
70.0
72.5
74.9
77.8

98.4
97.9
101.3
102.6
105.5

77.6
78.9
81.7
83.8
86.7

52.3
53.3
56.4
58.9
62.7

77.0
76.1
77.8
78.6
80.5

53.2
54.4
55.7
57.4
59.4

67.5
67.5
69.0
70.3
72.3

69.1
71.5
71.6
73.0
73.8

1965
1966
1967
1968
1969

80.3
82.2
83.8
86.7
86.4

108.1
108.7
105.3
106.0
104.1

89.2
90.7
90.7
92.9
92.1

67.0
71.0
72.5
76.5
78.7

83.5
86.4
86.5
88.2
91.1

62.0
65.3
68.9
72.1
75.6

75.1
78.3
79.9
82.3
85.4

74.2
75.7
79.6
81.7
83.0

1970
1971
1972
1973
1974

86.8
89.7
93.0
95.3
92.9

98.6
98.0
101.1
103.2
96.5

90.7
92.4
95.7
97.9
94.1

77.9
80.1
85.8
91.7
89.7

89.7
89.3
92.2
96.2
96.6

78.9
81.8
84.8
88.8
93.0

85.9
86.7
89.7
93.6
95.4

88.0
91.5
92.0
92.3
96.3

1975
1976
1977
1978
1979

94.7
97.8
100.0
100.6
99.3

91.7
96.1
100.0
101.9
100.0

93.6
97.2
100.0
101.1
99.6

87.6
93.6
100.0
105.7
108.0

92.5
95.7
100.0
105.1
108.7

95.6
97.4
100.0
103.7
107.9

93.6
96.3
100.0
104.6
108.4

103.3
101.8
100.0
98.7
99.2

1980
1981

98.4
99.8

95.1
94.4

97.3
97.9

106.2
108.5

108.0
108.8

111.7
115.0

109.2
110.9

103.4
105.7

Average annual percent change 7/
1948-73
1973-81

2.5
0.6

0.2
-1.1

1.7
0.0

3.9
2.1

1.3
1.5

3.6
3.3

2.1
2.1

2.3
1.7

1948-81

2.0

-0.1

1.3

3.4

1.4

3.6

2.1

2.2

See footnotes following table 6.




24

Table 3.

Manufacturing sector: Productivity and related measures, 1948-81. 1/

Inputs

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

Multifactor
Productivity

2/

Output

V

Hours of
All Persons
4/

Capital
5/

Combined Units
of Labor and
Capital Inputs 6/

Capital per
Hour of
All Persons

Indexes 1977=100
1948
1949

45.1
46.9

94.4
86.0

56.2
56.0

35.8
33.9

79.4
72.4

37.9
39.5

63.7
60.6

47.8
54.5

1950
1951
1952
1953
1954

49.4
51.1
52.0
52.9
53.7

94.9
99.6
95.7
98.6
89.2

59.9
62.3
62.2
63.5
62.3

38.6
43.0
44.5
47.5
44.1

78.2
84.2
85.4
89.8
82.1

40.7
43.2
46.4
48.2
49.5

64.5
69.1
71.4
74.8
70.8

52.1
51.3
54.4
53.7
60.2

1955
1956
1957
1958
1959

56.4
56.0
57.1
56.9
59.6

95.8
92.5
89.6
80.5
89.2

65.9
64.8
65.1
62.8
67.0

48.9
49.2
49.5
45.2
50.5

86.6
87.9
86.5
79.4
84.7

51.0
53.2
55.2
56.2
56.6

74.2
75.9
76.0
71.9
75.4

58.8
60. 5
63.8
70.7
66.9

1960
1961
1962
1963
1964

60.0
61.6
64.3
68.9
72.3

88.0
86.9
92.9
98.3
102.4

67.0
68.0
71.5
76.3
79.8

50.7
50.7
55.1
59.6
63.9

84.4
82.3
85.6
86.5
88.4

57.5
58.3
59.2
60.7
62.4

75.6
74.6
77.0
78.2
80.0

68.2
70.9
69.2
70. 1
70.6

1965
1966
1967
1968
1969

74.5
75.3
75.3
78.0
79.3

107.3
108.7
101.1
101.1
100.5

82.8
83.7
81.8
83.7
84.6

69.8
75. 1
75.0
79.1
81.7

93.6
99.8
99.6
101.4
103.1

65.1
69.2
74.2
78.2
81.3

84.3
89.8
91.7
94.4
96.6

69.5
69.3
74.5
77.1
78.9

1970
1971
1972
1973
1974

79.1
83.9
88.2
93.0
90.8

91.8
92.4
99.9
108.2
99.6

82.3
86.0
91.1
96.8
93.0

77.0
78.7
86.2
95.9
91.9

97.3
93.7
97.8
103.2
101.2

83.9
85.2
86.4
88.6
92.2

93.6
91.5
94.7
09. 1
98.8

86.2
90.9
88.3
85.9
91.1

1975
1976
1977
1978
1979

93.4
97.5
100.0
100.9
101.6

89.4
96.1
100.0
101.5
99.5

92.2
97.1
100.0
101.0
101.0

85.4
93.6
100.0
105.3
108.2

91.4
95.9
100.0
104.4
106.5

95.5
97.4
100.0
103.8
108.8

92.6
96.4
100.0
104.2
107.2

104.4
101.5
100.0
99.4
102.1

1980
1981

101.7
104.5

90.0
87.5

98.6
99.9

103.6
105.9

101.8
101.3

115.1
121.1

105.1
106.0

113.1
119.5

Average annual percent change 7/
1948-73
1973-81

2.9
1.5

0.6
-2.6

2.2
0.4

4.0
1.2

1.1
-0.2

3.5
4.0

1.8
0.9

2.4
4.2

1948-81

2.6

-0.2

1.8

3.3

0.7

3.6

1.6

2.8

See footnotes following table 6.




25

Table 4.

Private business sector: Productivity and related measures, 1948-81.

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

1/

Inputs
Multifactor
Productivity
2/

Output

Hours of
All Persons

Capital

3/

y

5/

Combined Units
of Labor and •
Capital Inputs 6/

Capital per
Hour of
All Persons

Percent change
1949

1.6

-5.6

-1.1

-1.0

-3.4

4.0

-0.8

7.6

1950
1951
1952
1953
1954

8.2
2.9
3.4
3.3
1.7

5.5
1.5
-0.8
1.3
-4.3

7.2
2.4
1.8
2.6
-0.4

0.4
5.9
3.3
4.4
-1.8

1.2
2.9
0.0
1.1
-3.4

3.7
4.4
4.2
3.1
2.7

2.1
3.5
1.5
1.8
-1.4

2. 5
1.4
4.2
2.0
6.3

1955
1956
1957
1958
1959

4.1
1.1
2.6
3.1
3.3

4.8
-n.o
-2.1
-3.7
5.3

4.4
0.3
o.o
0.7
4.0

8.1
2.6
1.0
-1.6
7.3

3.8
1.5
-1.5
-4.6
3.0

3.1
3.6
3.1
2.2
2.0

3.6
2.3
0.1
-2.3
3.2

-0.7
2.0
4.7
7.1
-1.9

1960
1961
1962
1963
1964

1.5
3.4
3.9
3.8
4.3

-0.8
-0.5
3.2
1.4
2.5

0.6
2.0
3.6
2.0
3.6

1.6
1.7
5.6
4.4
6.0

0.1
-1.6
1.6
0.6
1.6

2.4
2.2
2.3
2.9
3.4

0.9
-0.3
1.9
1.4
2.3

2.3
3.8
0.6
2.4
1.7

1965
1066
1967
1968
1969

3.6
3.1
2.3
3.5
0.2

2.4
0.2
-2.0
0.6
-1.7

3.1
1.9
0.3
2.4
-0.5

6.8
5.5
2.2
5.2
2.9

3.1
2.4
-0.1
1.7
2.7

4.3
5.3
5.3
4.6
4.7

3.6
3.5
1.9
2.7
3.4

1. 1
2.8
5.4
2.9
1.0

1970
1971
1972
1973
1974
1075

0.9
3.6
3.5
2.6
-2.4
2.2

-5.0
-0.5
3.0
2.0
-6.3
-4.7

-1.2
2.2
3.3
2.4
-3.8
-0.2

-0. 8
3.0
6.7
6.6
-2.1
-2.1

-1.7
-0.6
3.1
4.0
0.4
-4.2

4.3
3.5
3.6
4.6
4.5
2.7

0.3
0.8
3.3
4.2
1.8
-1.8

6.2
4.1
0.5
0.6
4.2
7.3

1976
1977
1978
1979

3.3
2.4
0.6
-1.0

4.5
4.0
1.8
-1.4

•3.8
3.0
1.0
-1.1

6.5
6.7
5.5
2.2

3.0
4.2
4.9
3.2

1.0
2.6
3.6
3.7

2.6
3.6
4.4
3.4

-1.1
-1.6
-1.2
0. 5

1980
1981

o
c
G
1

-5.0
-0.3

-2.2
1.1

-1.6
2.5

-0.8
0.7

3.6
2.9

0.7
1.5

4.4
2.1

1.8

See footnotes following table 6.




26

Table 5.

Private nonfarm business sector: Productivity and related measures, 1948-81.

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

If

Inputs
Multifactor
Productivity

2/

Output

V

Hours of
All Persons
4/

Capital
5/

Combined Units
of Labor and
Capital Inputs 6/

Capital per
Hour of
All Persons

Percent change
1949

2.2

-5.4

-0.6

-1.9

-4.0

3.7

-1.3

8.0

1950
1951
1952
1953
1954

6.3
1.9
2.5
1.7
1.5

6.0
2.2
-0.9
1.2
-4.7

6.2
2.0
1.2
1.5
-0.6

9.7
6.6
3.4
4.4
-2.0

3.2
4.6
0.9
2.6
-3.5

3.5
4.2
4.3
3.2
2.8

3.3
4.5
2.1
2.8
-1.4

0.3
-0.4
3.4
0.6
6.5

1955
1956
1957
1958
1959

4.0
0.3
1.8
2.4
3.5

5.0
-0.9
-2.0
-4.1
5.9

4.4
-0.1
0.4
0.0
4.3

8.4
2.8
1.2
-1.9
8.0

4.2
2.5
-0.6
-4.3
4.4

3.3
3.7
3.3
2.3
2.0

3.8
3.0
0.8
-2.0
3.5

-0.9
1.2
3.9
6.8
-2.3

1960
1961
1962
1963
1964

0.8
3.0
3.6
3.3
3.9

-1.1
-0.5
3.4
1.4
2.8

0.1
1.7
3.5
2.5
3.5

1.4
1.8
5.9
4.4
6.4

0.6
-1.2
2.2
1.1
2.4

2.6
2.3
2.4
3.0
3.5

1.3
0.1
2.3
1.8
2.8

2.0
3.5
0.2
1.9
1.1

1965
1966
1967
1968
1969

3.1
2.4
1.9
3.4
-0.3

2.5
0.5
-3.1
0.7
-1.8

2.9
1.7
0.0
2.4
-0.8

7.0
5.9
2.1
5.4
2.9

3.7
3.4
0.2
2.0
3.2

4.4
5.4
5.4
4.7
4.8

4.0
4.2
2.1
3.0
3.8

0.7
1.9
5.2
2.7
1.5

1970
1971
1972
1973
1974

0.4
3.4
3.7
2.5
-2.5

-5.3
-0.6
3.2
2.1
-6.5

-1.6
2.0
3.5
2.3
-3.9

-1.1
2.9
7.0
6.9
CJ
v
C
N
1

-1.5
-0.4
3.2
4.3
0.4

4.5
3.6
3.7
4.7
4.7

0.5
0.9
3.4
4.5
1.8

6.0
4.0
0.5
0.4
4.3

1975
1976
1977
1978
1979

2.0
3.3
2.2
0.6
-1.3

-5.0
4.9
4.0
1.9
-1.8

-0.5
3.8
2.9
1.1
-1.5

-2.4
6.9
6.8
5.7
2.1

-4.3
3.5
4.5
5.1
3.5

2.8
1.9
2.7
3.7
4.0

-1.9
2.9
3.8
4.6
3.7

7.4
-1.5
-1.8
-1.3
0.5

1980
1981

-0.9
1.4

-5.0
-0.8

-2.3
0.7

-1.6
2.2

-0.7
0.8

3.5
3.0

0.7
1.5

4.2
2.2

See footnotes following table 6.




27

Table 6.

Manufacturing sector: Productivity and related measures, 1948-81. 1/

Productivity
Year

Output Per
Hour of All
Persons

Output Per
Unit of
Capital

Inputs
Multifactor
Productivity
2/

Output

Hours of
All Persons

Capital

V

±1

5/

Combined Units
of Labor and
Capital Inputs 6/

Capital per
Hour of
All Persons

Percent change
-8.9

-0.4

-5.2

-8.9

4.0

-4.9

14. 1

1950
1951
1952
1953
1954

5.4
3.4
1.8
1.7
1.6

10.4
4.9
-3.9
3.0
-9.6

7.1
3.9
-0.1
2.1
-2.0

13.Q
11.4
3.3
6.9
-7.2

8.0
7.7
1.4
5.1
-8.6

'3.2
6.2
7.5
3.7
2.7

6.3
7.2
3.4
4.7
-5.3

-4. 5
-1.4
5.9
-1.3
12. 3

1955
1956
1957
1958
1959

5.0
-0.7
2.1
-0.4
4.8

7.5
-3.4
-3.2
-10.2
10.8

5.8
-1.6
0.4
-3.4
6.6

10.8
0.7
0.5
-8.6
11.7

5. 5
1.5
-1.5
-8.2
6.6

3.1
4.3
3.8
1.7
0.8

4.7
2.4
0.1
-5.4
4.8

-2.3
2.8
5.4
10.Q
-5.4

1960
1961
1962
1963
1964

0.7
2.7
4.3
7.2
4.8

-1.3
-1.2
6.9
5.7
4.2

0.1
1.5
5.1
6.7
4.6

0.3
0.1
8.6
8.3
7.1

-0.3
-2.5
4.1
1.0
2.2

1.6
1.4
1.6
2.4
2.9

0.3
-1.4
3.3
1.5
2.4

1.9
4.0
-2.4
1.4
0.6

1965
1966
1967
1968
1969

3.1
1.1
0.0
3.5
1.7

4.8
1.3
0.0
-0.6

3.7
1.2
-2.3
2.4
1.0

9.2
7.7
-0.2
5.5
3.4

5.9
6.5
-0.2
1.9
1.6

4.2
6.3
7.2
5.4
3.9

5.3
6.5
2.1
3.0
2.3

-1.6
-0.2
7.5
3.5
2.3

1970
1971
1972
1973
1974

-0.2
6.1
5.0
5.4
-2.4

-8.7
0.6
8.1
8.4
-7.9

-2.7
4.5
6.0
6.3
-3.9

-5.8
2.2
9.6
11.2
-4.2

-5.6
-3.7
4.3
5.5
-1.9

3.2
1.6
1.4
2.6
4.1

-3.2
-2.2
3.4
4.6
-0.3

9.3
5.5
-2.8
-2.8
6.1

1975
1976
1977
1978
1979

2.9
4.4
2.5
0.9
0.7

-10.3
7.4
4.1
1.5
-2.0

-0.9
5.3
3.0
1.0
-0.1

-7.1
9.6
6.9
5.3
2.7

-9.7
4.9
4.2
4.4
2.0

3.5
2.0
2.7
3.8
4.8

-6.2
4.1
3.8
4.2
2.8

14.6
-2.8
-1.5
-0.6
2.7

1980
1981

0.2
2.8

-9.5
-2.8

-2.4
1.4

-4.3
2.3

-4.5
-0.5

5.8
5.2

-1.9
0.9

10.7
5.7

1

4.0

c

1949

See footnotes following table 6.




28

F o o t n o t e s , T a b le s

SOURCE:

1 -6

Output data from Bureau of Economic Analysis (BEA), U.S. Department of
Commerce, and the Federal Reserve Board. Compensation and hours data
from the Bureau of Labor Statistics, U.S. Department of Labor, and
BEAo Capital measures are based on data supplied by BEA and U.S.
Department of Agriculture „
(1)

The private business sector includes all of Gross National Product
except the rest-of-world sector, the rental value of
owner-occupied real estate, the output arising in nonprofit
organizations, the rental value of real estate occupied by
nonprofit organizations, the output of paid employees of private
households, government, and the statistical discrepancy in
preparing the national income accounts. The private nonfarm
business sector also excludes farms, but includes agricultural
services„

(2)

Output per unit of combined labor and capital inputs,,

(3)

Gross Domestic Product originating in the sector, in constant
dollars 0

(4)

Paid hours of all employees, plus the hours of proprietors and
unpaid family workers engaged in the sector.

(5)

A measure of the flow of capital services used in the sector0

(6)

Hours of all persons combined with capital input, using labor and
capital shares of output as weights.

(7)

Average annual percent change based on compound rate formula.




29

J. R. NORSWORTHY
MICHAEL J. HARPER
KENT KUNZE
Bureau of Labor Statistics

The Slowdown in Productivity
Growth: Analysis o f
Some Contributing Factors
L abor productivity in the private business sector grew at an annual
rate of 1 percent from 1973 to 1978, about one-third of its rate of growth
from 1948 to 1965. The effects of this slowdown were both substantially
reduced economic growth and higher prices. A comprehensive analysis of
recent economic growth has been made by Edward F. Denison, whc^examined the effects of re g u la tio n on growth in a framework That assesses
the~Cbntrfbutions from various potential causal factors.1 Our_appr.oach is
different from his Tn several respects, depending primarily on the defini­
tion of output, and the measurement of capital input.1Several other studies
2
have focused on particular issues in the productivity puzzle, such as
analyses of the effects of capital formation, energy, labor force composi­
tion, and intersectoral shifts of labor.3
This paper investigates productivity in the private business sector for
which quarter!y.labor. productivity and-eost statistics are published by the
U.S. Bureau of Labor Statistics (BLSri The basic methodology weights
growth rates of capital and labor inputs by their shares in gross domestic
prqduelj3Lthis_S£ctor. Although-growthrinriabor^pigductivitv is the tar­
get for explanation, the framework includes the contribution of multifac­
tor productivity-growth— the Hicks-neutral residuaLJQie..measurement
techniques draw primarily on the work.oLDenison- and- Dale ,W^ Jorgen­
son,,_as outlined below.
The faclors-we examine as possibly contributing to the slowdown are
limited to those that can .be quantified and adapted for inclusion in a
national accounts framework. Therefore, we do not explore such issues as
deterioration-of The~~woTk-^thic^-aiid anv effect from such unmeasured
phenomena wilLpresumablv appear in the residual of our analysis. In an
alternative framawork—
based-on-regression analysis7dme~could try to
measure suc^phenemena-because the standards for quantifying them
could be relaxed.4 However, the collinearitvin single-equation regression
models-makesjthe coefficients associated with any single factor highly
variable,-depending greatly upon the other factors included in a particular
specification. A multiple-equation, simultaneous model might be at­
tempted; but it would be difficult to include a number of possible explana­

tory factors in a framework that allows for variable elasticities of substitution.
We examine, in addition, the existence and timing of the productivity
slowdown and its pervasiveness among major industry sectors of the
economy. And we estimate the contribution to this slowdown of changes
in the composition of the labor force, changes in capital-labor ratios,
trends in the ratio of hours worked to hours paid, interindustry shifts of
capital and labor, capital expenditures for pollution abatement, and in­
creases in energy prices. Most of these effects are analyzed by interpret­
ing them as augmenting or abating the effective input of capital or labor.
A general point about the analysis of the slowdown needs to be made at
the outset. For a particular phenomenon to contribute to a slowdown in
productivity growth, its effects must be greater in the slowdown period
than in the reference period. We therefore need data to estimate the
effects in both periods in order to determine any contribution to the slow­
down. It is not sufficient that a particular negative factor is at work during
the slowdown; it must be working demonstrably harder than before.

The Dimensions of the Slowdown
Peter Clark, after adjusting the labor productivity series for cyclical
movements, selected the time periods 1948-55, 1955-65, 1965-73, and
1973-77 for analysis.5 The endpoint years, except for 1977, are peaks in
Clark’s cyclically adjusted labor productivity. The year 1965 has addi­
tional claims as a watershed year: it marked the onset of major Vietnam
War deficit financing and increasing inflation. And at about that time the
first cohort from the postwar “baby boom” entered the labor force. We
Table 1. Rates of Growth of Labor Productivity, Output, Capital, and Hours, and
Ratio of Investment to Output, Private Business Sector, Selected Periods, 1948-78°
Annual average, in percent
Item

1. Edward F. Denison, “Effects of Selected Changes in the Institutional and Hu­
man Environment Upon Output Per Unit of Input,” Survey of Current Business, vol.
58 (January 1978), pp. 21-44.
2. Edward F. Denison, Accounting for United States Economic Growth, 19291969 (Brookings Institution, 1974).
3. For capital formation see Peter K. Clark, “Capital Formation and the Recent
Productivity Slowdown,” Journal of Finance, vol. 33 (June 1978), pp. 965-75; Eco­
nomic Report o f the President, January 1978, pp. 48-58; and J. R. Norsworthy and
Michael J. Harper, “The Role of Capital Formation in the Recent Productivity
Growth Slowdown,” Working Paper 87 (Bureau of Labor Statistics, January 1979).
For energy see Edward A. Hudson and Dale W. Jorgenson, “Energy Prices and the
U.S. Economy, 1972-1976,” Data Resources Review, vol. 7 (September 1978), pp.
1.24—
1.37; and George L. Perry, “Potential Output: Recent Issues and Present
Trends,” in Center for the Study of American Business, “U.S. Productive Capacity:
Estimating the Utilization Gap,” Working Paper 23 (Washington University, CSAB,
December 1977), pp. 1-20 (Brookings Reprint 336). For labor force composition
and intersectoral shifts of labor see Jack Beebe, “A Note on Intersectoral Shifts and
Aggregate Productivity Change,” Annals o f Economic and Social Measurement,
vol. 4 (Summer 1975), pp. 389-95; George L. Perry, “Labor Force Structure, Po­
tential Output, and Productivity,” BPEA, 3:1971, pp. 533-65; William D. Nordhaus,

S o u rc e : C o m p u te d b y a u th o r s u s in g d a ta fr o m
B u r e a u o f L a b o r S ta tis tic s .
y

1965-73

1973-78

3.32
3.71
2.62
0.38

2.32
3.77
3.67
1.44

1.20
2.62
2.05
1.42

12.3

13.5

12.8

th e U .S . B u r e a u o f E c o n o m ic A n a ly s is a n d

th e U .S .

a . O u tp u t, in v e s tm e n t, a n d th e c a p ita l s to c k a r e m e a s u r e d a t 1 9 7 2 p ric e s .
b . T h e m e t h o d o f a g g r e g a t i o n u s e d is d i r e c t a g g r e g a tio n . S e e t a b l e 4 .
c. M e a su re d a s h o u rs p a id

“The Recent Productivity Slowdown,” BPEA, 3:1972, pp. 493-526; Michael Grossman and Victor R. Fuchs, “Intersectoral Shifts and Aggregate Productivity
Change,” in American Statistical Association, Proceedings of the Business and Eco­
nomic Statistics Section (Washington, D.C.: ASA, 1972), pp. 66-75; and J. R. Nors­
worthy and L. J. Fulco, “Productivity and Costs in the Private Economy, 1973,”
Monthly Labor Review, vol. 97 (June 1974), pp. 3-9.
4. Robin Siegel, “Why Has Productivity Slowed Down?” Data Resources Review,
vol. 8 (March 1979), pp. 1.59-1.65.
5. Clark, “Capital Formation.”

Reprinted from Brookings Papers on Economic A ctivity, 2:1979.




1948-65

Labor productivity
Gross domestic product
Net capital stockb
Total hours of labor input0
Ratio of gross private domestic investment
to gross domestic product

30

Table 2. Growth of Labor Productivity and Share of Labor Input in the Total Private Business Economy, by Sector, Selected Periods, 1948-78

Annual average, in percent
Growth o f labor productivity
Sector
Private business
Agriculture, forestry, and fisheries
Mining
Construction
Manufacturing
Durable goods
Nondurable goods
Transportation
Communication
Electric, gas, and sanitary services
Trade
Wholesale
Retail
Finance, insurance, and real estate
Services
Government enterprises

1948-65

1965-73

1973-78

3.2
5.5
4.2
2.9
3.1
2.8
3.4
3.3
5.5
6.2
2.7
3.1
2.4
1.0
1.5
-0 .8

2.3
5.3
2.0
-2 .2
2.4
1.9
3.2
2.9
4.8
4.0
3.0
3.9
2.3
-0 .3
1.9
0.9

1.1
2.9
-4 .0
-1 .8
1.7
1.2
2.4
0.9
7.1
0.1
0.4
0.2
0.8
1.4
0.5
-0 .7

Share o f total labor hours
1948-65

1965-73

1973-78

100
12
1
6
30
17
13
5
1
1
23
6
17
5
12
2

100
6
1
7
32
19
13
5
2
1
25
7
18
6
14
2

100
5
1
7
30
18
12
4
2
1
26
7
19
6
16
2

S o u r c e : B u r e a u o f L a b o r S ta tis tic s . P r o d u c tiv ity d a t a f o r s e rv ic e s , c o n s tr u c tio n , a n d f in a n c e , in s u r a n c e , a n d r e a l e s ta te a r e u n p u b lis h e d .

n2 coefficients that describe the adjustment process. Particularly when out­
put changes are accompanied by substantial changes in relative prices, as
in 1973-74, the adjustment process may extend beyond the next cyclical
peak. At present, the cyclical adjustment issue probably cannot be dealt
with in a satisfactory way except in the context of an elaborate model
incorporating lagged simultaneous adjustment of inputs. Sufficient evi­
dence exists to suggest that any relatively simple method is inaccurate.
The distribution of the slowdown in labor productivity among major
industrial divisions shown in table 2 reveals different patterns in 1965-73
and in 1973-78.8 In manufacturing, the slowdown was about the same
magnitude in each period. Mining productivity began to decline in 1969
when the Federal Coal Mine Health and Safety Act was passed, and pro­
ductivity has continued to decline in recent years as energy prices have
risen and coal has played a larger role relative to petroleum mining. Pro­
ductivity growth in transportation slowed only slightly in 1965-73, but
fell much more in the recent period when energy prices may have retarded
an advance in productivity. Productivity growth in communications
slowed slightly in 1965-73 and then accelerated in 1973-78. (This in­
dustry is clearly not part of the productivity problem.) Utilities showed a
reduction in productivity growth in 1965-73 and a virtual halt in 197378. Energy prices, oil and gas shortages, and environmental regulations
are commonly cited as affecting this industry. Productivity growth accel­
erated in the trade industries in 1965—73 and fell off sharply in 1973—78.
In government enterprises, productivity declined in the base period, grew
slightly in 1965—73, and declined again in 1973—78. Agricultural produc­
tivity growth slowed in 1973-78.
Measures of output in the remaining three sectors are unreliable, and
they are included in the table only to complete the productivity picture in
the private business sector. The GNP Data Improvement Project— the
Creamer report— urged that output measures for construction be im­
proved because output is now essentially measured as deflated inputs
(labor and materials). Construction productivity, as measured, fell in
1965-73 and declined slightly less rapidly in 1973-78, after growing in
1948-65 at near the average rate for the private business sector. A report
by the U.S. Department of Commerce found no discernible cause for the

adapted Clark’s time periods (which were based initially on quarterly
data) by combining the first two periods and extending the last one to
1978. The present evidence that real output is leveling off or declining
during 1979 suggests that 1978 will be a reasonable endpoint for the
analysis.
For each of our reference periods, table 1 shows the rates of growth of
output, labor and capital input, and labor productivity in the private busi­
ness sector— the largest sector for which the BLS publishes productivity
statistics. The slowdown in the growth of labor productivity is evident in
the last two periods. The growth of the capital stock is examined carefully
below. It is worth noting here that it slowed substantially in the last period,
even though the ratio of investment to gross product was slightly higher
than in 1948-65. The growth rate of output does not explain variations in
this investment fraction the way a simple accelerator model would pre­
dict. However, accelerator effects might help to explain part of the slow­
down in capital formation between the last two periods.
Some investigators have chosen to examine the productivity slowdown
as a single phenomenon beginning in the middle to late 1960s. The argu­
ment for a break at the business cycle peak in 1973 seems compelling,
however. In addition to the sharp jump in energy prices that occurred, the
patterns of productivity growth rates— or slowdowns— in 1965-73 and
1973-78 are quite different. And Norsworthy and Harper have found
sharply different patterns of capital formation before and after 1973.6 We
therefore examine the productivity slowdown in two phases: 1965-73 and
1973-78.
By choosing our periods with endpoints that are years of relatively
high resource utilization, we avoid the need to make cyclical adjustments
in our data. Cyclical adjustment of output and input is an issue closely re­
lated to the choice of time periods for analysis. Clearly, productivity
growth is slower— and for quarterly measurements it is negative— during
economic recession. Measurement of average growth rates in output and
input between peaks or over relatively long periods implicitly assumes
that the various time periods between the endpoints are comparably
affected by negative cyclical influences. But this is clearly not true in the
time periods we analyze—-the years 1973-78 encompass a far more severe
recession in fewer years than did 1965-73. Only insofar as these recession
effects are captured in the factors we consider— for example, slower
growth of the capital-labor ratio— will they be captured in our analysis.
Nadiri and Rosen and Mohr have shown that the adjustment of any factor
input to changes in output depends not only on the output change itself,
but on the disequilibrium in other inputs.7 That is, with n inputs, there are

Demand, and Factor Productivity in 10 U.S. Manufacturing Industries,” Staff Paper
9 (Bureau of Labor Statistics, 1978).
8.
Output is based on gross product originating in the private domestic business
portion of each sector. Output and labor and capital inputs for nonprofit institutions
and household workers are excluded because output for those sectors is measured in
the national accounts by deflated labor compensation—thus productivity growth is
necessarily zero. This deduction is largely from the services sector. We also exclude
the imputation for rental value of owner-occupied dwellings both because the labor
input of homeowners and their families is not measured, and because final demand
categories such as home maintenance and repair and some utilities consumption
should properly be considered as intermediate inputs to the imputed output. This ex­
clusion affects the finance, insurance, and real estate sector.

6. Norsworthy and Harper, “Role of Capital Formation.”
7. M. Ishag Nadiri and Sherwin Rosen, “Interrelated Factor Demand Functions,”
American Economic Review, vol. 59 (September 1969), pp. 457-71; M. F. Mohr,
“A Quarterly Econometric Model of the Long-Term Structure of Production, Factor




31

Thus when capital and labor are the only inputs,

productivity decline.9 Within the finance, insurance, and real estate sector,
output is measured by labor input in the banking sector, where electronic
data processing has made major inroads. Any quality changes resulting
from this technological change are therefore not reflected in the output
measures for the sector. Measured productivity in that sector fell slightly
in 1965-73 after slow growth in 1948-65, and rose again in 1973-78. In .
the services sector, output is measured by labor input in several constitu­
ent industries, and inadequate deflation to account for quality change is
commonly cited as a problem. Measured productivity growth increased in
1965-73 and declined in 1973-78.10
The U.S. Bureau of Economic Analysis (BEA) does not publish data
on capital stocks for federal, state, and local government enterprises. Con­
sequently, we excluded output and labor input for government enterprises
from the private business and private nonfarm business sectors. Table 3

a = o —

Factors such as composition or quality change can make the effective
input of capital or labor differ from the measured input. Designating qK
as the change in factors influencing effective input of capital services and
qL as the change in factors influencing effective input of labor services, we
have
a = o — wKk — wLl — wKqK — wLqL.
To focus on the growth in labor productivity, we rearrange terms to
obtain
o — / = tV Q — 0 + WKqK + w^qL + o.
k <
The growth in labor productivity thus depends on growth in the capitallabor ratio, factors of composition or quality change, and change in totalfactor productivity. If all other factors are unchanged, labor productivity
will grow at the same rate as total-factor productivity.
A key assumption that underlies this approach to accounting for
growth in labor productivity is that the returns to various types of labor
and capital equal their contributions to output— that is, equal their margi­
nal products. This assumption, although questionable for any particular
point in time, is widely used in accounting for productivity growth, and is
more reasonable as a description of trends over longer periods of time
dhan a single year.
The particular factors whose contribution to labor productivity we
analyze can be described briefly. To measure the effect on labor produc­
tivity of shifts in labor among sectors, qLI, the growth rates of hours of
labor input in the sectors are aggregated using the proportion of total
labor compensation in each sector as weights. The effects of changes in
the composition of the labor force are computed by Divisia aggregation of
various categories of labor input— disaggregated by age, sex, education,
occupation, and class of worker. Divisia aggregation sums the growth rates
of each category of input, weighting each by its share of total labor input.
The index of the change in labor composition, qLC is then the difference
,
between the growth of the Divisia aggregate and the growth of the directly
aggregated (unweighted) labor input. The effect of shifts in the capital
stock among asset types, qK , is measured by aggregating the growth rates
C
of each type of capital asset weighted by each asset’s share in total non­
labor payments in the sector. The effect of intersectoral shifts in capital,
qKl, is measured by aggregating the growth rates of the capital stock in
each sector using the sector’s share of total nonlabor payments as a
weight. The effect of pollution abatement capital on the growth of the
capital stock, kPA, is also a kind of shift effect, and is treated as a deduction
from the capital stock.
Each of these factors affecting measured capital and labor inputs is
multiplied by the shares of labor and capital in nominal output— wL and
wK to compute the associated impacts on growth in labor productivity.
—
The framework for analyzing the effects of changes in various factors con­
tributing to growth in labor productivity thus can be expressed as

Table 3. Rates of Growth of Labor Productivity for the Private Business and Private
Nonfarm Business Sectors, Total and Excluding Government Enterprises,

Selected Periods, 1948-78
Annual average, in percent
Private nonfarm business

Private business

Total

1948-65
1965-73
1973-78

3.20
2.25
1.12

3.32
2.32
1.20

S o u rc e s : C o m p u te d b y a u th o rs u s in g d a ta fro m

Total

Excluding
government
enterprises

2.63
1.95
1.01

Period

Excluding
government
enterprises

2.77
2.02
1.09

th e B u r e a u o f E c o n o m ic A n a ly s is a n d B u r e a u o f L a b o r

S ta tis tic s .

shows the effects of the exclusion on the growth of labor productivity in
those sectors.
In summary, the pervasiveness of the slowdown suggests that an ex­
amination of major economic aggregates may be fruitful. At the same
time, growth in labor productivity by industry shows substantial differ­
ences between the 1965-73 and 1973-78 periods. An analysis that fails
to separate these two periods may miss important causal patterns. We
, therefore attempt to account for the slowdown in two phases: a slowdown
I of 1.00 percentage point a year in 1965-73 and a further slowdown of
1.12 percentage points a year in 1973-78.

Framework for Analysis

Our analysis separates growth in labor productivity into growth at­
tributable to changes in the capital-labor ratio, selected factors that alter
./ the effectiveness of measured capital and labor inputs, and residual or
otherwise unexplained growth, which may be considered as corresponding
j to total-factor productivity.
\
We begin by aggregating the growth rates of labor and capital inputs
weighted by their respective shares in output measured at current prices.
That is, the weight associated with the labor aggregate, wL, is the ratio of
total labor compensation to nominal output. Similarly, the weight asso­
ciated with the capital aggregate, wK, is the ratio of nonlabor payments
to nominal output. The measures of output in current and constant prices,
.labor compensation, nonlabor payments, capital stock, and labor input are
based on the national income and product accounts published by the De­
partment of Commerce. The flow of capital services is assumed to be
proportional to the net capital stock. The price of capital services is com­
puted as reported by Norsworthy and Harper.11
From the definition of total-factor productivity, A, we have
A = Oj ^

o — l — wK(k — 1) -f- WKqKc + WKqKi +

wk(—kpf)

+ wLqic +

w^ li

+ Wiqui + a,

where
(k — l) =
qK =
C
qKI =
kP =
A
qL0 =
qL[ =

WiXi,

where O is output;
is the share of input i in total-factor cost, with
2Wi = 1; and
is the quantity of input i used in producing O. Using
lowercase letters to denote percentage change, we obtain productivity
growth, a, from




— wLl.

rate of growth of the capital-labor ratio
effect of changes in the composition of capital
effect of intersectoral shifts in capital
rate of growth of pollution abatement capital
effect of changes in labor force composition
effect of intersectoral shifts in labor

9. H. Kemble Stokes, Jr., “An Examination of the Productivity Decline in the
Construction Industry” (U.S. Department of Commerce, Office of the Chief Econo­
mist, March 1979).
10. Interindustry shifts within the service sector account for a major part of mea­
sured productivity change.
11. Norsworthy and Harper, “Role of Capital Formation.”

a = o — X) W
i*,-.

i“l

32

Qlh — effect of changes in the ratio of hours worked to hours paid
a = change in total-factor productivity (residual).

is exact for the Cobb-Douglas specification, which requires strong sepa­
rability of the inputs being aggregated from other inputs appearing in the
production function. Translog aggregation, which is exact for a homothetic translog production function, requires weak separability of the inputs.
We performed econometric tests for each specification. The test for
the conditions required for direct aggregation failed by a wide margin for
all three sectors, while the test for translog aggregation passed for the pri­
vate nonfarm and private nonfarm business sectors and failed narrowly
for manufacturing.1 We therefore chose to use translog aggregation in
7
this study.
The choice of net or gross capital stocks of equipment and structures is
another issue in the measurement of the growth of the capital stock. For
productivity analysis, the issue comes down to whether net or gross capital
stock— or, indeed, some other measure— is the better indicator of real
capital input. In accounting terms, the difference between the gross and
net capital asset measures is the accumulated depreciation on the asset.
The method of depreciation and the service life of the capital asset are the
determinants of depreciation. There is precedent for using gross capital
stock, net capital stock, and a linear combination of the two.18 Denison
uses a linear combination of the net and gross capital stocks to measure
real capital input, whereas we use the net stock. Although the service
lives of capital assets are difficult to obtain, there is evidence that the net
capital stock from the national income accounts understates and the gross
stock overstates real capital input.19 The evidence is incomplete, but Deni­
son’s measure may be nearer to real capital input than that used here.
Evidence indicates that the results for 1965-73 are not sensitive to the
choice of measures: the net stock of equipment and structures in the pri­
vate nonfarm business sector grew at an average annual rate of 3.1 percent
in 1948-65 and 4.4 percent in 1965-73, while the gross stock grew at
rates of 2.7 and 3.9 percent in the respective periods. The changes in the
rates of growth therefore differ by only one-tenth of 1 percentage point.
The 1973-78 results, however, are sensitive to the choice between net and
gross measures.
In simplest terms, the translog aggregation of the capital stock that we
use is a method of correcting for aggregation bias because of changes in
the composition of the capital stock. The reasoning underlying the use of
the technique depends on the assumption that each asset type is used in
each sector in such quantity that its marginal product— the value of asset
services— is just equal to the price of the services of the asset. The price
of those services depends upon the purchase price of the asset, the cor­
porate tax rate, the service life of the asset (or the rate of depreciation),
other special tax treatment (such as capital gains or investment tax
credit), and the debt-equity structure of corporate liabilities.20 For exam­
ple, a shift in the composition of the capital stock from structures to equip­
ment (such as the one that took place from 1965 through 1977) repre­
sents an increase in the “quality” of the capital stock because the service
life of equipment is shorter than that of structures. Thus the depreciation
rate for the aggregate stock is higher, and the cost of capital services is
higher. The marginal productivity of the capital stock as a whole is there­
fore higher, and the flow of capital services in economic terms is greater.
The interindustry mix of the capital stock reflects differences in the rate

The residual or unexplained growth in labor productivity, a, is computed
as the difference between observed growth in labor productivity and the
contributions of the other effects. Thus it contains the effects of any errors
in measurement and of other factors not accounted for in the analysis.
The approach used here to measure sources of growth in labor pro­
ductivity is similar to the approaches used by Denison and by Frank M.
Gollop and Jorgenson in one respect— all depend on a share-weighting
scheme to estimate the contributions of various factors to productivity
growth.12 The focus on growth in labor productivity in this paper is an
expansion of a similar framework used by Christensen, Cummings, and
Jorgenson.13 Certain differences between our approach and the others
should be noted, however. Those relating to measurement of capital and
labor input are discussed in the appropriate sections below. Our concept
of output measurement is similar to that of most other investigators except
Denison. He measures output as net national income at factor cost and
thus excludes replacement investment from real output. Consistent with
this practice, he also excludes depreciation from the cost of capital, and
hence from the share of capital in the nominal value of output. To measure
output in this way seems less desirable than to include replacement invest­
ment because it is equivalent to assuming that the output that goes to re­
placing the capital stock could not be diverted elsewhere. Even at the
aggregate level, this is untrue— during the 1930s, net investment mea­
sured in the national income accounts was negative in at least one year.
And at the industry level, negative net investment in a given year is com­
mon. Denison’s approach reduces the measured effect of capital’s contri­
bution to productivity growth because, as noted above, the share of capital
is considerably smaller. In addition, the impact of productivity growth on
prices cannot be directly observed in Denison’s framework because output
prices include the full cost of capital. In a period such as 1965-73, when
the share of equipment in total investment and hence in the total capital
stock was rising, replacement investment was also rising because deprecia­
tion occurs faster for equipment than it does for structures. Thus output
in the private business sector would rise more rapidly in our accounting
framework than in Denison’s, other things being equal.

THE

C A PITA L STOCK

A number of issues arise in measuring the effects of capital input on
the growth of labor productivity. These include how to aggregate the cap­
ital stock; whether to use net or gross stocks; whether to include land, in­
ventories, and tenant-occupied housing; and whether to adjust for capac­
ity utilization. These issues are discussed extensively by Norsworthy and
Harper.14 Only the main outline of that argument is summarized here.
Issues in Measurement. Disagreement about the appropriate tech­
niques for aggregation of the capital stock— and, indeed, inputs in gen­
eral— for productivity analysis has characterized the discussion of pro­
duction theory in the economics literature.15 This particular type of
index-number problem turns on the validity of direct aggregation of the
components of the capital stock, measured in constant prices, as con­
trasted with translog or Divisia aggregation, which are both based on ag­
gregation of the growth rates of the components weighted by their shares in
total capital cost.16 In terms of the production function, direct aggregation

“The Explanation of Productivity Change,” Review o f Economic Studies, vol. 34
(July 1967), pp. 249-83. The application of the aggregation technique in time-series
analysis necessarily involves a discrete approximation to the continuous Divisia form.
The particular approximation— more than one is possible—used by Jorgenson and
his associates is based on the maintained hypothesis of a translog production or cost
function and thus seems best called a translog index. See Laurits R. Christensen, Dale
W. Jorgenson, and Lawrence J. Lau, “Transcendental Logarithmic Production Fron­
tiers,” Review o f Economics and Statistics, vol. 55 (February 1973), pp. 28-45.
17.. These tests are described in Norsworthy and Harper, “Role of Capital Forma­
tion.”
18. For gross capital stock see John W. Kendrick, Postwar Productivity Trends
in the United States, 1948-1969, General Series, 98 (National Bureau of Economic
Research, 1973); for net capital stocks see Laurits R. Christensen and Dale W.
Jorgenson, “U.S. Real Product and Real Factor Input, 1929-1967,” Review of
Income and Wealth, series 16 (March 1970), pp. 19-50; for a linear combination
see Denison, Accounting for United States Economic Growth.
19. Charles R. Hulten and Frank C. Wykoff, “Economic Depreciation and the
Taxation of Structures in U.S. Manufacturing Industries: An Empirical Analysis,”
in Dan Usher, ed., The Measurement of Capital (National Bureau of Economic
Research, forthcoming).
20. See Christensen and Jorgenson, “U.S. Real Capital and Real Factor Input.”

12. Denison, Accounting for United States Economic Growth, and Frank M.
Gollop and Dale W. Jorgenson, “U.S. Productivity Growth by Industry,” in John
W. Kendrick and Beatrice N. Vaccara, eds., New Developments in Productivity
Measurement (National Bureau of Economic Research, forthcoming).
13. Laurits R. Christensen, Diane Cummings, and Dale W. Jorgenson, “Eco­
nomic Growth, 1947-1973: An International Comparison,” in Kendrick and Vac­
cara, eds., New Developments.
14. Norsworthy and Harper, “Role of Capital Formation.”
15. The disagreement figures prominently in the debate between Edward F.
Denison, Dale W. Jorgenson, and Zvi Griliches, which is reproduced in “The Mea­
surement of Productivity,” Survey o f Current Business, vol. 52 (May 1972), pt. 2,
pp. 1-111.
16. The term “Tornquist index” is also used. The Divisia index, properly speak­
ing, is a continuous index, and some of the superior mathematical properties claimed
for it apply only in the continuous form. See D. W. Jorgenson and Z. Griliches,




33

prices and adjusted for price changes are reported by BEA. Correspond­
ing measures of land input are not available from that source. In this paper
we adopt the measures used by Kendrick in his estimates of the input of
land for the aggregate sectors.23
It is important to measure the capital stock that corresponds as closely
as possible to the output it produces. In his analysis of productivity growth
in the nonfarm business sector, Clark used the capital stock for the pri­
vate nonfarm sector of the economy and found that some slowdown in
labor productivity was attributable to capital formation in 1965-73.24 In
table 5 that capital stock is adjusted to conform to the definition of the
private nonfarm business sector by eliminating the capital in nonprofit in­
stitutions and including tenant-occupied residential capital.25 These ad­
justments increase the acceleration in capital formation between 194865 and 1965-73 from 0.74 percentage point to 1.31 percentage points a
year, enough to alter sharply Clark’s verdict on the role of capital in the
1965-73 slowdown. Inclusion of land and inventories modifies the pattern
only slightly. To adjust real capital input— the flow of capital services—
for changes in capacity utilization means that part of the corresponding
growth (or decline) in output can be traced to the change in capacity utili­
zation. Denison argues extensively and convincingly that this cannot be
done.28 He also argues that adjustment of the entire capital stock by utili­
zation rates in manufacturing is inappropriate because those rates inaccu­
rately reflect utilization rates for other sectors, and for assets other than
machinery. A careful reading of Denison’s argument— which is too exten­
sive to reproduce or even adequately summarize here— is compelling for
us and presumably for Jorgenson, who revised his measurement tech­
nique to eliminate adjustment for capacity utilization.27
We also make no separate adjustment for technological improvement
embodied in the capital stock. Insofar as these advances are reflected in
a higher price for the asset, the adjustment for changes in the asset mix
will capture the effect. If the improvements are achieved at no cost, the
quantity of the asset used in production will be correspondingly adjusted
so that the marginal product of the improved asset is equal to its service
price, as noted above. Thus in either case the equilibrium nature of the
model captures embodied technological change in the quantity or “qual­
ity” of the capital stock.
Effects of Capital Spending for Pollution Abatement. The effects of
investment in pollution abatement capital (PAK) on productivity growth
is assumed to operate only through the capital stock. A reliable estimate
of the contribution to the 1965— slowdown cannot be made because
73
data for investment in PAK are not available before 1968. The unofficial
BEA estimates of PAK investment and net stock are sufficient to fill out
the 1965-73 period, and this period can be used as a reasonably good
baseline with which to judge the effects of PAK expenditures in 1973-78
on productivity growth.28 Even the unofficial estimates begin in 1955. We
quite arbitrarily projected the estimated investment growth back to 1948
to obtain a baseline for estimating the contributions to the 1965-73 slow­
down. The data are poor and the technique mechanical; however, the re­
sulting changes in the rates of growth of the capital stock, shown in table 6,
are so small for the earlier periods in all but the manufacturing sector that
substantial changes in technique would make little difference. The effects
on the growth of labor productivity in the private business, private non­
farm business, and manufacturing sectors are estimated by weighting the
capital devoted to pollution abatement by the share of capital in total
output in thd^ectors.

of return on assets among industries. Because we only consider four asset
categories, whereas the BEA capital stock information is based on more
than twenty classes of equipment alone, there may also be systematic dif­
ferences in depreciation rates among industries reflecting the different
average service lives of the stocks of equipment and structures. Even in
the equilibrium model on which this aggregation technique is based, such
differences may occur in the average price of capital services across in­
dustries reflecting different capital stock composites. Therefore differen­
tial rates of growth of the capital stock by industry can lead to changes in
the value of the flow of aggregate capital services. As noted below, the
asset and industry dimensions of changing capital stock composition can
be separated in the translog aggregation process, and reported and ana­
lyzed separately.
The translog aggregation procedure makes it possible to isolate the
separate contributions of changes in asset type and changes in interindus­
try composition of the capital stock to the growth of the capital aggregate.
We may express the growth rate of the translog index for the capital
aggregate, kT, as
k r = k -f- qKA +

<k i ,
]

where
k = growth rate of the capital stock directly aggregated
qKA = growth contributed by changes in the asset mix (among equip­

ment, structures, land, and inventories)
qKi = growth contributed by changes in the industry mix of the capital
stock.21
An additional term, not shown in the expression for kT above, accounts
for the interaction between qKI and qKA. Where it is not shown explicitly,
we distributed the value of this term between the values of qKI and qK
A
in the tables presented below.
Direct and translog aggregation of the capital stock for the private
business sector are compared in table 4. The translog aggregate grows
more rapidly in all time periods, particularly in 1965-73 when there was
a substantial shift to equipment purchases in the manufacturing sector,
presumably in response to the investment tax credit. Assets and interin­
dustry shift generally follow the annual growth rates in magnitude. The
size of the total capital composition, or quality effect, is important; it pro­
vides between 10 and 20 percent of the average annual growth rate in
each period. The notion that aggregation effects of this sort can be ignored
seems to be refuted effectively. The rates of growth changed and so did
their intertemporal pattern: the increase in the rate of capital formation
in 1965-73 is greater for the translog aggregate.
In measuring total real capital input for productivity analysis, it is im­
portant to include land and inventories as well as measures of equipment
and structures.22 Stocks of inventories measured in current and constant
Table 4. Rates of Growth of Capital Stock, by Method of Aggregation, and
Contributions to Growth from the Effect of Capital Composition, Private Business
Sector, Selected Periods, 1948-78

Annual average, in percent
Effect o f capital composition
Method o f aggregation
Period

Direct

Translog

Total
effect

1948-65
1965-73
1973-78

2.62
3.67
2.05

3.14
4.48
2.31

Inter­
Asset
sectoral
composition shifts

0.51
0.82
0.24

0.30
0.41
0.18

0.34
0.51
0.10

Interaction
between asset
composition
and shifts
- 0 .1 3
- 0 .1 0
- 0 .0 4

23. John W. Kendrick, The National Wealth of the United States: By Major Sec­
tor and Industry, Report 698 (The Conference Board, 1976).
24. Clark, “Capital Formation,” p. 974.
25. Aggregates in table 5 are based on direct aggregation of capital stocks.
26. Edward F. Denison, “Some Major Issues in Productivity Analysis: An Exam­
ination of Estimates by Jorgenson and Griliches,” Survey o f Current Business, vol.
49 (May 1969), pt. 2, pp. 1-29.
27. Ibid., and Gollop and Jorgenson, “U.S. Productivity Growth.”
28. A more complete discussion of the quality of PAK data and their meaning is
found in John E. Cremeans, “Capital Expenditures by Business for Air and Water
Pollution Abatement, 1973 and Planned 1974,” Survey of Current Business, vol. 54
(July 1974), pp. 58-64; and his “Conceptual and Statistical Issues in Developing
Environmental Measures— Recent U.S. Experience,” Review o f Income and
Wealth, series 23 (June 1977), pp. 97-115.

S o u r c e s : C o m p u te d b y a u th o r s . N e t c a p ita l s to c k s e rie s f o r e q u ip m e n t, s tr u c tu r e s , a n d in v e n to r ie s a r e
f r o m th e B u r e a u o f E c o n o m ic A n a ly s is . D a ta o n la n d a r e f r o m

J o h n W . K e n d r i c k , T h e N a tio n a l W e a lth o f

th e U n ite d S ta t e s : B y M a jo r S e c to r a n d In d u s tr y , R e p o r t 6 9 8 ( T h e C o n f e r e n c e B o a r d , 1 9 7 6 ) , e x t r a p o l a t e d
f o r 1 9 7 5 -7 8 b y th e a u th o rs .

21. Only three industry sectors are recognized in the capital stock and investment
data available from the U.S. Bureau of Economic Analysis: manufacturing, farm,
and nonfarm nonmanufacturing. Because the definition of asset is a general one,
finer detail for each industry typically leads to a reallocation of capital “quality”
change— as the sum of the q terms above is often called—from asset to industry.
See Gollop and Jorgenson, “U.S. Productivity Growth.”
22. See Denison, Accounting for United States Economic Growth; Gollop and
Jorgenson, “U.S. Productivity Growth”; and Kendrick, Postwar Productivity.




34

Table 5. Reconciliation of Nonresidential Equipment and Structures to Business Capital, Private Nonfarm Business Sector, Selected Years,
1948-78 *
Net stock (billions o f 1972 dollars)

Item

1965

1948

Non re sid en tia l eq ui p m en t a n d
structures

1973

Rate o f growth (percent)

1948-65

1978

1965-73

1973-78

304.41
25.23

867.47
88.58

991.86
92.37

4.05
5.69

4.79
4.02

2.72
0.84

532.41

778.88

899.50

3.88

4.88

2.92

128.42

156.73

197.46

202.00

1.19

2.93

0.46

407.60
158.11
565.71

eq ui p m en t a n d structures

597.08
64.67

279.18

Minus: Ca pi tal of nonprofit institutions
Equals: Nonre sid en tia l business

689.14
261.09
950.23

976.34
358.47
1,334.81 .

1,101.49
388.25
1,489.75

3.14
3.03
3.10

4.45
4.05
4.34

2.44
1.64
2.23

Plus: Te n an t-o cc up ied residential
capital

Equals: Business e qu ip m e n t an d
structures

Plus: L a n d a n d inventories
Equals: Business capital
S o u rc e : C o m p u te d b y a u th o rs u s in g d a ta fro m
a.

th e B u r e a u o f E c o n o m ic A n a ly s is . F ig u r e s a r e r o u n d e d ,

T h e a g g r e g a te s a r e b a s e d o n d ir e c t a g g r e g a tio n o f c a p ita l s to c k s .

sixty-one industries; and the manufacturing sector, twenty-one industries.
The raw data for the disaggregation was compiled from records of the U.S.
Bureau of the Census, special labor force reports published by the BLS,
and for the last years, from tapes from the Current Population Survey.32
The growth rate in the adjustment for labor composition, qL0, is de­
fined as the growth in labor services adjusted for all categories of labor,
h, less the growth in unadjusted hours worked, l:

For the last period, the growth of the capital stock is affected notice­
ably by the adjustment for pollution abatement. For the periods before
1973, the table demonstrates that PAK expenditures had a minimal effect
on the capital aggregates. The effects in particular sectors were obviously
greater than what is shown in these aggregate data. Denison examines the
proposition from a broader perspective and still finds no major impact,
although his is an aggregate perspective also.29

qLc = h — /,
A D JU STM EN T
FORCE

FOR THE

CO M PO SITIO N

AND FO R IN T E R IN D U S T R Y

OF TH E

LABOR

where labor services is a function of the various categories of labor input,
L {\

S H IFT

We adapt the method used by Gollop and Jorgenson to analyze the
effects of the composition of the labor force and interindustry shifts.30 Our
procedure also follows Denison’s analysis closely.31 Denison does not ac­
count for different occupation groups nor is he always able to weight all
the separate characteristics by their specific relative wage as we do; how­
ever, this is because of a lack of data rather than a difference in approach.
The basic technique for translog aggregation of the various compo­
nents of the labor force is the same as that for aggregation of the capital
stock: each category of labor input is assumed to be paid the value of its
marginal product in each year. Thus relative increases in the proportion
of higher paid labor categories to total labor input are taken to represent
increases in effective input. This assumption underlies the adjustment by
Denison as well as by Gollop and Jorgenson for changes in effective
labor input.
To account for changes in the composition of labor input, the total
hours for each sector analyzed here— the private business, private non­
farm business, and manufacturing sectors— are disaggregated according to
sex, age, education, occupation, and employment class of worker (selfemployed or employee) for each year from 1948 to 1978. Total compen­
sation for each sector was disaggregated in the same manner. In all, there
are 1,600 disaggregations for each sector (two groups for sex, two for
worker employment class, five for education, eight for age, and ten for
occupation).
The interindustry disaggregation was based on the industry detail from
the national income and product accounts: the private business sector is
composed of sixty-two industries; the private nonfarm business sector,

H = f(L u L2 . . . , Ln).
,
Assuming / is a linear homogenous logarithmic function, the growth in
labor services is the derivative with respect to time:

h = £ vJi,
il
=

where

”• - f

h= £

iI
=

Sector
Private business
Private nonfarm business
Manufacturing

1965-73

Excluding
pollution
abatement
Total
capital

h = qLc + /,

29. Denison, “Effects of Selected Changes,” p. 42. The effects of pollution abate­
ment and health and safety regulations are analyzed by Denison in a different man­
ner. He concludes that by 1975 the annual impact of these activities as well as
private expenditures for crime prevention may have contributed as much as 0.26
percentage point a year to the slowdown measured from 1969 to 1975, reaching
0.47 percentage point from 1973 to 1975.
30. Gollop and Jorgenson, “U.S. Productivity Growth.”
31. Denison, Accounting for United States Economic Growth, pp. 30-50, 219-59.
For a comparison of the analyses by Denison, Gollop and Jorgenson, and Ken­
drick, see Kent Kunze, “Evaluation of Labor Force Composition Adjustment,” in
Measurement and Interpretation of Productivity (National Academy of Sciences,
forthcoming).
32. The disaggregation of the hours and compensation was resolved by use of a
multiproportional matrix model. The annual hours and compensation are controlled
at the industry level for employees, with only the hours and compensation for the
self-employed and unpaid family workers adjusted according to the March Current
Population Survey.

Excluding
pollution
abatement
Total capital

3.14
3.24
2.93

4.48
4.59
3.93

2.31
2.37
2.16

3.11
3.21
2.86

4.37
4.47
3.64

1973-78

2.05
2.09
1.47

S o u r c e : C o m p u te d b y a u th o r s u s in g d a ta f r o m th e B u re a u o f E c o n o m ic A n a ly s is ,
a . T h e a g g r e g a te s a r e b a s e d o n d ir e c t a g g r e g a tio n o f c a p ita l s to c k s .




Vi(h - / ) + /.

The difference (/, — l) is interpreted as the growth rate of the propor­
tion of total hours worked by the fth category of workers. The growth
rate of labor services can thus be expressed as the sum of the rates of
changes in qL and /. That is,
C

Annual average, in percent

Excluding
pollution
abatement
Total capital

S v<= *
■

We further decompose labor services into qro and /. Adding and then
subtracting the growth rate in unadjusted hours from the right-hand side
yields

Table 6 . Rates of Growth of the Capital Stock, Total and Excluding Pollution
Abatement Capital, by Sector, Selected Periods, 1948-78“

1948-65

and

35

where

Table 8. Rates of Growth of Direct Effects of Labor Characteristics on Labor
Composition, by Sector, Selected Periods, 1948-78

qLc = S

t1
-

Annua! average, in percent

V - I).
i(/,-

0.17
0.08
0.14

0.23
0.30
-0 .1 1

0.18
0.03
0.06

-0 .0 2
0.18
-0 .1 2

0.46
0.95
1.05

0.31
0.28
0.25

- 0 .0 6
- 0 .0 7
- 0 .2 3

- 0 .0 5
-0 .0 5
0.02*

0.04
-0 .3 0
- 0 .0 8

0.33
0.85
1.00

0.06
0.11
0.24

- 0 .0 4
-0 .0 8
- 0 .0 6

-0 .0 3
0.00
0.02*

0.17
- 0 .1 6
-0 .1 7

0.49
0.81
0.75

0.30
0.36
0.52

Education

th e B u r e a u o f th e C e n s u s a n d

C a lc u la te d f o r th e 1 9 7 3 -7 6 p e rio d .

attainment increased and added to effective labor input in the 1965-73 and
1973-78 periods for all sectors. Education and occupation are highly in­
terrelated factors, so that adjusting for education alone also captures a
significant amount of the contribution from the changing occupation mix.

H O U R S W O R K E D V E R S U S H O U R S PA ID

Labor productivity is generally measured using hours paid as the labor
input measure. The data are taken from the current employment statistics
(CES) program’s survey of nonagricultural establishments, which has
far greater coverage than any currently available survey of hours
worked.34 A 1976 report by the BLS found that no available survey pro­
vides data on hours worked that are sufficiently accurate to serve as a
basis for quarterly or annual measures of labor productivity.35
Insofar as hours paid exceed hours worked, the level of labor produc­
tivity will therefore be understated. Measured growth in labor productiv­
ity will be affected only if the ratio of hours worked to hours paid changes
through time; the measured slowdown in productivity growth will be af­
fected only if the rate of change of that ratio is altered. Recent work by
Stafford and Duncan,36 based on quite small samples, shows that the di­
vergence between hours worked and hours paid accounts for as much as
one-third of the productivity slowdown. This suggests that it is worth­
while to use the best available data to attempt to quantify the effect.
The BLS report made rough estimates of hours worked from 1952 to
1965, based on exclusion of leave from the CES data on hours paid, and
from 1966 to 1975, based on the Employer Expenditures for Employee
Compensation survey.37 From these data we estimated average annual
rates of change in the ratio of hours worked to hours paid for 1952-65,
1965-73, and 1973-75 in private business, private nonfarm business, and
manufacturing. The results, shown in table 9, are not striking. There is a
small, persistent but variable decline in the ratio of hours worked to hours
paid in each sector, except for manufacturing in the last period. The effects
on growth of labor productivity were estimated by assuming that the aver­
age annual growth rates for 1952-65 and 1973-75 characterized the
periods 1948-65 and 1973-78— a rather weak technique. The resulting
values were weighted 'by the share of labor in total output in the three
sectors.

34. The ideal target concept is hours actually worked. In this paper we use the
term to denote hours at the workplace, a concept that excludes paid leave (vacation,
holiday, and sick leave).
35. Bureau of Labor Statistics, “Report of the BLS Task Force on Hours
Worked” (BLS, March 1976). Modification of the survey to include the collection
of data on hours worked is now planned.
36. Frank P. Stafford and G. I. Duncan, “The Use of Time and Technology by
Households in the United States,” in Ronald G. Ehrenberg, Orley Ashenfelter, and
Ronald L. Oaxaca, Research in Labor Economics, vol. 3 (JA I Press, forthcoming).
37. The Employer Expenditures for Employee Compensation survey covered
6,000 establishments, primarily large ones, from 1966 to 1974. While the data are
not comparable to the time-use diaries cited by Stafford and Duncan, the sample size
and frequency is considerably larger. See ibid.

-0 .0 3
- 0 .0 3
0.07

S o u rc e : C o m p u te d b y a u th o r s a s e x p la in e d in th e te x t, u s in g th e m e th o d d e s c rib e d in F r a n k M . G o llo p
B e a t r i c e N . V a c c a r a , e d s . , N e w D e v e lo p m e n ts in P r o d u c tiv ity M e a s u r e m e n t ( N a t i o n a l B u r e a u o f E c o n o m i c
t h e U .S . B u r e a u o f t h e C e n s u s a n d t h e B u r e a u o f L a b o r

33. The data have not been developed at this time to measure the interaction for
1977 and years following. To use only the 1973-76 period would be inappropriate.




0.08
-0 .2 7
-0 .2 3

a.

a n d D a l e W . J o r g e n s o n , “ U .S . P r o d u c t i v i t y G r o w t h b y I n d u s t r y , 1 9 4 7 - 1 9 7 3 ," in J o h n W . K e n d r i c k a n d
R e s e a rc h , fo rth c o m in g ). T h e b a s ic d a ta a r e f r o m
S ta tis tic s .

- 0 .0 2
- 0 .0 0
-0.13*

S o u r c e : C o m p u te d b y a u t h o r s a s d e s c r ib e d in th e te x t, u s in g d a t a f r o m
th e B u r e a u o f L a b o r S ta tis tic s ,

Manufacturing

0.20
0.07
0.11

-0 .1 1
- 0 .0 7
- 0 .2 3

Manufacturing
1948-65
1965-73
1973-78

Interindustry
Labor Interindustry
Labor
Labor
Interindustry
shifts
Period composition
shifts
shifts
composition
composition
1948-65
1965-73
1973-78

Age

Private nonfarm business
1948-65
1965-73
1973-78

Annual average, in percent
Private nonfarm business

Employment
class o f worker

Private business
1948-65
1965-73
1973-78

Table 7. Rates of Growth of Adjustments to Total Hours for Changes in Labor
Composition and for Interindustry Shifts, by Sector, Selected Periods, 1948-78

Private business

Occupa­
tion

Sex

Sector and period

The ratio of hourly compensation between categories is assumed to be
equal to the ratio of marginal products for each category of labor.
Two sets of indexes are computed for each of the major sectors: one
for changes in sex, age, education, occupation, and class of worker; the
other for changes in labor input among industries, qLl.
Separation of the industry adjustment from the adjustment for labor
composition assumes independence between them. This assumption was
investigated by calculating a measure of labor composition using industry
as one of the characteristics. If independence exists, no difference occurs
between this measure and the sum of the two measures we have used, qLl
and qLC There was virtually no difference for the private business and
.
private nonfarm business sectors in either the 1948-65 or the 1965-73
periods.33 This was not the case in the manufacturing sector, where a sig­
nificant interaction seemed to occur between qL and qLI. For all sectors,
G
the measured interaction term was added to the adjustment for labor
composition.
Table 7 indicates the annual growth rates for adjusted labor composi­
tion and adjusted interindustry shifts as computed above. (These growth
rates have not been weighted by labor’s share, wL.) The contribution to
labor productivity provided by the changing composition of the labor
force decreased by more than 50 percent for all sectors from the 1948-65
to 1965-73 period and increased in 1973-78. The contribution of inter­
industry shift, on the other hand, increased significantly from the first to
second period for the private business and private nonfarm business sec­
tors, then decreased substantially in 1973-78. Interindustry shift has had
little effect in the manufacturing sector.
To obtain a better understanding of the cause for the changes in labor
composition, we also examined the separate direct effects of age, sex, edu­
cation, occupation, and class of worker, as shown in table 8. These
growth rates show the composition adjustment separately for the specific
characteristics. The effects are not simply additive to qL because they are
C
not independent; however, they do show which characteristics exhibited
the largest effect on the change in the labor composition and the direction
of the effects.
The growth rates presented in table 8 show that age was the major
factor contributing to the downward adjustment from labor composition
for the first period of slowdown. In all three sectors this characteristic
went from a positive to a negative annual growth rate, corresponding
directly to the large increase of young workers as the postwar baby-boom
cohort entered the labor market. For the private nonfarm and private
business sectors the age factor reversed itself in the third period, but the
increase in female entrants to the labor force seemed to compensate for
this reversal. Especially rapid entry of females took place in nonfarm­
nonmanufacturing industries, an area that has historically shown a smaller
increase in productivity. This development did not affect the manufactur­
ing sector. However, the age factor did continue to depress the composi­
tion of the labor force in manufacturing for the third period. Educational

36

perhaps the strongest proponent of this view and his quantitative esti­
mate of the effects of R&D is the largest.42 Kendrick regresses the totalfactor productivity (TFP) residual on a measure of the stock of ac­
cumulated knowledge. The quantitative estimates from this procedure
depend upon how one quantifies knowledge and on how one defines
TFP: if, as in Kendrick’s case, it is defined as the ratio of output to the
sum of share-weighted factor inputs, the effect will be relatively large; if,
as in our analysis, factor-augmenting effects are removed from TFP, the
effect will be smaller. In either case, the regression will attribute to R&D
the effects of all unaccounted factors insofar as they have similar inter­
temporal patterns. On the other hand, to the extent that the effects of
R&D can be seen in capital or labor or change the capital-labor ratio,
some of the effect may be missed by attributing it to other factors in the
analysis. It is not clear what approach, if any, can be relied upon to cap­
ture all the effects. Thus, although there seems to be a consensus that the
decline in R&D expenditures is partially responsible for the slowdown in
productivity growth, we found no satisfactory way to include the effect in
our analysis.

Table 9. Rates of Change in the Ratio of Hours Worked to Hours Paid, by Sector,
Selected Periods, 1952-75

Annual average, in percent
-

Sector

1952-65

Private business
Private nonfarm business
Manufacturing
S o u rc e s:

C o m p u te d

by

a u th o rs

1965-73

1973-75

- 0 .0 8
- 0 .0 6
-0 .0 6

-0 .2 2
-0 .2 1
- 0 .4 0

-0 .1 4
- 0 .1 2
0.03

fro m

d a ta in

B u re a u

o f L a b o r S ta tis tic s ,

“ R e p o rt o f th e

B L S T ask

F o r c e o n H o u r s W o r k e d ” ( B L S , M a r c h 1 9 7 6 ).

E N E R G Y AND

PR O D U CTIV ITY

We can make only a limited appraisal of the impact of higher energy
prices on the growth of labor productivity in the private business and
private nonfarm business sectors. Data on energy use are not available
by sector, but rather by the following categories: industrial, commer­
cial, transportation, and residential. These categories have not been
mapped into the major economic sectors with sufficient accuracy to justify
their inclusion in the productivity accounting framework. In addition,
our framework uses a concept of output based on gross product originat­
ing, so that flows of intermediate products— including energy— are ex­
cluded, although value-added is included in the energy-producing sectors.
It is possible to appraise the effects of energy price increases based
on the energy share in output in the major sectors, as Denison has done.
However, his procedure implicitly assumes that the elasticity of substitu­
tion between energy and other factors is one, and strong evidence exists
to the contrary, at least for the manufacturing sector. Bemdt and Wood
and Hudson and Jorgenson find complementarity between energy and
capital in U.S. manufacturing;383Griffin and Gregory, using cross-section
9
and time-series data for several countries, find substitution.30 Our own
recent investigation relied on a dynamic adjustment model of the manu­
facturing sector in an attempt to remove the short-term complementary
use of capital and energy suggested by Griffin and Gregory as a major
cause of the Berndt and Wood findings. We found stronger complemen­
tarity in the long-run than in the short-run version of the model.40
Using this model of the manufacturing sector, we undertook a simula­
tion exercise for the 1973-78 period to assess the effects of increases in
energy prices on the growth of labor productivity as these effects operate
through changing the capital-labor ratio. Whatever actual effect energy
prices have had on this ratio is included in the total estimated effect of
capital formation on productivity. Here we suggest how much of that
may be attributable to higher energy prices. The simulation assumes that
energy prices rose at the same rate as the implicit price deflator for manu­
facturing rather than at the 22.3 percent rate that actually occurred. On
this basis, the model suggests that the capital-labor ratio would have in­
creased at an annual rate of about 2.3 percent instead of 1.7 percent. Thus
labor productivity would have risen about 0.18 percentage point a year
faster in manufacturing during 1973-78 if the relative price of energy
had not changed. Hudson and Jorgenson also find a large reduction in
investment for. the 1972-76 period resulting from higher energy prices.
Their study, which uses a more complete model of the economy, includes
complementarity between energy and capital.41

Accounting for the Slowdown
As the preceding discussion indicates, some hypotheses about the
causes of the productivity slowdown defy quantification. In table 10 we
present the estimated effects of those factors that could be incorporated
into this analysis for the private business, private nonfarm business, and
manufacturing sectors. All three sectors show significant declines in labor
productivity for both slowdown periods: the, total effect of those slow■
downs is smallest in manufacturing and greatest in private business,
where the farm-to-nonfarm shifts of labor and capital contributed sub­
stantially to growth before 1965. In private business and private nonfarm
business, total-factor productivity growth, the “other factors” category in
the table, declines very little between 1965-73 and 1973-78.
The changes in the growth rates from table 10 are presented in table
11 as a way of detailing the contributions to the productivity slowdown
from the various factors analyzed. One conclusion is immediate— two
slowdowns occurred with two different patterns of contributing causes:
the 1965-73 slowdown is largely unexplained by factors quantified in
this analysis; the 1973-78-slowdown is largely accounted for by the rela­
tive weakness in capital formation.
In the private business sector, the broadest aggregate, the total effects
from capital formation augmented productivity growth in the first slow­
down period; the effect of changes in capital composition more than
compensated for the slight impacts of expenditures for pollution abate­
ment and the capital-labor ratio. The latter effect was due entirely to
slower growth in the capital-labor ratio in the farm sector, where the
growth of the capital-labor ratio slowed largely because the rapid migra­
tion of labor from the farm sector had ended. Labor effects in the first
slowdown period in the private business sector were small, although they
contributed somewhat to the slowdown. Favorable interindustry shift
effects were more than offset by a decline in the ratio of hours worked to
hours paid and changes in the composition of the labor force. The domi­
nant effect in the first slowdown period comes from other factors, which
account for more than 90 percent of the total aecline in the growth of
labor productivity.
Different factors account for the productivity slowdown in the second
period.,.Capital effects account for 0.79 percentage point out of a total
decline of 1.12 percentage points. In this period the decline in growth of
the .eapitaHabor. ratio contributes the largest effect, but changes in the
asset and interindustry composition also add to the slowdown, and capi­
tal spending for pollution abatement makes a small negative contribution
as well. Labpx.effects contribute somewhat more to the 1973— slow­
78
down than in the earlier period, but the pattern is quite different. Changes
in the composition of the labor force resulting largely from increased

R E S E A R C H AND D E V E L O P M E N T

A number of investigators have argued that research and development
expenditures have important effects on productivity growth. Kendrick is
38. Ernst R. Berndt and David O. Wood, “Technology, Prices, and the Derived
Demand for Energy,” Review of Economics and Statistics, vol. 57 (August 1975),
pp. 259-68; Hudson and Jorgenson, “Energy Prices.”
39. James M. Griffin and Paul R. Gregory, “An Intercountry Translog Model of '
Energy Substitution Responses,” American Economic Review, vol. 66 (December
1976), pp. 845-57.
40. J. R. Norsworthy and Michael J. Harper, “Productivity Growth in Manufac­
turing in the 1980’s: Labor, Capital, and Energy, in American Statistical Association,
Proceedings of the Business and Economic Statistics Section (Washington, D.C.:
ASA, forthcoming). The study was based on a four-factor model (capital, labor,
energy, and intermediate materials) of manufacturing using energy data from
Bureau of the Census, Census o f Manufactures, for 1958, 1963, 1967, 1972, and
1977; and Bureau of the Census, Annual Survey of Manufactures, for intermediate
years.




41. Hudson and Jorgenson, “Energy Prices,” p. 1.33.
42. John W. Kendrick, The Formation and Stocks of Total Capital, General
Series, 100 (National Bureau of Economic Research, 1976). Kendrick represents the
stock of accumulated knowledge by the capitalized value of research and develop­
ment expenditures.

37

Table 10. Rates of Growth of Labor Productivity and of Capital and Labor Effects on Productivity, by Sector, Selected Periods, 1948-78

Annual average, in percent
Private business
Item

Private nonfarm business

1948-65

1965-73

1973-78

1948-65

1965-73

1973-78

Labor productivity growth

3.32

2.32

1.20

2.77

2.02

1.09

Capital effect
Capital-labor ratio
Asset composition
Intersectoral shifts
Pollution abatement capital
Total

0.76
0.06
0.12
0.00*
0.94

0.75
0.10
0.17
- 0 .0 3 “
0.99

0.21
0.05
0.03
- 0 .0 9
0.20

0.68
0.05
0.00
-0.01*
0.72

0.84
0.08
0.00
-0.04*
0.88

0.12
0.15

0.06
0.20

0.10
-0 .0 7

0.12
- 0 .0 2

- 0 .0 5
0.22

- 0 .1 4
0.12

- 0 .0 9
- 0 .0 6

2.16

1.21

1.06

Manufacturing

Labor effect
Labor force composition
Interindustry shifts
Ratio of hours worked to hours
paid
Total
Effect o f other factors

1948-65

1965-73

1973-78

3.13

2.47

1.70

0.20
0.04
0.00
- 0 .0 9
0.15

0.54
0.03
n.a.
-0 .0 3 *
0.54

0.72
0.05
n.a.
-0.09*
0.68

0.44
0.03
n.a.
- 0 .1 9
0.28

0.01
0.12

0.04
- 0 .0 8

0.14
- 0 .0 2

0.05
- 0 .0 2

0.08
0.05

- 0 .0 4
0.06

- 0 .1 4
-0 .0 1

- 0 .0 8
- 0 .1 2

- 0 .0 4
0.08

- 0 .0 7
- 0 .0 4

0.02
0.15

1.99

1.15

1.06

2.51

1.83

1.27

S o u rc e s : C o m p u te d b y a u th o r s a s e x p la in e d in th e te x t, u s in g d a ta fr o m th e B u re a u o f E c o n o m ic A n a ly s is a n d th e B u re a u o f L a b o r S ta tis tic s ,
a . A s in d i c a t e d i n t h e te x t, e s tim a te s f o r th e s e p e r i o d s a r e b a s e d o n p a r t i a l d a t a ,

V

n .a . N o t a v a ila b le .

education have a positive effect on productivity growth, as does the ratio
of hours worked to hours paid (though, again, the data underlying this
latter estimate are weak). Interindustry shifts of the labor force have a
strong negative influence. Other factors play a much smaller role than in
the 1965-73 slowdown; only about 13 percent of the 1973-78 slowdown
in the private business sector is not accounted for by the measured capital
and labor effects.
i

downward push on productivity from interindustry shifts more than off­
setting small contributions in the other direction from the composition of
the labor force and changes in hours worked. As in the private business
sector, the measured capital and labor effects account for most of the
1973— slowdown in productivity growth in the private nonfarm busi­
78
ness sector.

In the private nonfarm business sector the pattern in the first slowdown
period is'generally similar to that for the private business sector, although
capital effects are even more favorable to productivity growth because
the capital-labor ratio grows more rapidly in 1965-73 than in 194865. The pattern of labor effects is quite similar to that in private business,
although the net impact is slightly smaller. And other factors are again
the dominant slowdown factor. Indeed, after adjusting for capital and
labor effects, the contribution to the slowdown of other factors is some­
what larger than the slowdown in labor productivity itself.

The productivity slowdown pattern in the manufacturing sector is
similar to that for private nonfarm business in 1965-73: capital effects
contribute to faster productivity growth, and total labor effects reduce
it. During this period, the acceleration of the capital-labor ratio in­
creased productivity growth by about 0.2 percentage point a year, but was
partially offset by expenditures for pollution abatement capital and a slight
asset effect. Labor effects made a small contribution to the slowdown,
largely through changes in the composition of the labor force. Other
factors not accounted for in the analysis dominate the productivity decline
in manufacturing in the first slowdown period.

In the second period, capital effects contribute nearly 80 percent of the
observed slowdown in labor productivity. As in the private business sec­
tor, the dominant impact comes from slower growth in the capital-labor
ratio. Capital spending for pollution abatement and changes in the asset
mix each have a small effect. Labor effects contribute somewhat, with a

In the 1973-78 period, some differences emerge between the manu­
facturing sector and the private nonfarm business sector. The productiv­
ity slowdown is somewhat smaller in manufacturing. Capital effects,
dominated by slower growth in the capital-labor ratio, are more strongly
influenced by expenditures for pollution abatement capital. The effect,

Table 11. Contributions to the Slowdown in the Growth of Labor Productivity, by Sector, 1965-78*
A n n u a l average, in percentage points

Private business
Item

1965-73 1973- 78
slowdown slowdown

Private nonfarm business
Total

1965-73
1973-78
slowdown slowdown

Total

Manufacturing
1965-73
1973-78
slowdown slowdown

Total

Change in labor productivity growth

- 1 .0 0

-1 .1 2

- 2 .1 2

-0 .7 5

-0 .9 3

-1 .6 8

-0 .6 6

-0 .7 7

-1 .4 3

Contribution from capital effect
Capital-labor ratio
Asset composition
Intersectoral shifts
Pollution abatement capital
Total

-0 .0 1
0.04
0.05
-0 .0 3
0.05

- 0 .5 4
- 0 .0 5
- 0 .1 4
- 0 .0 6
- 0 .7 9

- 0 .5 5
-0 .0 1
- 0 .0 9
- 0 .0 9
- 0 .7 4

0.16
0.03
0.00
-0 .0 3
0.16

-0 .6 4
- 0 .0 4
0.00
-0 .0 5
-0 .7 3

-0 .4 8
-0 .0 1
0.00
-0 .0 8
-0 .5 7

0.18
0.02
n.a.
-0 .0 6
0.14

-0 .2 8
- 0 .0 2
n.a.
- 0 .1 0
- 0 .4 0

-0 .1 0
0.00
n.a.
-0 .1 6
- 0 .2 6

Contribution from labor effect
Labor force composition
Interindustry shifts
Ratio of hours worked to hours
paid
Total

- 0 .0 6
0.05

0.04
-0 .2 7

- 0 .0 2
- 0 .2 2

-0 .1 1
0.14

0.03
- 0 .2 0

-0 .0 8
- 0 .0 6

-0 .0 9
0.00

0.03
0.07

- 0 .0 6
0.07

-0 .0 9
-0 .1 0

0.05
-0 .1 8

- 0 .0 4
-0 .2 8

- 0 .1 0
-0 .0 7

0.06
-0 .1 1

- 0 .0 4
-0 .1 8

-0 .0 3
- 0 .1 2

0.09
0.19

0.06
0.07

Contribution from effect o f other
factors

-0 .9 5

-0 .1 5

- 1 .1 0

- 0 .8 4

- 0 .0 9

-0 .9 3

-0 .6 8

- 0 .5 6

- 1 .2 4

S o u r c e : D e r i v e d f r o m t h e r e s u l t s o f t a b l e 10.
a . T h e 1 9 6 5 - 7 3 s l o w d o w n is m e a s u r e d r e la tiv e t o t h e 1 9 4 8 - 6 5 b a s e p e r i o d ; t h e 1 9 7 3 - 7 8 s l o w d o w n , r e l a t i v e t o t h e 1 9 6 5 - 7 3 s l o w d o w n ,
n .a .

N o t a v a ila b le .




38

however, is still small in manufacturing, where a major impact of envi­
ronmental regulations would be expected to be felt. Capital effects, how­
ever, explain only about half of the 1973-78 slowdown in manufacturj ing, a much smaller proportion than in private nonfarm business. The
labor effects augmented productivity growth. Thus, in the second period,
j other factors play a larger role in the slowdown of the manufacturing secj tor than in the other sectors.
I
The slowdown patterns for the nonfarm nonmanufacturing sector that
are implied by the nonfarm and manufacturing results in tables 10 and
11 are summarized in table 12. The capital effects were determined by
weighting the capital effects in each sector by the relative size "of their
capital stock. A similar procedure was used for labor productivity and
labor effects based On the nonfarm nonmanufacturing share in the hours
of private nonfarm business labor.43
In this sector, productivity again slows noticeably in both periods. In
the first period, total capital effects work against the slowdown, while
total labor effects contribute to it and are noticeably larger than those
for all nonfarm business. As in the private nonfarm business sector, other
x factors are the primary source of the decline in productivity growth. Al­
most all the second slowdown is explained by capital and labor effects
that parallel those in total private nonfarm business, so that other factors
play a minor role.
For any of the major sectors analyzed here, to view the productivity
slowdown as a single phenomenon beginning in the mid-1960s would dis­
tort the temporal pattern of contributions to it, and would likely lead to
poor policy prescription. From the evidence of the recent period, the un­
explained decline in multifactor productivity growth is largely behind us,
while the problem of capital formation is current. It also appears that the
changing composition of the labor force has contributed somewhat less to
the slowdown in either period than some other estimates have suggested
and, correspondingly, may offer somewhat less hope for reversal in the
future.

Table 13. Rates of Growth of Input Prices, Private Nonfarm Business Sector,
Selected Periods, 1948-78

Annual average, in percent

Period

2.57
1.78
0.80

0.78
0.95
0.11

0.11
- 0 .1 0
-0 .2 2

o f p u r c h a s e d f u e ls in m a n u f a c tu r in g

-0 .7 9
-0 .9 8

0.17
-0 .8 4

-0 .2 1
-0 .1 2

- 1 .7 7

-0 .6 7

-0 .3 3

The weak productivity growth of recent years has corresponded with
a rapid and continued rise in employment from the trough of the 197375 recession through early 1979. This phenomenon, which has been
widely observed and described as puzzling, is consistent with the much
closer movements in the prices of capital and labor and the complemen­
tarity between capital and energy. Under these conditions, increases in
output would be achieved with relatively greater expansion of labor input
and less expansion of capital (and hence energy) than under the price
conditions that prevailed since 1948 in general, and in the 1965-73 period
in particular. This tentative explanation is consistent with findings by
Hudson and Jorgenson for the 1973-76 period.46
The main conclusions of our investigation of the slowdown in the
growth of labor productivity can be summarized briefly.
There are two distinct phases to the slowdown in the growth of labor
productivity: 1965-73 and 1973-78. Differences are apparent both in
the pattern of productivity growth among industries and in the factors
contributing to the decline.

-0 .7 7

The 1965-73 slowdown is largely unexplained by the factors we have
considered. Capital formation was not a cause; changes in the composi­
tion of the labor force played a relatively minor role. Although R&D ex­
penditures slowed during this period and may well have contributed to
the productivity slowdown, we devised no satisfactory means to take this
factor into account. Intersectoral shifts of capital and labor did not con­
tribute.

Contribution to slowdown
(percentage points)

ta b le s 10 a n d

The 1973-78 slowdown is dominated by the effects of reduced capital
formation. Some effect is also attributable to interindustry shifts in labor
and capital. The sharp rise in energy prices may show up in a framework
such as ours through its impact on capital formation and may help ex­
plain the relative weakness in capital formation in recent years.
V*

11 a s d e s c r i b e d i n t h e t e x t .

43. Direct computation would have been preferable. However, the difference in
patterns between the nonfarm business and manufacturing sectors did not emerge
until it was too late to compute these effects directly. Although the total effects for
capital, labor, and other factors reported in table 12 would change very little when
directly computed, detailed effects of changes in factor composition and interindus­
try shifts would be revealed.
44. Data for the private nonfarm business sector are shown because the slowdown
in capital formation in agriculture began before 1973-78, and the relative rise of
wages in the agriculture sector further obscures the relative price movements that
prevailed in private nonfarm business.




n a tu ra l

The relative price explanation for the acceleration in the capital-labor
ratio in the 1965-73 period also explains the deceleration in 1973-78,
when the relative price change was so small. The rapid rise in energy
prices that took place in late 1973 and early 1974 may be another impor­
tant factor contributing to the slowdown in this last period. If capital and
energy are complements, the rise in energy prices would have retarded
capital formation.45

Rate o f growth (percent)

S o u rc e : D a ta a r e in fe rre d fro m

1 9 7 3 - 7 7 . I t is b a s e d o n t h e r a t i o o f t o t a l c o s t

in d e x o f e le c tr ic ity , c o a l, c o k e , f u e l o il, a n d

labor grew about 2 percentage points a year faster than the price of capital
services in 1948-65, more than 4 points faster in 1965-73, and 1 point
faster in 1973-78. These differences measure the relative price change of
labor as compared to capital: the price incentive to substitute capital for
labor was thus about twice as strong in 1965-73 as it was in the earliest
period, and about four times as great as in 1973-78. A factor holding
down the relative price of capital services in 1965-73 was the investment
tax credit for equipment that went into effect in the mid-1960s.

-0 .7 5
-0 .0 2

Total

to a tr a n s lo g

B u r e a u o f t h e C e n s u s , C e n su s o f M a n u fa c tu r e s . T h e w h o l e s a l e p r i c e i n d e x w a s u s e d t o

in te r p o la te s o m e p r ic e c o m p o n e n ts b e tw e e n C e n s u s b e n c h m a r k s .

1.68
0.93
0.91

1965-73 slowdown
1973-78 slowdown

-0 .7 3
4.73
22.29

g a s q u a n titie s fro m

Effect o f
other factors

1948-65
1965-73
1973-78

4.60
6.58
8.98

S o u rc e : J . R . N o rs w o rth y a n d M ic h a e l J . H a rp e r, “ T h e R o le o f C a p ita l F o r m a tio n in th e R e c e n t P r o ­

Annual average
Labor
effect

2.84
2.20
7.95

a . T h e e n e r g y p r ic e s e r ie s is f o r 1 9 5 4 - 6 5 , 1 9 6 5 - 7 3 , a n d

Table 12. Capital and Labor Effects on the Growth of Labor Productivity, Private
Nonfarm Nonmanufacturing Sector, Selected Periods, 1948-78

Capital
effect

Price o f
energy input“

d u c tiv ity G r o w t h S lo w d o w n ,” W o r k in g P a p e r 8 7 , ( B u r e a u o f L a b o r S ta tis tic s , J a n u a r y 1 9 7 9 ).

Because slower capital formation appears to have been a major cause
of the slowdown in labor productivity in the 1973-78 period, it is impor­
tant to understand why. Table 13 attempts to shed light on this ques­
tion.41
The acceleration of the capital-labor ratio in 1965-73 may be ex­
plained by price-induced substitution of capital for labor. The price of

Total labor
productivity

Labor
compensation
per hour

1948-65
1965-73
1973-78

Factors Affecting Capital Formation

Item and period

Price o f
capital services

45. There are, of course, other dimensions to the problem, and therefore to a sat­
isfactory explanation for it. For example, an accelerator model of capital accumula­
tion is examined in Peter K. Clark, “Investment in the 1970s: Theory, Performance,
and Prediction,” BPEA, 1:1979, pp. 73-113. The slowdown in output growth of more
than 1 percentage point a year between 1965-73 and 1973-78 would also explain
part of the slowdown in capital formation in the latter period.
46. Hudson and Jorgenson, “Energy Prices.”

39

Fart III. Productivity Trends
nn SndMstries and the
Federal Government

number of studies of labor requirements for defense in­
dustries, such as synthetic rubber and shipbuilding.
After the war, the industry studies program resumed on
a regular basis, and was supplemented by a number of
industry studies based on the direct collection of data
from employers. Budget restrictions after 1952
prevented the continuation of direct collection of data.
Consequently, the preparation of industry measures is
largely limited to those industries where readily
available data can be used to construct measures.
In recent years, public interest in productivity has
grown, and increases in output per employee hour have
been recognized as important indicators of economic
progress and a means to higher income levels, rather
than merely a threat to job opportunities.
The industry studies cover a variety of manufacturing
and nonmanufacturing industries at the 2-, 3-, and
4-digit Standard Industrial Classification level.
Measures for these industries are published on an an­
nual basis and are provided for most years between 1947
or 1958 and the most recent year for which data are
available.
Coverage has been expanded to include industries in
trade and services, and with the increasing importance
of the public sector, to various functional areas in the
Federal Government. Productivity measurement in the
Federal Government was initiated by a request from the
Joint Economic Committee in the fall of 1970 to the
General Accounting Office in conjunction with the Of­
fice of Management and Budget and the Civil Service
Commission (now Office of Personnel Management). A
joint Federal productivity measurement task force con­
sisting of-these agencies, with technical assistance and
support from b l s , was established. This task force col­
lected data and constructed indexes for fiscal years
1967-71. In July 1973, the Office of Management and
Budget endorsed the continuation of the project to
measure Federal productivity, and b l s assumed full
responsibility for collecting input, output, and related
information, in addition to the development of produc­
tivity measures.
Since July 1973, the Bureau has been expanding
coverage to include organizational units not previously
covered, improving the quality of some of the input and
output data, and refining the methodological pro­

This section includes a brief historical introduction to
productivity measurement in industries and in the
Federal Government. Methods, sources, and uses of the
measures are discussed and an excerpt from the most
recently published compilation of industry productivity
measures serves as an example of recent results.
One essay analyzes the difficulties in measuring pro­
ductivity in service industries and explores their solu­
tion. The complexities of measuring output per hour in
government are also discussed. The articles on labor
productivity trends in several industries examine the fac­
tors underlying productivity change over time.
B a c k g ro u n d
Studies of output per employee hour in individual in­
dustries have long been a part of the b l s program. A
study of 60 manufacturing industries in 1898, prompted
by congressional concern that human labor was being
displaced by machinery, was presented in the report
Hand and Machine Labor; this provided striking
evidence of the savings in labor resulting from
mechanization in the last half of the 19th century. The
impact of productivity advance upon employment re­
mained an important focus of b l s throughout the.
1920’s and 19305 Also during this period, the Bureau
s.
began the preparation and publication of industry in­
dexes of output per employee hour, which were based
on available production data from the periodic Census
of Manufactures and employment statistics collected by
BLS.

In 1940, Congress authorized the Bureau of Labor
Statistics to undertake continuing studies of productivi­
ty and technological changes. The Bureau extended
earlier indexes of output per employee hour developed
by the National Research Project of the Works Progress
Administration, and published measures for selected in­
dustries. This work, however, was reduced in volume
during World War II, owing to the lack of meaningful
production and employee hour data for many manufac­
turing industries.
The advent of World War II also caused a change in
the emphasis of the program from problems of
unemployment to concern with the most efficient
utilization of scarce labor resources, b l s undertook a



40

cedures used to construct productivity indexes. The
measurement program is part of a multifaceted effort
sponsored by the Office of Personnel Management and
includes analysis, enhancement, and diffusion of pro­
ductivity improvement ideas.
The Bureau is also expanding its productivity
measurement program by explicitly accounting for
other inputs besides labor in the industry measures and
developing a supplementary set of productivity
measures. The new measures are referred to as multifac­
tor productivity measures.
The industry multifactor series are designed to
measure changes in productivity by relating changes in
an industry’s output not only to changes in labor input
but also to changes in capital and intermediate pur­
chases. In addition to providing indicators of produc­
tivity change useful for analysis in their own right, such
measures also are helpful in analyzing the causes of
change in output per employee hour or labor productivity.

liethods and Sources
industries
Output per employee hour
Bls computes an index of output per employee hour
by dividing an output index by an index of aggregate
employee hours. For most industries, measures are
prepared separately relating output to (a) all employee
hours, (b) production worker hours, and (c) nonproduc­
tion worker hours. (The standard definitions of produc­
tion workers and nonproduction workers are used.)
Three corresponding measures also are computed
relating output to the number of employees. For in­
dustries in trade and services, measures are prepared
relating output to the hours of all persons involved in
producing that output, including self-employed and un­
paid family workers.
Output
Bls industry output indexes are based on quantifiable
units of products or services of the industry combined
with fixed period weights. Whenever possible, physical
quantities are used as the unit of measurement. For
those industries lacking quantity data, constant-dollar
value of shipments, sales, or revenue data are used to
develop the output series. This procedure is used almost
exclusively for the nongoods-producing industries. For
manufacturing and mining industries, quantity data on
physical output are usually most comprehensive for
years covered by a census. To make maximum use of the
comprehensive census data, output indexes are derived
from data for two consecutive censuses; these indexes
are referred to as benchmark indexes. For intercensal
years, annual indexes are based on either physical out­
put data (generally in less detail than for census years)



or, if such data are not available, value of output ad­
justed for price change (the value of output in constant
dollars). The annual series subsequently are adjusted to
the benchmark levels for the census years.
Sources. Industry output indexes are prepared from
basic data published by various public and private agen­
cies, using the greatest level of detail available.
Data from the Bureau of the Census, U.S. Depart­
ment of Commerce, are used extensively in developing
output statistics for manufacturing, trade, and service
industries. The Bureau of Mines, U.S. Department of
the Interior, compiles most of the information for the
mining and cement industries. Other important Govern­
ment sources include the U.S. Department of Energy,
the Department of Agriculture, the Fish and Wildlife
Service, U.S. Department of the Interior, the Interstate
Commerce Commission, the Internal Revenue Service,
and the Civil Aeronautics Board. Important sources of
trade association data include the Textile Economics
Bureau, Inc., N ational Association of Hosiery
Manufacturers, Inc., National Canners Association,
Rubber Manufacturers Association, and the American
Iron and Steel Institute.
For deflated value series, industry price indexes are
derived from producer and consumer price indexes
developed by the Bureau of Labor Statistics.
Employee hours
An index of employee hours is computed by dividing
the aggregate employee hours for each year by the baseperiod aggregate. Employee hours are treated as
homogeneous and additive with no distinction made
between hours of different groups of employees. Data
on changes in qualitative aspects of employee hours,
such as skill, efficiency, health, experience, age, and sex
of persons comprising the aggregate are not used and
generally are not available. For mining and manufactur­
ing industries, employee hour indexes are constructed
for employees, production workers, and nonproduction
workers. For service and trade industries, indexes are
constructed for the hours of all persons, which includes
paid employees, partners, proprietors, and unpaid fami­
ly workers.
Sources. Industry employment and employee hour in­
dexes are developed from basic data compiled by the
Bureau of Labor Statistics or the Bureau of the Census.
For trade and service industries, these data are sup­
plemented with data from the Internal Revenue Service.
Federal Gowerninent
Indexes of output per employee year, output, and
employee years for selected functional areas of Govern­
ment activity4 and for the more than 400 participating

41

ment. Therefore, the overall statistics do not represent
‘‘Federal
productivity” but rather, the weighted
average of the productivity changes of the measured
Federal organizations included in the sample.

organizations are constructed in a manner similar to
that described for industries. At the present time, these
measures cover about 67 percent (1.9 million employee
years) of the Federal civilian work force.
Ideally, a productivity index should relate final out­
puts to their associated direct and indirect input(s), and,
in fact, the output data are final from the perspective of
the funtional areas within which these data are
classified. However, since the outputs of one organiza­
tion may be consumed wholly or partially by another
Federal organization in the production of its final out­
puts, all output indicators in the Federal sample may
not be final from the perspective of a higher level
organization; for example, the entire Federal Govern-

Presentation
Bls industry indexes are published annually in the
bulletin, Productivity Measures fo r Selected Industries.
A limited amount of the most current data is provided
in an annual news release. As new industry indexes are
developed, they are presented as articles in the Monthly
Labor Review. The articles contain an analysis of pro­
ductivity, output, and employment trends in the in­
dustry. Technical notes describing the methodology
used to develop the indexes are available on request.
Unpublished indexes for all 4-, 3-, and 2-digit sic
manufacturing industries are available for analytical
purposes upon request. Federal Government indexes are
published annually in the Monthly Labor Review.
Indexes of output per employee hour also are publish­
ed in the Statistical Abstract o f the United States and in
the Handbook o f Labor Statistics. Some indexes for
earlier years are published in Historical Statistics o f the
United States.

4 The 28 functions are:
Audit of operations; Buildings and grounds; Communications;
Education and training; Electric power production and distribution;
Equipment maintenance; Finance and accounting; General support
services; Information services; Legal and judicial activities; Library
services; Loans and grants; Medical services; Military base services;
Natural resources and environmental management; Personnel in­
vestigations; Personnel management; Postal services; Printing and
duplication; Procurem ent; Records m anagem ent; R egula­
tion—compliance and enforcement; Regulation— rulemaking and
licensing; Social services and benefits; Specialized manufacturing;
Supply and inventory control; Traffic management; and Transpota­
tion.




42

Highlights of Recent
Trends in Output Per
Employee Hour
Current developments
From 1980 to 1981, output per employee hour in­
creased in nearly two-thirds of the industries for which
data are presented in this report. In contrast, from 1979
to 1980, productivity fell in about two-thirds of the in­
dustries. The gains in industry productivity in 1981 were
consistent with the performance of the nonfarm busi­
ness sector as a whole, where productivity increased
by 1.4 percent.
Among the large manufacturing industries, the steel
industry recorded a gain of 9.0 percent in productivity
after two consecutive years of productivity declines.
Buoyed by strong sales to the oil and gas industry, steel
output was up 9.8 percent in 1981, while hours grew
by only 0.7 percent.
Another large manufacturing industry, motor vehi­
cles, experienced a 4.7-percent growth in output per
employee hour in 1981. This was in sharp contrast to
the three previous years, in which productivity fell.
Rebounding somewhat from a very poor year in 1980,
output in the industry increased by 5.9 percent while
employee hours grew by only 1.2 percent.
Among the other large manufacturing industries, the
tire and inner tubes industry posted a gain in produc­
tivity of 13.3 percent. Output in this industry was up
by 8.6 percent, sustained by demand from the replace­
ment market, while hours continued to decline. Aside
from demand strength, many old and inefficient tire
plants were closed in 1980, which contributed to the
sharp productivity gain 1981.
Other large manufacturing industries which recorded
increases in productivity in 1981 included: Synthetic fi­
bers (6.3 percent), bakery products (6.2 percent), gray
iron foundries (5.9 percent), fluid milk (4.7 percent),
fabricated structural metal (4.7 percent), corrugated and
solid fiber boxes (2.7 percent), paper, paperboard, and
pulp mills (2.0 percent), and electric motors and gen­
erators (1.3 percent).
However, three relatively large manufacturing indus­
tries posted productivity declines in 1981. In sawmills
and planing mills, output per employee hour was down
2.6 percent, while the footwear industry recorded a falloff in productivity of 3.6 percent. The construction
machinery and equipment industry also posted a drop
in productivity—5.4 percent in 1981.
All the mining industries covered experienced pro­
ductivity gains in 1981. Coal mining posted its second
Reprinted from BLS Bulletin 2155 (1982),
Productivity Measures f o r Selected Industries, 1954-81.




43

consecutive large productivity gain, rising 9.2 percent
in 1981. Although coal output was down slightly (-1.3
percent) from the previous year, hours continued to
decline sharply, resulting in the productivity gain. Pro­
ductivity advances in the other mining industries were
not as great as for coal. Iron mining (usable ore) rose
6.7 percent, copper mining (recoverable metal) in­
creased 5.5 percent, and nonmetallic minerals gained
1.7 percent. Both copper and iron mining had large
output increases in 1981, in contrast to sharp declines
in 1980. The productivity gain in nonmetallic minerals,
on the other hand, was based on a drop in output, due
to poor demand from the construction industry, and an
even larger decline in hours.
Productivity changes were mixed among transporta­
tion and utility industries. Productivity was up 5.4 per­
cent in telephone communications as output grew 5.6
percent. In railroads, productivity rose 5.2 percent. Out­
put in the railroad industry declined for the second
straight year, dropping 0.7 percent, while hours con­
tinued to fall, declining 5.6 percent. After four consecu­
tive years of declines, productivity in the trucking in­
dustry advanced 4.7 percent even though output
dropped 4.9 percent; the number of employees was
down by 9.2 percent. On the other hand, productivity
registered a small decline of 0.3 percent in the air trans­
portation industry. The general economic downturn and
the air traffic controllers’ strike contributed to the falloff in productivity. Productivity also fell in both gas
(-3.8 percent) and electric utilities (-0.7 percent). Out­
put was down in gas utilities as many consumers cur­
tailed usage due to rising prices, while hours were up
due to the growing number of customers. Output was
up in electric utilities only 0.8 percent, well below the
long-term rate of 6.6 percent, while hours grew 1.6 per­
cent, resulting in the productivity falloff. Productivity
dropped sharply (-8.3 percent) in petroleum pipelines
as output fell for the second consecutive year due to
declining demand for petroleum products, while hours
increased.
In trade and services, productivity changes also were
varied. Productivity grew 1.5 percent in retail food
stores, as output was up 1.9 percent and hours grew
0.4 percent. New car dealer productivity was up 1.4
percent. Gasoline service station productivity rose 1.2
percent. Output was down 2.1 percent in this industry,
as demand was off due to increased gasoline prices and

more fuel efficient cars, while hours fell even more as
marginal stations were closed and self-service stations
became more prevalent. Productivity in both eating and
drinking places and hotels and motels declined 1.9 per­
cent as small gains in output were more than compen­
sated for by larger gains in hours. Productivity in drug
stores fell 2.7 percent as output declined 1.9 percent
and hours were up slightly. In the laundry and clean­
ing industry, productivity fell 3.2 percent due to a con­
tinued, decline in demand for the industry’s services re­
sulting in an output drop of 7.2 percent, while hours
fell 4.2 percent.

terial handling procedures, a new technique for firing
tile became widespread. These changes resulted in sig­
nificant labor savings for the industry. The wood office
furniture industry posted a 7.0-percent annual growth
in productivity, far outstripping its rate of 1.1 percent
in the previous period. Productivity was enhanced by
substantial output growth in the late seventies as wood
furniture captured a large share of the market from
metal office furniture.
Among the industries that experienced productivity
declines during the 1976-81 period, the brick and struc­
tural clay tile industry dropped the most (-5.7 percent
per year). Other industries in which productivity de­
clined substantially (more than 2 percent per year) dur­
ing the period were: Cosmetics and other toiletries (-3.4
percent), laundry and cleaning services (-3.2 percent),
hydraulic cement (-2.5 percent), petroleum pipelines
(-2.2 percent), and blended and prepared flour (-2.1 per­
cent). (See table 1.)

Long-term ftreimdte
With one exception, all of the industries in this re­
port recorded average annual increases in productivity
over the long term (1947-81 for many of these indus­
tries). The metal forming machine tools industry re­
corded an average annual change in productivity of -0.1
percent over the 1958-81 period. A combination of fac­
tors worked against a productivity increase in the ma­
chine tools industry, including the tendency of machine
tool firms to keep highly skilled workers on the payroll
even when output falls during cyclical slowdowns. Ad­
ditionally, machine tools are not mass produced al­
though they may make mass production processes pos­
sible in user industries. The parts and components of a
finished machine tool are usually made in relatively
small batches and require comparatively large amounts
of labor.
On the other hand, the wet corn milling industry
posted the highest long-term rate of growth in produc­
tivity (7.1 percent per year). Growth in this industry’s
output and productivity was modest until 1972 when
an explosion in output resulted from the rapid market
penetration of high fructose and glucose corn syrups.
The industry, previously highly capital intensive, be­
came even more so with large infusions of capital.
Although the rates of productivity change varied
widely among the industries during the 1976-81 period
(see chart 1), three-fourths of the industries recorded
lower average annual gains in this more recent period
than during the 1947-76 period. This experience matches
the productivity record of the nonfarm business
economy as a whole. From 1947 to 1976, productivity
in this sector grew 2.3 percent per year while from 1976
to 1981 the annual growth rate was only 0.1 percent.
The slower growth in productivity among the indus­
tries included in this report was associated with less
rapid output growth in the later period.
However, there were a few notable exceptions to the
falloff in productivity growth during the 1976-81 pe­
riod. In addition to wet corn milling, the ceramic wall
and floor tile industry recorded a substantial improve­
ment in the rate of productivity growth in the last few
years—7.4 percent a year. Coupled with changes in ma­



measures
Millwork SIC 2431. Labor productivity in the
mill work industry rose at an average annual rate of 1.4
percent from 1958 to 1980, a modest advance when
compared with total manufacturing. Over this period,
output in millwork increased at a rate of 2.7 percent
annually and employee hours at 1.3 percent. The pro­
ductivity rise partly reflected low growth in capital in­
vestment, particularly over the past decade, and, evi­
dently, slow diffusion of modernized production tech­
nologies. These factors, combined with instability in
demand for the industry’s products, retarded produc­
tivity growth.
Approximately three-fourths of the industry’s output
is used in residential housing, including additions and
alterations; small amounts of output are used in com­
mercial and educational buildings, prefabricated
wooden buildings, and in trailers and other transporta­
tion equipment. Millwork output is thus linked mainly
to residential construction where fluctuations have been
frequent and substantial.
The basic technology used in millwork plants dates
from the 1930’s and 1940’s; however, advances in pro­
duction techniques have been numerous. For example,
the versatility of machines fabricating moldings has been
greatly extended so that a large variety of complex
molding profiles can be cut and grooved at great speeds
without loss of precision and insignificnt loss in setup
time. Significant advances have also been made in ad­
hesive applications. High-speed production requires
rapid curing, and the gluing process is therefore usu­
ally an integral part of the production process. Certain
radio-frequency gluing devices have reduced curing
time from 20 to 2 minutes. A modest trend toward au­
tomated systems in the industry is underway, furthered
44

C h art 1. G ro w th in o u tp u t p er e m p l o y e e h o u r in s e l e c t e d industries, 1976-81
(Average annual percent change)

Fluid milk
Telephone communications

■ ■/
'

-L

Iron mining, usable ore
Malt beverages
Prim ary copper, lead, and zinc

AH

.
0

■ .••••• , y • , y y y . ; ;- . . . y . y y ;
■
p-“ -

:
"v

.

1
. ' . __

■:■
■.~
■

.■
___!■-, .

r

:

.. - - 1 - : - . V

Coal mining
Tires and inner tubes
Flour and other grain mill products

.■ ?.

-

■ ■ n B B H |m
m
f
Metal cans y .V■a . < :;./;;.. :a:I a
Air transportation _____
Steel '-H A \C n':.
-r
~
' ■
Crushed and broken stone

Fabricated structural metal
Electric lamps

■■ .

Paints and allied products
All tobacco products
Machine tools
Franchised new car dealers
Retail food stores
Motor vehicles
Aluminum rolling and drawing
Gas and electric utilities
Construction machinery and equipment
Ball and roller bearings
JjglS teel foundries
Bpetroleum pipelines
(H ydrau lic cement
__MLaundry and cleaning services
(B rick and structural clay tile

by the declining costs of numerical conrols, which more
and more entail one-station systems featuring micro­
computers. This trend also involves computer-controlled materials handling systems, robotized transfer
and palletizing, and carousels interfacing with con­
veyors, robots, and other material handling devices.

tially below the 2.8-percent rate for all manufacturing.
The productivity gain in the office furniture industry
resulted from growth in output averaging 5.5 percent
annually and employee hours averaging 3.6 percent.
The growth in productivity in the industry has been
comparatively low, in large part because of relatively
short production runs engendered by product
proliferation.
A number of factors have shaped the demand for of­
fice furniture and the industry’s output growth. Some

Office furniture (SIC 252). Between 1958 and 1980,
output per employee hour in the office furniture indus­
try advanced at an annual rate of 1.8 percent, substan­



45

of these factors are the growth of the white-collar work
force, the amount of available office space, replacement
demand, and the introduction of new products. Un­
doubtedly, the most important factor influencing the
long-term output growth has been the increase of the
white-collar work force. During the period under study
(1958-80), white-collar employment increased from
about 27 million to nearly 53 million and currently ac­
counts for slightly more than one-half of the total em­
ployed work force.
Productivity growth in the industry has been en­
hanced by the increased use of more automatic machin­
ery; new materials such as particleboard, quick-setting
glues, and improved finishes; better workflow layout;
and computerized recordkeeping.
Cosmetics and other toiletries (SIC 2844). Productivity
increased at an average annual rate of 4.0 percent from
1958 to 1980—substantially above the 2.8-percent rate
for the manufacturing sector. This growth was associ­
ated with average annual increases of 7.3 percent in
output and 3.1 percent in employee hours. Productiv­
ity gains have resulted primarily from a trend toward
fewer and larger plants producing a greater level of
output, and continued improvements in production and
packaging operations.
The changes in output per employee hour have not
been steady. Since 1958, annual increases in productiv­
ity have ranged from a high of 14.9 percent to a low
of 0.4 percent. Declines in productivity occurred in 6
years, the most recent in 1980 when a drop of 11.4 per­
cent was recorded. The decline in 1980 was the largest
single-year decline noted.
Productivity should benefit from continued improve­
ments in the production process and in the equipment
used. Increased utilization of computer technology may
also contribute to productivity gains. Substantial de­
mand for the industry’s products is expected to continue.

to give the workpieces their final shape. However, ad­
vances in the feeding mechanisms permit preforming of
the pieces to such an extent that they can enter the dies
without the usual need for heating.
Farm and garden machinery (SIC 352). Productivity
in farm and garden machinery manufacturing grew at
a rate of 2.6 percent from 1958 to 1980, with output
increasing at 4.2 percent per year and hours growing
1.5 percent per year. The rate of productivity gain was
about the same as the average for all manufacturing in­
dustries over the period measured.
Productivity growth in the farm machinery industry
can be divided into three distinct periods. During 195865, productivity grew at a rate of 1.7 percent; in 196574, it accelerated to a rate of 3.3 percent per year; and
in 1974-80, it slowed to 0.2 percent. The higher rate of
gain during the 1965-74 period can be associated with
a number of years of very high output and productiv­
ity growth, fueled by dramatic increases in farm income.
The introduction of new technology including nu­
merical control of machine tools, computer-controlled
warehousing facilities, and industrial robots, which have
been adopted mainly by the larger firms in the indus­
try, has aided productivity growth. On the other hand,
productivity growth has been moderated by the vari­
able nature of demand, which is affected by cyclical
changes and changes in farm income. For example, al­
most every decline in productivity can be associated
with a drop in output. In turn, industry output declines
tend to be roughly coincident with downturns in the
economy.
Pumps and compressors (SIC 3561,3563). Output per
employee hour in pump and compressor manufacturing
rose at an average annual rate of 2.1 percent between
1958 and 1980—compared with a rate of 2.8 percent
for manufacturing as a whole. Output increased 4.7 per­
cent a year, employee hours 2.6 percent.
Pumps and compressors are used throughout manu­
facturing and in many nonmanufacturing industries, as
well as agriculture. Pumps are the second most com­
mon machine in use after the electric motor. Compres­
sors generate compressed air, which may be regarded
as a form of energy ranking in breadth of use only be­
low electricity, gas, and water. Growth in the output
of pumps and compressors was, in general, related to
expansion in industrial and public utility demand, gains
in residential and associated public works construction,
and intensified needs of energy-related extractive and
pipeline industries. Foreign trade, too, played an im­
portant role in sustaining output: About one fifth of
pump and compressor production was exported be­
tween 1972 and 1978.
Small lot production is the rule in pump and com­
pressor establishments. Pumps and compressors are

Hand and edge tools (SIC 3423). As measured by out­
put per employee hour, productivity in the hand and
edge tool industry grew at an average annual rate of
only 1.3 percent during 1958-80, compared with 2.8
percent for all manufacturing. Output increased at a
rate of 3.3 percent and employee hours at a 2.0 percent
rate.
Productivity gains have been linked to the continu­
ing mechanization of the production process. The use
of robots in material handling operations has been a
factor in this regard. Improvements in the ovens used
to heat the metal for the forging operation have con­
tributed to faster production rates. The adoption of
horizontal impact forging equipment by some plants has
resulted in improved production efficiencies. The adop­
tion of cold forming techniques is also aiding produc­
tivity. In the cold forming process, dies are still used



46

often large machines, manufactured to customer speci­
fications. While many of these machines are composed
of standard parts, the economies associated with mass
production are generally not available. The production
process must constantly be adapted so as to cope with
the many design, casting, and machining requirements
that arise. Such adaptation was facilitated by the ad­
vent of numerically controlled machine tools in the six­
ties, and the introduction of computer-aided design
(CAD) into engineering practice. Numerical controls
(NC) and CAD have been important sources of pro­
ductivity advances in the industry.

banking services, somewhat retarded productivity im­
provement, partly because scale economies became less
favorable.
The sources of the industry’s strong output growth
were the boom conditions of the early seventies and
the financial needs they generated; rapid increases in
check transactions; relatively greater reliance by busi­
ness on external funds; and continuously heavy demand
for consumer and real estate credit. Also, commercial
banks expanded their share of major types of such credit,
as well as of time deposits. Moreover, they emphasized
the retailing aspects of their services and consequently
accelerated branching. Trust department functions also
grew apace as pension and welfare funds proliferated.
The banking industry has been radically transformed
by electronic data processing. Although computer de­
velopments during the fifties embodied the principle of
machine readability, it was the introduction of magnetic
ink character recognition (MICR) in 1958 that made
the breakthrough of electronic data processing in bank­
ing possible. The computer became an indispensable
and major factor in improving banking productivity,
and the technology has rapidly permeated the industry.

Commercial banking (SIC 602). Output per employee
hour in commercial banking rose at an average annual
rate of 1.3 percent between 1967 and 1980—nearly the
same as for the nonfarm business sector as a whole (1.4
percent). Output over the period examined rose at a rate
of 6.0 percent per year, while employee hours grew 4.6
percent. The rise in banking productivity was associated
with strongly expanding customer services and with ad­
vances in computer technology and their rapid diffusion
throughout the industry. However, the spread of branch
banking, while enhancing the convenience of access to




47

Talbl® 1. S®l®€t®d industries: Emptoymsnt, 1881, and average anrtusS rate® ©if <§ton§® ta@utpuit p@r employ®® to u r, 1878-81
Output per employee hour:
Average annual rate of change,
1976-81 (percent)1

Employment, 1981
(thousands)
SIC cod®

Industry
All
employees

Production
workers

Non­
production
workers

All
employees

Production
workers

Non­
production
workers2

m siiS
S inig
1011
1011
1021
1021
111,121
121
14
142

Iron mining, crude o re...............................
Iron mining, usable o re .............................
Copper mining, crude ore ........................
Copper mining, recoverable m e ta l.........
Coal mining..................................................
Bituminous coal and lignite mining.........
Nonmstallic minerals, except fuels.........
Crushed and broken stone........................

21
21
36
36
222
219
119
37

17
17
23
23
184
181
91
30

4
4
8
8
38
38
28
7

5.3
4.6
0.1
-0.8
3.4
3.4
-0.1
1.8

5.5
4.8
0.1
-0.8
4.0
4.1
0.5
2.0

4.5
3.8
0.1
-0.8
0.1
0.1
-2.4
0.9

E^anufaeturlmg
2026
203
2033
204
2041
2043
2044
2045
2046
2047, 48

Fluid milk.....................................................
Preserved fruits and vegetables...............
Canned fruits and vegetables..................
Grain mill products.....................................
Flour and other grain mill products.........
Cereal breakfast foods.............................
Rice milling.................................................
Blended and prepared flour......................
Wet corn milling ........................................
Prepared feeds for animals and fowls_
_

96
240
91
140
20
17
5
8
13
71

(3)
197
75
97
10
13
4
3
10
48

(3)
43
16
43
10
4
1
2
3
23

6.2
40.1
4-0.2
43.6
2.8
42.0
44.3
4-2.1
410.3
43.5

(3)
40.6
40.6
43.5
1.9
41.2
40.5
4-3.2
411.4
44.0

(3)
4-2.2
4-5.5
43.8
5.8
46.7
4-4.1
41.1
47.5
42.8

205
2061,62,63
2061,62
2063
2055
2082
2086
2111,21,31
2111,31

219
32
20
12
56
50
137
58

127
23
14
9
45
32
45
45

©2
©
6
3
11
1©
92
13

-0.2
2.1
43.8
43.0
(3)
4.2
3.5
0.7

-0.5
1.5
42.9
44.0
(3)
4.2
5.5
1.5

0.4
5.0
47.1
4-3 .9
(3)
4.1
2.4
-3.0

2121

Bakery products........................................
S ugar...........................................................
Raw and refined cane sug ar....................
Beet s u g a r.................................................
Candy and confectionery products.........
Malt beverages..........................................
Bottled and canned soft drinks.................
All tobacco products.................................
Cigarettes, chewing and smoking
tobacco ...................................................
Cigars...........................................................

51
7

39
6

12
1

0.5
3.0

1.3
2.8

-3 .8
5.5

2251,52
2281
2421
2431
2435,36
2435
2436
251
2511,17
2512

Hosiery.......................................................
Nonwool yarn mills....................................
Sawmills and planing mills, g en eral.......
Millwork.......................................................
Veneer and plywood.................................
Hardwood veneer and plywood...............
Softwood veneer and plywood.................
Household furniture...................................
Wood household furniture........................
Upholstered household furniture...........

65
82
172
67
64
25
39
299
140
89

58
75
152
54
57
22
35
250
122
73

7
7
20
13
7
3
4
49
18
18

1.5
1.0
-0.3
4-1.7
4-0.3
42.6
4-1.5
40.1
4-0.7
41.2

2.0
1.3
0.5
4-1.0
40.4
42.8
4-0.5
40.5
4-0.2
41.3

-2.3
-3.1
-5.0
4-5.0
4-5.6
40.5
4-8.4
4-1.9
4-4.1
40.8

2514
2515
252
2521
2522
2611,21,31,61
2843
2651
2653
2823,24

Metal household furniture........................
Mattresses and bedsprings......................
Office furniture..........................................
Wood office furniture.................................
Metal office furniture.................................
Paper, paperboard, and pulp mills...........
Paper and plastic b a g s .............................
Folding paperboard boxes........................
Corrugated and solid fiber boxes.............
Synthetic fibers..........................................

31
31
54
22
32
207
50
43
103
102

25
24
42
19
24
202
39
34
76
74

6
7
12
3
8
65
11

4-1.7
42.7
44.7
47.0
43.2
2.3
4-1.0
-1 .8
2.7
0.2

4-2 .0
43.5
45.3
42.9
2.5
4-1 .4
-1 .5
3.1
6.1

4-0 .8
40.3
42.7
4-2.8
44.3
4-2.2
-2.2
1.6
8.6

2834
2841
2844
2851
2911
3011
314
3221
3241
325

Pharmaceutical preparations..................
Soaps and detergents...............................
Cosmetics and other toiletries.................
Paints and allied products........................
Petroleum refining.....................................
Tires and inner tubes.................................
Footwear.....................................................
Glass containers........................................
Hydraulic cem ent.......................................
Structural clay products............................

158
44
62
63
174
107
143
@
8
30
41

77
29
41
31
105
75
123
5©
24
31

81
15
21
32
32
20
©
6
10

42.6
42.0
4-3.4
0.9
4-0.3
3.1
-0.0
3.3
-2.5
-0.8

42.1
41.8
4-3.8
1.8
41.3
3.2
-0.4
3.5
-2.4
(5)

42.8
4-2.6
(5)
4-3.8
2.7
-2.1
1.7
-3.2
-3.2

3251,53,59
3251
3253

Clay construction products......................
Brick and structural clay t i l e ....................
Ceramic wall and floor til© ........................

30
17
©

23
13

7
4
2

-2.2
-5.7
47.4

-1.4




48

7

9
27

28

m

-4.0
40.0

1.4

(3)

-5.6
-12.4
411.4

Table 1. 8©!©@t®dl Industries: Employment, 1i§H, aradl a^©rag© annual rat©Q ©! ©hang® B output par ®mpl®y©® hour,
n
D® —
F@ ©H= G®nfiBnu©d
Output per employee hour:
Average annual rate of change,
1976-81 (percent)1

Employment, 1981
(thousands)
Industry

SIC code

All
employees

Production
workers

Non­
production
workers

All
employees

Production
workers

Non­
production
workers1
2

^anutoeturSng^CemftBmuQd]
3255
3271,72
3273
331
3321
3324, 25
3331,32,33

Clay refractories......................................
Concrete products...................................
Ready-mixed concrete............................
Steel.........................................................
Gray Iron foundries.................................
Steel foundries........................................
Primary copper, lead, and zinc................

11
85
88
505
121
64
22

8
82
(3)
391
08
49
17

3
23
(3)
114
23
15
5

4.3
4-o.e
4- 1 .2
2.0
-1.0
-1.8
3.6

4.6
4-0.1
<
3)
2.3
<
5)
-1.6
3.9

3.4
4-3.2
<
3)
1.1
-5.8
-2.3
2.4

3331
3334
3351
3353,54, 55
3411
3423
3441
352
3523
3524

Primary copper........................................
Primary aluminum.....................................
Copper rolling and drawing.....................
Aluminum rolling and drawing ................
Metal cans.................................................
Hand and edge tools...............................
Fabricated structural m etal.....................
Farm and garden machinery...................
Farm machinery......................................
Lawn and garden equipment...................

14
37
29
72
58
46
101
156
135
21

11
28
22
52
50
35
72
105
00
15

3
9
7
20
8
11
29
51
45
6

5.1
-0.1
1.8
-0.6
2.6
40.7
1.3
4-0.7
4-1.6
44.0

5.4
0.6
2.3
-0.1
2.8
41.4
0.9
4-1.1
45.1

4.2
-2.6
(5)
-2.5
1.7
4-1.8
2.5
4-2.3
4-3.1
41.2

3531
3541,42
3541
3542
3531,53
3561
35S3
3532
3612
3621

Construction machinery and equipment.
Machine tools ..........................................
Metal cutting machine tools.....................
Metal forming machine tools...................
Pumps and compressors........................
Pumps and pumping equipment..............
Air and gas compressors........................
Ball and roller bearings............................
Transformers............................................
Motors and generators............................

143
104
80
24
04
32
32
57
54
125

95
87
52
15
58
37
10
44
38
98

48
37
28
9
38
25
13
13
16
20

-1.0
0.6
1.4
-2.0
40.9
41.0
40.5
-1.5
45.1
-0.4

-0.1
0.7
1.1
-0.7
40.9
41.0
40.5
-1.8
44.5
0.3

-3.1
0.3
2.0
-4.7
40.9
41.0
40.4
-0.3
48.4
-2.8

3631,32,33, 30
3031
3632
3633
363®
3641
3045,46,47,48
3651
371

Major household appliances...................
Household cooking equipment..............
Household refrigerators and freezers ...
Household laundry equipment................
Household appliances, n.e.c...................
Electric lamps..........................................
Lighting fixtures......................................
Radio and television receiving sets.........
Motor vehicles and equipment................

93
23
34
21
15
33
65
82
784

74
18
27
17
12
29
49
57
583

19
5
7
4
3
4
16
25
201

2.4
2.8
4.3
0.2
1.3
1.1
4-0.7
44.4
-0.3

2.1
1.9
3.5
0.1
1.7
1.1
(4
.)
45.1
1.1

3.5
5.3
8.0
0.6
-0.4
1.1
4-2.9
42.2
-4.4

1.4
40.6
(3)
(3)
<
3)
(3)
(3)
(3)
(3)
(3)

(4

S
)

Other
401 class I
401 class I
4111,413, 414 pts
4213 part
4213 part
4511,4521 part
4612,13
4811
481,92,93
401,493 part

Railroad transportation, revenue traffic..
Railroad transportation, car miles..........
Bus carriers, class I .................................
Intercity trucking......................................
intercity trucking (general freight)...........
Air transportation.....................................
Petroleum pipelines.................................
Telephone communications...................
Gas and electric utilities..........................
Electric utilities........................................

457
457
36
670
413
358
22
1,076
777
501

395
395
(3)
(3)
<
3)
<
3)
15
r782
7632
(3)

62
02
(3)
(3)
(3)
<
3)
7
8204
8145
(3)

3.1
41.5
4 a2.4
-0.8
-1.7
®
2.1
-2.2
5.8
-0.8
-0.0

3.4
*1.7
(3)
(3)
(3)
(3)
-1.0
(3)
7-0.5
(3)

492,493 part
54
5511
5541
58
5912
602
7011
721

Gas utilities...............................................
Retail food stores0 ...................................
Franchised new car dealers.....................
Gasoline service stations0 .......................
Eating and drinking places0.....................
Drug and proprietary stores0...................
Commercial banking...............................
Hotels, motels, and tourist courts9 .........
Laundry and cleaning services9 ..............

210
2,420
708
691
5,097
528
1,482
1,179
419

(3)
(3)
(3)
<
3)
(3)
(3)
(3)
(3)
(3)

(3)
(3)
(3)
(3)
(3)
(3)
(3)
(3)
(3)

-0.5
-0.1
0.2
2.1
-1.8

(3)
(3)
(3)
(3)
<
3)
(3)
(3)
(3)
(3)

1 Based on the linear least squares trends of the logarithms of the in­

dex numbers.
2 Rates of change for nonproduction worker hours are subject to a
wider margin of error than other rates shown.
3 Not available.

4 1976-80.




1.2

4-0.6
-0.6
-3.2

(3)
(3)
(3)
(3)
(3)
(3)
(3)
(3)
(3)

5 Less than 0.05 percent.
8 Output per employee.
7 Nonsupervlsory personnel.
8 Supervisory personnel and force account construction workers.
0 Data relate to all persons, including proprietors, partners, and un­
paid family workers.

49

Measuring productivity
in service industries
The growth o f the service economy
presents special challenges
fo r productivity analysts; output
is often difficult to quantify, and measurement
o f labor input requires great care
Je r o m e A . M

ark

Linking output to input

The increased importance of service industries over the
last two decades and current concern over productivity
growth have stimulated interest in productivity mea­
sures for this expanding sector of the economy.
The service sector, as defined here, encompasses the
major industry groupings of trade, finance, insurance,
communications, public utilities, transportation, and
government, as well as business and personal services.
It accounts for almost three-fourths of the Nation’s em­
ployment and provides the greatest potential, as well as
some of the greatest difficulties, for developing produc­
tivity measures.
Over the last decade, the Bureau of Labor Statistics
has been expanding the number of service industries for
which it publishes productivity measures, and at present
provides measures for 16 industries, representing almost
a third of the employment in the sector. The Bureau is
continuing to develop additional measures, and hopes
eventually to extend coverage to most of the service sec­
tor.
This article describes that effort, discusses some of
the problems of measuring productivity, particularly la­
bor productivity in service industries, and explains how
the Bureau is working to resolve some of the problems.

Productivity measures relate real physical output to
real input. They range from single factor measures, such
as output per unit of labor input or output per unit of
capital input, to measures of output per unit of multi­
factor input. Such measures also reflect changes in tech­
nology, scale of production, educational levels of
workers, managerial techniques, and many other factors
in addition to the contributions of the particular inputs.
Although b l s is currently developing multifactor pro­
ductivity measures, at present, the published productivi­
ty measures relate output to labor input. This is the
most extensively developed and widely used productivi­
ty measure because of its relevence to economic analy­
ses and because, as a practical matter, labor is the most
easily measured input.
Problem s of m easuring output
In many ways, the problems of measuring output in
the service industries are similar to those of measuring
output in the goods-producing industries. That is, the
output indicator must be quantifiable and independent
of the input measures. If an output measure for an ac­
tivity is based on an input measure, as is the case in
some instances in the national accounts, obviously no
change in productivity can be ascertained. In the case of
general government, for example, output in the national

Jerome A. Mark is Assistant Commissioner for Productivity and
Technology, Bureau of Labor Statistics.

Reprinted from the
M onthly Labor Review, June 1982.




50

by unit labor requirements, and in sufficient detail to
adjust for quality changes. In practice, however, such
data are not generally available for service industries
(and, in many cases, for goods-producing industries as
well). As a result, approximations based on alternative
approaches must be used.
The principal alternative is to remove the change in
price from the change in total value of the volume of
services. This approach is tantamount to price
weighting quantities of services provided. Insofar as
price relationships among the various component serv­
ices of a service industry are similar to the unit labor
requirements or unit labor costs, this is a close approxi­
mation of the desired measure. And because it is easier
to measure price change for a specified group of services
than it is to measure the number of services provided
directly, this is the approach most generally followed.
However, the adjustment requires data in sufficient
detail to adequately represent the price trends of the
components included in the price change. Otherwise,
price movements of the covered areas will be implicitly
imputed to the uncovered areas. But because the rela­
tionship among the price movements of similar services
is much stronger than the relationship among quantity
changes, this alternative still has greater viability than
imputing quantity changes for uncovered services.
In practice, BLS uses the two approaches to develop
output measures for service industries. In some in­
stances, quantity data are available, particularly for util­
ities and transportation industries. In others, price de­
flation is employed, and for some, deflation at lower
levels of aggregation is combined with labor input
weighting at higher levels. For example, in developing
the measure for gasoline service stations, gasoline sales,
repair, and other services are deflated separately and
summed, but in the case of retail food stores, sales by
major department are deflated and combined with em­
ployee labor cost weights.

income and product accounts is measured in terms of
compensation of government employees. The deflated or
constant-dollar measure is derived from changes in em­
ployment. Hence, changes in the output measure are
closely related to changes in the input measure.
It is also important to distinguish between intermedi­
ate and final services. In productivity measurement, we
attempt to ensure that the indicators represent output
flowing from the industry being measured rather than
intermediate steps in the service flow. In this sense, pro­
ductivity measurement differs from work measurement,
which generally refers to the analysis of the operation of
an activity and the labor requirements at each interme­
diate stage. Productivity measurement refers only to the
final service and its relationship to input.
For example, in the trucking industry,-a count of the
ton-miles of freight moved would be the appropriate in­
dicator of the final output— that is, the result of all the
activities of the industry. The intermediate steps, such
as pickup and delivery, platform work, billing, and col­
lecting, are considered to be subsumed in the final out­
put.
In the case of an organization or an industry
providing one type of service, output is merely a count
of the units of this service, however defined. In the
more usual case of an industry producing a number of
heterogeneous services, the various units must be
expressed in some common basis for aggregation. For
example, the output of franchised new-car dealerships
should be a combination of the number of cars sold and
the repair activities of the dealers, with appropriate
weighting.
To obtain a productivity measure that is an average
of the changes of individual components, the appropri­
ate weights for combining the various elements in the
output measure are in terms of their factor input re­
quirements. In a labor productivity measure, the
weights are unit labor requirements.
Homogeneity among services, after considerations of
quality and specifications, is indicated by similarity in
unit labor requirements. In this way, the output mea­
sure for the development of labor productivity statistics
differs from more traditional production measures based
on total price or value-added price weighting.
When there are quality changes within the service,
adjustments must be made in the output measure to ac­
count for the fact that the output is no longer the same
homogeneous unit. However, the indicator of quality
change for labor productivity measurement dif­
fers from the usual concept of quality change associated
with consumer price measurements in that it reflects dif­
ferences in producers’ labor requirements or labor costs
rather than consumer utility differences.
Ideally, then, the output measure should incorporate
data on the number of services provided, differentiated



M easuring labor input
With regard to labor input measures, the principal
problems are data gaps. Information is needed on hours
worked by all persons— nonsupervisory workers, super­
visory workers, and self-employed and unpaid family
workers—in an individual industry. But although data
on hours worked are collected by various government
agencies as part of such ongoing programs as the Bu­
reau’s occupational safety and health surveys, they tend
to be limited in scope, or otherwise inconsistent with
the output data developed.
The principal source of data on employment and
hours is the BLS Current Employment Survey of estab­
lishments. This payroll series provides good measures of
the employment and hours of nonsupervisory workers.
However, it is collected on an hours paid basis, rather
SI

than on an hours worked basis. To the extent that
hours paid for but not worked are changing, this mea­
sure has limitations. To overcome this problem, the Bu­
reau is measuring hours at work as a proportion of
hours paid for a sample of establishments in the survey
and will use these data to adjust the industry hours
paid series.
In general, data on the hours of supervisory workers
are poor. Although employment data on supervisory
workers are available from the payroll survey, hours
data are not. Other sources, such as the censuses of
population, are used to estimate this component of the
labor input measure.
Data on the number of self-employed, an important
component of the input series measure for retail indus­
tries, come from the Internal Revenue Service ( ir s ). The
IRS data lag current estimates by 3 years, but may be
projected forward with special tabulations from the
Current Population Survey ( c p s ).
These CPS tabulations break out the numbers and
hours of self-employed and unpaid family workers at
the 3-digit Standard Industrial Classification level. Al­
though the sample size at this level is small and the sta­
tistical error is high, the data are the only continuous
series of the number and hours for unpaid family work­
ers and for the hours of the self-employed.
The measures derived from these data are unweighted
hours; that is, the hours of various types of employees
are treated as being equally productive. This would not
be a problem if the proportions of workers at different
levels of productivity were constant over time. Howev­
er, to the extent that there are changes in the composi­
tion of the work force, such as age, sex, and
occupational mix, it may be desirable to adjust the la­
bor input measure for these changes which otherwise
would be reflected in the productivity measure.
Data gaps hamper the making of these adjustments.
Industry data on employment and hours by age and oc­
cupation are limited, although various sources, such as
the CPS and BLS occupational employment surveys, pro­
vide some pieces. And while worker groups may be dif­
ferentiated into productivity levels according to their
wages or compensation, pay is a factor which may re­
flect other than productivity differences.1

Table 1. Average annual rates of change in output per
hour of all employees in selected service industries,
1965-73 and 1973-80
[In percent]

SIC Code

1965-73

1973-80

4111 ;4131 ;414 (parts) .
4213 (parts)..................
4511 .............................
4612,13........................

Transportation:
Railroad transportation, revenue
traffic ......................................
Bus carriers ...............................
■ Intercity trucking1 ........................
Air transportation1 ......................
Petroleum pipelines ....................

4.2
-1.5
2.7
5.3
7.9

2.2
-0.4
0.5
4.3
0.0

4811 .............................

Communications:
Telephone communications.........

4.7

7.0

491 ;492;493 ..................
491 ;493 (part) .............
492;493 (part) .............

Public utilities:
Gas and electric utilities.............
Electric utilities.............................
Gas utilities.................................

4.9
5.4
3.9

0.7
1.3
-0.4

2.2
2.6
4.9
1.1
62

-1 .2
0.5
3.1
-1 .0
19

1.8
1.7

1.3
-1.1

401 ...............................

5 4 .................................
5511 .............................
5541 .............................
5 8 .................................
5 9 1 2 .............................

7011 .............................
721 ...............................

Trade:
Retail food stores ......................
Franchised new car dealers . . . .
Gasoline service stations ...........
Eating and drinking places .........

Services:
Hotels, motels, and tourist courts
Laundry and cleaning services ..

10utput per employee.

tries in the goods-producing sector.
In addition, a measure for commercial banking is be­
ing developed, and work has begun on measures for the
insurance and hospital industries. In a related area, pro­
ductivity measures for Federal agencies which provide
functions such as recordkeeping, insurance, libraries,
building and grounds maintenance, and medical services
have been published.
It is not possible within the confines of this article to
discuss all of the productivity measures prepared by the
Bureau, but reference to some of the more important
and interesting ones in each of the major areas can il­
lustrate the difficulties encountered in constructing such
statistics.
Trade. The Bureau has published measures for retail
trade industries since 1975 (with the data beginning in
1958). At present, statistics are published for five im­
portant industries— retail food stores, new car dealer­
ships, gasoline service stations, eating and drinking
places, and drugstores. Work is underway on a measure
for apparel stores, including shoe stores, to be published
separately. The effort to develop productivity measures
in the wholesale area has not yet succeeded.
For most retail trade industries, data on gross sales
in current dollars, deflated by the appropriate price in­
dexes, are used to estimate real output. This method, as
mentioned earlier, can yield good estimates of real out­
put. However, such measures can reflect shifts among
services with different values, but having the same labor
requirements. Therefore, the overall industry productivi­

Measures for service industries
At present, BLS publishes indexes of output per unit
of labor input for industries in each major service activ­
ity— trade, communications, transportation, utilities,
and business and personal services, a total of 16 sepa­
rate measures. Data for these industries, presented in ta­
ble 1, indicate a wide range of productivity growth
since 1973, the year in which a productivity slowdown
for the general business economy appeared to begin. In
many cases, the growth rates exceeded those for indus­



Industry

52

with the movement of goods and passengers are usually
greater for short hauls than for long hauls. Therefore, a
shift from a long haul to a short haul trip or vice versa
could be reflected as a change in productivity although
only the mix of trips had changed.
For the two major freight-carrying industries, rail­
roads and trucking, undifferentiated ton-mile informa­
tion is reported for total freight operations. In trucking,
the ton-mile data are also reported separately for three
types of carriers— general, contract, and others. But
output measures should reflect the kinds of commodities
handled and the average distance they are moved. The
preferred way to develop these measures would be to
combine the tonnage and the average haul of each com­
modity by its respective labor requirements and aggre­
gate the results for all commodities transported. Un­
fortunately, this cannot be done with available data.
However, supplementary information on tonnage for
railroads is available from the ICC for about 200 com­
modity lines, ranging from agricultural and mining
products to motor vehicles and scientific instruments.
Until recently, similar information was also available
for the trucking industry. BLS uses these data to adjust
the overall measure of freight ton-miles for changes in
the composition of goods carried.
Although this commodity adjustment is a significant
improvement, refinements to the undifferentiated tonmiles cannot be developed to the extent desired. For ex­
ample, separate labor requirements data are not avail­
able for weighting the individual commodity groups.
The commodity index adjustments are therefore made
in terms of unit revenue weights, the underlying as­
sumption being that differences between labor require­
ments among commodities are similar to differences in
terms of unit revenues. This does not seem unreasonable
because labor costs constitute more than half of each
industry’s total operating costs, although the proportion
could conceivably differ by commodity. For railroads,
the adjusted freight ton-mile measure is combined with
a measure of revenue passenger-miles to obtain the total
industry output index.
For air transportation and trucking, employment is
the only available measure of labor input. Thus, the
productivity measures for these two industries should
be interpreted with caution, for if changes occur in the
average workweek, the trends in productivity would not
show the true relationship between output and labor
time expended on the output.
The transportation industries fo r which BLS publishes
productivity measures all are regulated to some degree
by the Federal Government. Recent efforts to reduce
the paperwork burden, coupled with the effects of de­
regulation, have acted to eliminate some of the operat­
ing statistics previously published. As a result, some

ty index can show movements without any change in
component elements.
In retail industries, a large portion of the value of
sales has been provided by the manufacturer and the
wholesaler of the product sold. A net output measure
would be desirable, because it would most closely corre­
spond to the value added by the retailer. However, a
gross or total sales measure will yield the same results
as a net or value-added measure if the value added as a
percent of sales (gross margin) does not change over
time. Available data indicate that, among retail
industries for which productivity data are published,
gross margins have not changed significantly over time.
To incorporate labor input weights, the indexes for
most of the retail trade industries are developed in two
stages. First, deflated output measures based on sales
volume are developed for detailed merchandise lines.
These are aggregated to higher levels and then com­
bined with labor costs weights. For example, in retail
food stores, sales for 13 key merchandise lines are de­
flated using specially prepared price indexes based on
CPI components. The merchandise lines are aggregated to
five department lines— meat, produce, frozen food, dry
groceries, and dairy and all others. These are then ag­
gregated with labor cost weights from Department of
Agriculture data to develop the overall output measure
for groceries. The labor input data for retail trade pro­
ductivity statistics are generally derived from the Bu­
reau’s establishment survey, supplemented by IRS and
c p s data.
Transportation. BLS publishes productivity measures for
five transportation industries— railroads, intercity truck­
ing, intercity buses, air transportation, and petroleum
pipelines. These measures cover 57 percent of transpor­
tation employment.
Conceptually, productivity measures for the transpor­
tation industries are easier to develop than those for
other non-goods producing industries. This is because
transportation industry output— the movement of goods
or passengers or both from one point to another—
is more easily quantified. Output units in transportation
have two dimensions, amount and distance; they reflect
not only how much has been transported, but also how
far. As such, ton-miles, passenger-miles, barrel-miles,
and so forth are the primary output indicators for these
industries.
Although the basic information for developing good
transportation productivity measures is available and is,
of course, being used, there are some data gaps that
place certain limitations on the BLS measures. For ex­
ample, it is sometimes impossible to adjust the produc­
tivity measures adequately for changes in the average
length of haul. The unit labor requirements associated




53

productivity measures have had to be extended on the
basis of more limited information. The outlook for
expanding the data base, at least in the near future, is
not favorable. However, BLS is cooperating with other
government agencies to ensure that adequate statistics
for transportation industries remain available.
Communications. The BLS productivity measure for tele­
phone communications covers about four-fifths of the
employment in the communications sector. The output
index is derived from revenues of all telephone compa­
nies reporting to the Federal Communications Commis­
sion. The revenues are stratified by major source—
local, toll, or miscellaneous— and deflated by specially
prepared price indexes for these different services. The
labor hours data are based on the Bureau’s estab­
lishment payroll survey.
At one time, BLS published a productivity measure,
the numerator of which was derived from the number
of local and long-distance telephone calls, aggregated
on the basis of revenue weights. This measure was dis­
continued in the mid-195Q’s because of concern that the
labor input measure was not consistent with the output
measure. For example, private line services, such as
leased telephone lines, radio and TV transmission, tele­
type, and so forth, were reflected in employee hours but
not in the output measure as defined. The same was
true for calls between stations transmitted through pri­
vate switchboards and directory services.
A different type of productivity index for the industry
was initiated in 1973, with data back to 1951. The nu­
merator of this measure was derived from annual reve­
nue data stratified by major services and deflated, until
last year, by price indexes furnished by American Tele­
phone and Telegraph Co. Beginning in 1982, the BLS
producer price index for telephone communications will
be used to deflate the revenue data, and productivity in­
dexes published for the industry since 1972 will be re­
vised in accordance with the new procedure.
The BLS deflated revenue measure of the output of the
telephone communications industry is fairly comprehen­
sive. It includes revenues from private Sine services,
which have grown in importance over the years, as well
as those arising from the maintenance of private switch­
boards by telephone carriers. St also accounts for TV,
radio, and computer data transmission by telephone in­
dustry facilities, and for directory services. However,
certain measurement problems remain unresolved, in­
cluding the unsatisfactory treatment of differences in in­
tensity of the use of telephone equipment by customers.
Intensity of use differences occur when revenue does not
vary in proportion to the number of calls made because
of flat charges, as in the case of local telephone service
or WATS Sines. Implicitly, the BLS output measure as­



54

sumes that the maximum permissible usage takes place
under any flat charge system used in the industry.
Business and personal services. In the area of business
and personal services, which includes not only business,
personal, and repair services, but also education, social
services, and political organizations, BLS currently
publishes only two measures of productivity, one for
hotels and motels, and the other for laundry and dry
cleaning services. These measures cover 13 percent of
the total employment in the sector.
Because physical quantity information as not available
for these two industries, output measures are developed
using price-deflated value techniques. The techniques
are similar to those described earlier, in that both reve­
nues and employee-hour weights are used to aggregate
the output indicators into a total industry output index.
On the input side, the hours of all persons are used
as the measure of labor time. As in the trade sector,
partners, proprietors, and unpaid family workers make
up a significant portion of the work force. Currently,
this group accounts for about 15 percent of all persons
employed in laundries and 20 percent of the workers in
hotels and motels.
BLS efforts to expand coverage in the business and
personal service area have been hampered by two major
problems. First, because many business service catego­
ries are quite broad, it is impossible to account ade­
quately for changes in the mix of their component
services. For example, we cannot publish a productivity
index for automotive repair shops because there are al­
most no data available on the types of repairs that are
made. The second problem is that not enough services
are covered by the Consumer Price Index and, conse­
quently, the deflated value of the output of many un­
covered areas would have to be imputed.
Finance. In the finance area, BLS is developing a bank­
ing measure in terms of the three major services com­
mercial banks render their customers—deposits, loans,
and trust services. While banks also provide non-fundusing services, such as safe deposit and customer pay­
roll accounting, lack of adequate data preclude deriving
a measure for theim, However, because the proportion
of employees engaged in such services is very small, the
overall output measure is little affected by the omission.
There has been much controversy over the years as to
the appropriate measure of the output of banking. Some
analysts have advocated a ‘ liquidity5’ approach, others,
“
’
a “transactions” approach. In the former, the banks are
viewed as holders of money, and their output is
equivalent to the net interest they receive on the volume
of deposits held. This interest is the income depositors
are willing to forgo to maintain deposits rather than in­

ment. Included in the loan output measure are commer­
cial and residential mortgage loans; consumer loans;
single-payment loans; credit card loans; and commercial
and “other” loans. The number of loans can usually be
derived by dividing the dollar value of total loans in a
given category by the average face value of a loan. For
the category of commercial loans, the actual number of
loans extended has been available since the mid-1970’s.
An experimental output measure for the trust depart­
ment services of commercial banks is derived from the
trend in the number of accounts. Trust accounts are
stratified into five major categories, including benefit
trusts, personal trusts, and estates.
After output estimates are developed for depository,
loan, and fiduciary segments, they are aggregated to the
industry level using employment weights.

vesting directly in assets less readily converted to cash,
that is, the value to customers of the liquidity they en­
joy from bank services. This approach can be extended
to all types of savings accounts, on the principle that
the forgone net interest is the value of the bank’s ser­
vices.
The other approach views banking output as a series
of transactions; the volume of the bank’s output is pro­
portional to the volumes of the transactions handled.
BLS has adopted this second approach for its produc­
tivity measure.
Accordingly, the final output of banks is defined as
an array of depository, lending, and fiduciary services.
Estimates of the number of transactions for each of the
three service functions must be derived. Because no di­
rect count of the number of transactions is available in
many instances, estimates are made from data on the
total value of transactions and surveys of average trans­
action amounts.
Deposit activity is measured in terms of the number
of checks transacted and the number of time and sav­
ings deposits and withdrawals. (An electronic funds
transfer is treated as a transaction on par with one in­
volving payment by check.) The data for demand de­
posit activities are from Federal Reserve counts and of­
ficial benchmark surveys. For time and savings deposit
activity, the output measure is based on data published
by the Federal Deposit Insurance Corporation and on
the Functional Cost Analysis conducted annually by
the Federal Reserve.
Lending services provided by banks are also mea­
sured in terms of units. As in the case of deposit and
trust activity, BLS does not use banks’ financial data to
arrive at the component output measures. Use of such
data would be highly misleading even if appropriate de­
flators could be found. For example, an increase in the
aggregate deflated value of loans might simply reflect
the making of a few large loans; similarly, a decrease
might indicate the repayment of a few large loans, even
as the number of small loans increased.
Twelve types of loan output are measured, for the
most part using data generated by the Federal Reserve
and the Department of Housing and Urban Develop­




S o m e o f t h e MAJOR p r o b l e m s in developing labor
productivity measures in the service activities and how
BLS has tried to meet some of these problems have been
highlighted above. Considerable work in this very im­
portant area has been conducted and the outlook for
improvements in certain subareas is optimistic. For ex­
ample, as price measures are improved and hours
worked data become available, and as work in the area
of government productivity measurement progresses,
BLS will be able to provide a better picture of what is
happening to productivity in more activities within the
sector. Additional measures in communications, finance,
insurance, and real estate, and business and personal
services can and will be developed, and indexes for
wholesale trade are very possible. However, there are
severe conceptual as well as data problems in measuring
productivity in such industries as education and social
services and in the important field of medical services,
and progress in these areas is expected to be much
slower.

' In connection with work on multifactor productivity measure­
ment, BLS is exploring the possibility of making adjustments for
changes in work force composition.

55

Productivity Clhiariig® im th@
Bituminous 0®al Industry,

R o s e M. Z e i s e l

Productivity in bituminous coal mining reached a
peak in 1968-69 and declined every year thereafter ex­
cept 1973 and 1978. From 1970 to 1979, output per
employee hour declined 4.1 percent (annual average)
and output per production worker hour fell 3.7 percent,
as hours rose sharply and output only moderately. In
contrast, productivity grew rapidly in the previous two
decades. The high productivity growth rates of the
1950’s reflected very sharp declines in hours, while in
the 196Q’s, average hours dropped only moderately but
output rose substantially.
Several factors contributed to rapid productivity
growth in those 20 years. Of major importance was the
diffusion of technological advances in deep mining, par­
ticularly continous mining, and proportionately greater
output from more productive surface mines. Moreover,
in that period, when coal’s competitive position was
poor, only the more efficient mines were still operating,
raising the industry’s level of productivity. Also, a high
degree of labor-management cooperation resulted in a
20-year period free of contract strikes.
The increase in surface mining contributed greatly to
productivity growth. Output per miner day in surface
mines in the 1960’s averaged 31 tons, compared with 14
in underground mining. And coal mined on the surface
increased as a share of total output from 24 percent in
1950 to 38 percent in 1969. But in the 1970’s, the pro­
ductivity decline reflected declines in both surface min­
ing and underground mining. Output per miner day in
surface mines declined from 36 tons in 1969 to 25 tons in
1979; underground productivity fell from a peak of
almost 16 tons in 1969 to 8 tons in 1979. Although many
different explanations have been advanced to explain
the decline, research by the Department of Energy and
the Department of Labor indicates that environmental
and safety regulations, periodic disruptions in produc­
tion, and an increase in the number of less efficient
mines and less experienced workers could be some of the
factors.
Following enactment of the 1969 Coal Mine Health
and Safety Act, additional employee hours in

underground mines per unit of output were required to
comply with the new regulations. Mining methods had
to be changed, and the production process using con­
tinous miners was particularly affected. The diversion
of some labor and capital to activities which did not in­
crease output had an adverse effect on productivity.
Although the impact of the legislation was less by the
mid-1970’s, greater enforcement probably continued to
depress productivity growth. Similarly, the implementa­
tion of State reclamation laws could be one of the major
causes for the productivity decline in surface mining.
The impact of the Federal Surface Mine Control and
Reclamation Act of 1977 on the productivity of surface
mines will not be known until the act is more fully im­
plemented.
Labor disruptions in the 1970’s included three major
contract strikes and many wildcat strikes. The practice
of stockpiling before a strike, and catching up after­
ward, reduces efficiency.
In addition, smaller, less efficient mines entered the
industry as coal prices rose in response to the oil em­
bargo and to changing market conditions. From 1973 to
1976, the number of active deep mines increased by 40
percent. Smaller mines tend to have lower productivity,
and their entry contributed to the decline in the
mid-1970’s. Also, a younger, inexperienced work force
probably depressed productivity growth.
The decline in bituminous mining productivity con­
trasted sharply with trends in other sectors of the
economy. From 1970 to 1979, output per employee hour
in bituminous coal declined at an average annual rate of
4.1 percent, while in manufacturing it rose at a rate of
2.2 percent, and in nonfarm business, 1.3 percent. In
contrast, from 1950 to 1970, productivity growth in coal
mining had far exceeded growth in other sectors.
The decline also contrasted sharply with the trends in
the coal industries of Europe and the U.S.S.R.,
although data are not fully comparable. From 1960 to
1978, while productivity in the U.S. coal industry
declined by 23 percent, it increased in Poland and West
Germany by 150 and 124 percent, respectively; in

Reprinted from BLS Bulletin 2072 (1981), Technology, Productivity,
and L abor in the Bituminous Coal Industry, 1950-79.




56

Belgium by 76 percent; and the United Kingdom by 48
percent. Nevertheless, the level of productivity in the
U.S. industry in 1978 was still more than twice as high
as in West Germany and Poland.
The productivity outlook for the U.S. industry is mix­
ed. As the shift to capital-intensive western surface
mines accelerates, productivity will be favorably af­
fected. Western surface mines with thicker seams and

less overburden are more productive than eastern sur­
face mines. Also, with the shift West, the proportion of
miners who are union-affiliated declines, which is likely
to reduce the impact of work stoppages. On the other
hand, State and Federal safety and environmental pro­
tection laws could have an adverse effect on surface
mine productivity as they are more fully enforced.

T a b le 1 6 . O u tp u t per m in e r d ay b y m e th o d o f m in in g , 1 9 5 0 -7 9

T a b le 18. O u tp u t per e m p lo y e e h o u r in coal m in in g and selected sectors o f the
e c o n o m y , 1 9 5 0 -7 9

(S h o r t to n s )
Year

U n d e rg ro u n d
m in in g

T o ta l

( In d e x , 1 9 6 7 = 1 0 0 )

S u rfa c e
m in in g

C oal
m in in g 1

Year
6 .7 7
7 .0 4
7 .4 7
8 .1 7
9 .4 7

5 .7 5
6 .0 8
6 .3 7
7.01
7 .9 9

1 5 .6 6
1 6 .0 2
16.81
1 7 .7 3
1 9 .8 0

...............................................................................
..............................................................................
..............................................................................
..............................................................................
..............................................................................

9 .8 4
1 0 .2 8
1 0 .5 9
1 1.3 3
1 2 .2 2

8 .2 8
8 .6 2
8.9 1
9 .3 8
1 0 .0 8

2 1 .1 7
2 1 .3 7
2 1 .8 7
2 1 .8 4
2 2 .9 4

1 9 6 0 ..............................................................................
1 9 6 1 ..............................................................................
1 9 6 2 ...............................................................................
1 9 6 3 ...............................................................................
1 9 6 4 ..............................................................................

1 2 .8 3
1 3 .8 7
1 4 .7 2
1 5 .8 3
1 6 .8 4

1 0 .6 4
11.41
1 1.97
1 2 .7 8
1 3 .7 4

2 3 .3 1
2 5 .2 9
2 7 .3 1
2 9 .3 0
3 0 .0 5

1965
1966
1967
1968
1969

...............................................................................
...............................................................................
...............................................................................
...............................................................................
...............................................................................

1 7 .5 2
1 8 .5 2
1 9 .1 7
1 9 .3 7
1 9 .9 0

1 4 .0 0
1 4 .6 4
1 5 .0 7
1 5 .4 0
15.61

3 2 .7 6
3 4 .2 3
3 5 .8 7
3 4 .6 4
3 6 .0 0

1 9 7 0 ...............................................................................
1 9 7 1 ..............................................................................
1 9 7 2 ...............................................................................
1 9 7 3 ...............................................................................
1 9 7 4 ...............................................................................

1 8 .8 4
1 8 .0 2
1 7 .7 4
1 7 .5 8
1 7 .5 8

1 3 .7 6
1 2 .0 3
11.91
1 1.66
1 1.31

3 5 .8 3
3 5 .8 8
3 6 .3 3
3 6 .6 7
3 3 .1 6

1975
1976
1977
1978

1 4 .7 4
1 4 .4 6
1 4 .8 4
1 4 .2 6
1 3 .5 0

9 .5 4
9 .1 0
8 .6 9
8 .2 5
7 .9 0

2 6 .6 9
2 6 .4 0

1 9 5 0 ..............................................................................
1 9 5 1 ...............................................................................
1 9 5 2 ...............................................................................
1 9 5 3 ...............................................................................
1 9 5 4 ..............................................................................
1955
1956
1957
1958
1959

19 7 9 ”

...............................................................................
...............................................................................
...............................................................................
...............................................................................
..............................................................................................

N o n fa r m
business
s e c to r2

M a n u fa c tu r in g
s e c to r 2

1950
1951
1952
1953
1954

6 7 .2
6 8 .4
6 9 .7
7 0 .7
7 1 .8

6 5 .8
6 7 .9
6 9 .2
7 0 .4
7 1 .4

...............................................
...............................................
...............................................
...............................................
...............................................

5 2 .3
54.1
5 5 .0
5 9 .9
6 0 .9

7 0 .6
7 1 .5
7 3 .5
7 5 .4
7 7 .8

7 4 .6
7 5 .0
7 6 .5
7 7 .9
8 0 .5

7 5 .0
7 4 .4
7 6 .0
7 5 .6
7 9 .0

...............................................
...............................................
...............................................
...............................................
...............................................

6 5 .5
73.1
7 8 .6
8 3 .3
8 8 .8

7 9 .0
81 .4
8 5 .1
8 8 .3
91 .7

81 .2
8 3 .3
8 6 .9
8 9 .8
9 2 .9

7 9 .8
8 1 .6
8 5 .2
9 1 .1
9 5 .6

...............................................
...............................................
...............................................
...............................................
...............................................

9 4 .0
9 8 .2
1 0 0 .0
1 0 2 .2
1 0 2 .2

95 .1
9 8 .0
1 0 0 .0
1 0 3 .3
1 0 3 .6

9 6 .0
9 8 .4
1 0 0 .0
1 0 3 .2
1 0 3 .0

9 8 .4
9 9 .8
1 0 0 .0
1 0 3 .7
1 0 5 .0

1970
1971
1972
1973
1974

...............................................
...............................................
...............................................
...............................................
...............................................

9 8 .9
8 8 .7
8 2 .4
8 3 .7
8 0 .9

1 0 4 .4
1 0 7 .8
1 1 1.5
1 1 3 .6
1 10 .2

1 0 3 .2
1 0 6 .4
1 10.1
1 1 2 .0
1 0 8 .6

1 0 5 .0
1 1 0 .5
1 1 5 .7
1 1 8 .9
1 1 3 .0

1 9 7 5 ...............................................
1 9 7 6 ...............................................
1 9 7 7 ...............................................
1 9 7 8 ...............................................
1 9 7 9 P ........................................

N O T E : O u t p u t p e r m in e r d a y re p re s e n ts to ta l o u t p u t f o r th e ye a r d iv id e d b y th e to ta l
n u m b e r o f m in e r d a y s w o r k e d .

61 .2
6 3 .0
6 4 .8
6 6 .8
6 7 .9

1965
1966
1967
1968
1969

2 6 .5 9

3 7 .7
3 7 .4
3 9 .2
4 2 .2
4 8 .4

1960
1961
1962
1963
1964

2 5 .0 0
2 4 .8 0

...............................................
. . . .............................
...............................................
...............................................
...............................................

1955
1956
1957
1958
1959

p = p r e lim in a r y .

SOURCE:

P riv a te
b u siness
s e c to r2

6 9 .9
6 8 .5
6 7 .2
71 .5
6 5 .9

1 1 2 .6
1 16 .6
1 18 .7
1 1 9 .3
1 1 8 .3

1 10.7
1 1 4 .6
1 1 6 .4
1 1 6 .9
1 15 .7

1 18 .8
1 2 4 .0
1 2 7 .7
1 2 8 .2
1 2 9 .2

1 Based o n h o u rs o f a ll e m p lo y e e s .
2 Based o n in d e x e s o f real p r o d u c t and in d e x e s o f th e h o u rs o f all p e rso n s engaged.
T h e in d e x o f real p r o d u c t is based o n a m aasure o f value a d d e d in c o n s ta n t d o lla r s and
d iffe r s in c o n c e p t fr o m th e p h y s ic a l o u t p u t in d e x f o r c o a l m in in g . H o u rs o f a ll p e rs o n s
c o v e r e s tim a te d h o u rs o f a ll p e rs o n s engaged, in c lu d in g p r o p r ie to r s and u n p a id f a m ily
w o rk e rs , a n d are based p r im a r ily o n d a ta fr o m th e B L S s u rv e y o f bu sin e ss e s ta b lis h m e n ts ,

U .S . D e p a r tm e n t o f E n e rg y , E n e rg y I n fo r m a tio n A d m in is tr a tio n .

p = p r e lim in a r y .
SOURCE:

T a b le 1 7 . P ro d u c tiv ity in coal m in in g , selected S tates, 1 9 6 9 and 1 9 7 8
O u t p u t p e r m in e r d a y
S ta te
19 6 9

19 7 8

Change
19 6 9 -7 8

T a b le 19. U n d e rg ro u n d coal m ines, o u tp u t p er e m p lo y e e s h ift. U n ite d S tates and
selected co u n tries, 1 9 6 0 and 1 9 7 8

P e rce n ta g e o f
to ta l 19 7 8
p r o d u c t io n f r o m
s u r f a c e m in e s

S h o r t to n s
C o u n tr y
19 6 0

S h o rt to n s

W est:
M o n t a n a ............................
N o r th D a k o t a ..................
W y o m in g .............................

B u re a u o f L a b o r S ta tis tic s .

19 7 8

P e rc e n t
change,
19 6 0 -7 8
-2 2 .5

9 7 .6
7 8 .4

2 9 .0
2 3 .7

13 .9
1 4 .4

-1 5 .1
-9 .3

4 8 .9
5 6 .2

2 5 .9

13 .0

-1 2 .9

7 1 .1

16 .0

8 .5

-7 .5

8 .2 5

1.2 3

2 .1 7

7 6 .4

F r a n c e ...........................................................................................

10 0 .0

6 1.8

10 .0
2 .2
2 2 .6

10 .6 4

B e l g i u m .......................................................................................
8 7 .6
7 6 .2
3 9 .2

U n ite d S t a t e s ............................................................................

1 .3 7

2 .0 0

4 6 .0

10 0 .0

2 3 .6

9 8 .8

H u n g a r y .......................................................................................
East:
I l l i n o i s ................................
K e n tu c k y .........................
O h i o ....................................
W est V i r g i n i a ..................
SOURCE:

1 .1 0

4 2 .9

1 .2 5

1 2 .5 4

1 1 0 3 .2

P o l a n d ...........................................................................................

1.5 9

3 .9 7

1 4 9 .7

S p a in ............... ..............................................................................

.6 7

1.4 3

1 1 3 .4

U n ite d K i n g d o m .....................................................................

1.6 8

2 .4 8

4 7 .6

U .S .S .R ...........................................................................................

1.7 0

2 2 .5 2

2 4 8 .2

W est G e r m a n y ............................................................................

U .S . D e p a r t m e n t o f E n e r g y , E n e r g y I n f o r m a t i o n A d m i n i s t r a t i o n .




.7 7

N e t h e r la n d s ................................................................................

1 .7 7

3 .9 6

1 2 3 .7

1 1 9 7 4 d a ta .
'1 9 7 6 d a ta .
NO TE:
D a ta are n o t s t r ic t ly c o m p a ra b le a m o n g c o u n tr ie s a n d m a y be used o n ly as
b ro a d in d ic a to r s o f lo n g -te rm tre n d s . U .S . d a ta are o n a m in e r-d a y basis and in c lu d e b i t u ­
m in o u s c o a l a n d lig n ite p r o d u c tio n . E u ro p e a n a n d U .S .S .R . d a ta are o n an e m p lo y e e - s h ift
basis a n d in c lu d e b it u m in o u s a n d a n th r a c ite c o a l, b u t e x c lu d e lig n ite p r o d u c t io n e x c e p t
f o r d a ta f o r F ra n c e in 1 9 6 0 in w h ic h lig n ite a c c o u n ts f o r o n ly a s m a ll p e rc e n ta g e o f to ta l
u n d e rg ro u n d o u t p u t . A ll d a ta in c lu d e s u rfa c e o p e r a tio n s o f u n d e rg ro u n d m in e s .
S O U R C E S : U .N . E c o n o m ic C o m m is s io n
E n e rg y In f o r m a t io n A d m in is tr a tio n .

57

fo r

E u ro p e ;

U .S .

D e p a r tm e n t o f

E n e rg y ,

Productivity in commercial banking:
computers spur the advance
Nevertheless, output per employee hour
paralleled the trend o f the economy
during 1967-80, with the annual rate
o f growth decelerating after 1973
H

orst

Br a n d

and

Jo h n D

uke

The computer was among the major forces that spurred
labor productivity advance in commercial banking in
1967-80. The computer also facilitated great increases
in banking output. Labor requirements per unit of out­
put, however, declined rather slowly during the period.
Output per employee hour in commercial banking
rose at an average annual rate of 1.3 percent between
1967 and 1980—nearly the same as for the nonfarm
business sector as whole (1.4 percent).1 Data for a pro­
ductivity measure for years prior to 1967 are inade­
quate, and none was calculated. Output over the period
examined rose at a rate of 6.0 percent per year, employ­
ee hours, at a rate of 4.6 percent. The rise in banking
productivity was associated with strongly expanding
customer services and with advances in computer tech­
nology and their rapid diffusion throughout the indus­
try. However, the spread of branch banking, while
enhancing access to banking services, somewhat retard­
ed productivity improvement, partly because scale econ­
omies became less favorable.2
The labor productivity trend in banking paralleled
not only the long-term rate for nonfarm business but
also the significant differences in rates of change be­
tween the 1967-73 and 1973-80 periods. Over the earli­
er span, productivity in banking rose at an average

annual rate of 2.1 percent, compared with 1.9 percent
for all of nonfarm business. Subsequently, the rate de­
celerated to 0.7 percent a year; for nonfarm business, to
0.9 percent.
Year-to-year swings in the productivity trend were
pronounced, ranging from a drop of 6.9 percent in 1974
to a spurt of 6.1 percent in 1976. During the 12-year
period, years of decline occurred 4 times, characterized
by employment increases in the face of slowed advances
(1969 and 1979) or declines in output (1974 and 1980).
In such years, restrictive monetary policy (as in 1969
and 1979) or recession (as in 1974 and 1980)
constrained the demand for funds. In years when pro­
ductivity gains ran substantially ahead of the long-term
trend average, strong cyclical recoveries or peaks in the
demand for banking services occurred (as in 1971, 1973,
1976, and 1977).3

Measuring productivity
The labor productivity measure for commercial bank­
ing has been developed in accordance with the usual
procedures of the Bureau of Labor Statistics for measur­
ing changes in the relation between the output of an in­
dustry and the employee hours expended in producing
that output. Commercial banking produces a variety of
outputs, that is, services to the public. These services
have been summed on the basis of weights which reflect
— or are close substitutes for—labor requirements per
unit of service. The output index was then divided by

Horst Brand and John Duke are economists in the Division of
Industry Productivity Studies, Bureau of Labor Statistics.
Reprinted from the
M onthly L abor Review, December 1982.




38

transactions, especially demand deposits, also lost some
momentum during the second half of the seventies.

an index of employee hours for commercial banking, so
as to obtain an index of output per employee hour, or
labor productivity. The labor productivity measure for
banking, then, measures the change over time in the ra­
tio of the weighted output of the composite of services
to the public to employee hours.
Output has been defined in terms of the three major
banking activities: (1) demand deposit transactions, in­
volving the crediting and debiting of checks written by
the public, and time and savings deposits transactions,
involving deposits upon and withdrawals from accounts
held by the public; (2) lending for commercial, con­
sumer, and real estate purposes; and (3) fiduciary, in­
volving the administration of trusts and estates, and the
purchase and sale of securities on their account.
The output measure for constructing the indexes of
labor productivity in banking has been obtained from
data on the quantity of these various services rendered
by the banks to the public. As noted, in aggregating
these services, the labor requirement per unit of each of
the major categories of service in a base period was
used as the basis for combining the dissimilar activities.
Where labor requirement data were not available, prox­
ies were employed.
The labor inputs used in constructing the productivi­
ty measure for commercial banking have been derived
from BLS data for employment and employee hours, as
reported by banking establishments on the basis of their
payroll records. The labor input series, therefore, is an
hours paid, rather than an hours worked, measure. No
adjustment has been made for differences in skill, expe­
rience, or other factors of labor quality, data for such
adjustments not being available.4

Deposits. Periods of speedup and slowdown aside, de­
mand and time deposits rose rapidly over the long term.
The number of demand deposit transactions more than
doubled. The velocity of transactions (measured by the
number of times a dollar of debits is charged against de­
posits in a given period) nearly tripled.5Furthermore, the
importance of demand deposits, a major source of
lendable funds, declined in relation to the banks’ total li­
abilities, from 43 percent in 1967 to 27 percent in 1979.6
Intensifying demand deposit activity, especially during
1967-73, contributed to pressures to introduce such la­
bor-saving procedures and equipment as electronic funds
transfers ( e f t ) . 7 Thus, according to a study conducted by
the Federal Reserve Bank of Atlanta, the number of
checks written by the public rose at an average annual
rate of 7.2 percent during the first half of the 1970’s and
declined to a rate of 5.6 percent during the second half.8
In addition to the cash-economizing efforts by the
public, evident from the tendency to hold relatively low
check balances after the mid-sixties,9 certain kinds of fi­
nancial transactions have generated large amounts of
account activity. For example, the number of shares
traded on the New York Stock Exchange in the seven­
ties averaged nearly 3 times the volume of the sixties.
Such trading usually involves multiple funds transfers
through the banking system. The number of commodity
futures contracts traded on commodity exchanges near­
ly tripled between the first and the second half of the
1967-79 period.1 Such trading also entails numerous
0
funds transfers through the banks. The underwriting of
stock and bond issues, usually by syndicates, which also
rose in the mid- and late seventies, spells the pooling of
lender funds and ultimate transfer to the borrower;
“(Debits) totaling several times the amount of the fi­
nancing involved may be recorded in this process. . . .” n
There were some developments that tended to retard
the growth of transactions and check volume—for ex­
ample, mergers, which cause book credit and debits to
replace bank transactions; bank credit cards, which
tend to consolidate individual payments; and the long­
term trend towards the output of services relative to
goods, making for fewer intermediate transactions.
These tendencies were largely offset, however, by the
upswings in manufacturing and construction, which re­
sult in numerous intermediate transactions.
Time deposits generally expanded rapidly following
the progressive liberalization of permissible rates under
the Federal Reserve’s Regulation Q. Liberalization
strengthened the banks’ position in retaining and
attracting funds which would otherwise have been in­
vested elsewhere. Savings and other time deposits held
at the commercial banks by individuals, partnerships,

Output of banking services
Output of commercial banks as measured by BLS rose
at an average annual rate of 6.0 percent between 1967
and 1980—twice as fast as output of the total private
business sector. Sources of the strong growth were the
boom conditions of the early seventies and the financial
needs they generated; rapid increases in check transac­
tions; relatively greater reliance by business on external
funds; and continuously heavy demand for consumer
and real estate credit. Also, commercial banks expanded
their share of major types of such credit, as well as of
time deposits. Moreover, they emphasized the retailing
aspects of their services and consequently accelerated
branching. Trust department functions also grew apace
as pension and other employee benefit funds proliferated.
Banking output rose at a higher rate during the
1967-73 period (7.8 percent a year) than during the
1973-80 span (4.6 percent annually). Output was damp­
ened considerably more in the recession that bottomed
in 1975 than in 1970. Loan demand rose more rapidly
prior to the 1975 recession than after. The rate of deposit



59

and corporations climbed 106 percent betvyeen 1968
and 1980, while demand deposits rose 52 percent. Time
deposits accounted for 60 percent of total commercial
banking deposits in 1980, as against 54 percent in 1968
(and 35 percent in 1960). Some observers have noted
that, in view of such technological advances as electron­
ic funds transfers, the distinction between time and de­
mand deposit accounts has become less significant.1
2

Table 1. Productivity and related indexes for commercial
banking, 1967-80
[1977 = 100]

Year

Output

Employee hours

Employees

1967
1968
1969
1970

...........................
...........................
........... ...............
...........................

83.8
85.3
84.0
85.5

52.2
56.3
60.0
64.5

62.3
66.0
71.4
75.4

63.0
66.7
72.0
76.6

1971
1972
1973
1974
1975

...........................
...........................
...........................
...........................
...........................

88.6
90.3
95.9
89.8
90.0

69.1
74.3
83.2
82.9
84.6

78.0
82.3
86.8
92.3
94.0

79.0
82.9
87.5
92.6
94.2

1976
1977
1978
1979
1980

Loans. Expansion of loan output was another source of
output growth. The rate of increase of loan output had
begun to accelerate prior to 1967, and some of the un­
derlying factors— for example, the emphasis on retail
banking— have, of course, a long history. Loan volume
being highly susceptible to the impact of the business
cycle and of monetary policy on the demand for funds,
year-to-year movements proved to be much more erratic
for lending than for the volume of deposit transactions.
The long-term trend was influenced by the increasing
propensity of business to contract for term loans (that
is, loans with maturities of more than 1 year); the con­
tinued accent upon retail banking; and banks’ growing
share of mortgage and consumer credit.
Nonfinancial business became more dependent upon
funds raised in credit markets than it had been earlier
(when corporations had relied more heavily upon inter­
nally generated funds). Between 1967 and 1980, the ra­
tio of credit market borrowing by nonfinancial business
to its capital expenditures averaged 44 percent, com­
pared with 29 percent for the earlier sixties. The compo­
sition of commercial and industrial loans shifted toward
term loans, indicating that banks were financing a
growing proportion of the plant and equipment outlays
as well as of inventories of nonfinancial business.1
3
Banks also stepped up their consumer credit opera­
tions. Here, too, growth, of course, originated in earlier
years. The share of disposable income devoted to in­
stallment borrowing began to rise in the early sixties; at
16 percent in 1967, it continued to rise gradually to 20
percent in 1979. (In 1980, a recession year, the ratio
dropped.) Furthermore, the commercial banks expanded
their share of holdings of total consumer credit out­
standing from 42 percent in 1967 to 49 percent in 1973,
remaining at about that level from then on. This gain
was linked in part to a shift away from retail store cred­
it, together with growing consumer acceptance of bank
credit cards and check credit.1
4
Growth in banks’ real estate loans was in large part
tied to the expansion in residential and commercial con­
struction of the early seventies and to the strong recov­
ery of both after their slump in the mid-seventies.
Banks also captured a larger share of total mortgage
holdings, rising from 19 percent in 1967 to 25 percent
in 1979 (as the share of insurance companies, in partic­
ular, declined). Growth in this area of lending was in



Output per
employee hour

...........................
...........................
...........................
...........................
...........................

95.0
100.0
100.7
98.5
92.7

91.8
100.0
105.4
108.1
106.1

96.6
100.0
104.7
109.7
114.5

96.8
100.0
104.9
110.5
115.7

1.3

6.0

4.6

4.5

1967-80 average
annual rate of change
(in percent) ...........

recent years also strongly influenced by household bor­
rowing against equity in existing homes.1
5
Trust services. Long-term gains in the trust department
output of commercial banks have been associated with
the growth in the number of fiduciary accounts and the
activity these accounts generate.
Between 1968 (when pertinent data first became
available) and 1980, the number of such accounts rose
54 percent.1 The increase was linked to a more than
6
threefold rise in employee benefit accounts, reflecting
the spread of corporate retirement and other employee
benefit plans, as well as of pension plans initiated by
self-employed persons (Keogh plans).1 The number of
7
personal trust accounts rose by two-thirds; they still
constitute the single most important trust department
service, representing more than three-fifths of bank-ad­
ministered trust accounts. Their rise has in part been re­
lated to the desire to shelter current income from
taxation, notably as inflation has tended to push in­
comes into more heavily taxed brackets.1
8

Employment and changing skills
Employment in commercial banking, currently num­
bering 1.5 million persons, rose 84 percent between
1967 and 1980, dr at an average annual rate of 4.5 per­
cent. Average weekly hours tended to decline some­
what, from 37.1 in the first 5 years of the period to 36.5
since then— owing chiefly to the employment of more
part-time workers.1 In no year did aggregate employee
9
hours decline, but their most vigorous rise occurred
over the first half of the review period (5.6 percent an­
nually). That high rate was not equaled even during the
cyclical recovery following the 1975 slump. From 1974
to 1980, gains averaged 3.8 percent annually.
60

Ncrasupervisory jobs accounted for nearly four-fifths
of commercial banking employment in 1980. Of these
jobs, office and clerical positions again accounted for
four-fifths of employment in the top 100 banks, or 37
percent of total banking employment in 1980. Women
staffed 85 percent of these jobs and about one-third of
all officer positions. They accounted for two-thirds of
banking personnel in 1980, compared with 41 percent
of all payroll employment.2 The prevalence of relatively
0
low-skilled jobs in banking is reflected by the ratio of
average hourly earnings in the industry to average hour­
ly earnings in the private economy. Despite the growth
of positions in computer programming and systems
analysis, that ratio has tended to decline, from 0.87 in
the sixties to 0.73 in 1980.
Supervisory jobs in commercial banking have in­
creased in both absolute and relative terms. Such jobs
accounted for 23 percent of employment in 1980, as
against 17 percent in 1967, an increase of 144 percent.
Nonsupervisory jobs rose 65 percent. The ratio of
nonsupervisory to supervisory employees thus dropped
from 5:1 in 1967 to slightly more than 3:1 in 1980. The
increase in supervisory workers was in large part linked
with the expansion of branching and the attendant
needs for managerial personnel. It was also related to a
rise in the number of loan officers, especially for install­
ment loans, and of credit analysts, who are frequently
charged with supervisory responsibilities in addition to
their regular work.
Skills needed by commercial banking employees have
changed considerably, even during the relatively short
period examined here. For example, the number of
bookkeeping operators has dropped by more than one
half since 1969 (and by more than 90 percent since
I960)— owing to the spread of electronic bookkeeping
machines and computers, which require substantially
fewer operators.2 Also, tellers have tended to become
1
less specialized as branch banking has spread. The six
usual teller classifications—note, commercial-savings,
commercial, savings, vault, and all-round— have in
many banks been reduced to one all-round teller classi­
fication. The practice of classifying tellers by commer­
cial or saving transactions has been declining.2
2
Most bank employees perform tasks related mainly to
the banks’ depository functions and loan administra­
tion. A high school education is generally considered
adequate preparation for entry level jobs. Bank officers,
on the other hand, usually supervise the various finan­
cial and customer services. Loan officers, in particular,
are expected to be knowledgeable about the industries
from which the individual bank draws its customers
and to be sensitive to the often unique problems cus­
tomers present— problems which frequently require
handling on a personal basis. Officers usually have a
college degree or an M BA.23



61

The labor inputs of commercial banks thus vary
widely in terms of education, training, and skill com­
plexity. Also, wide differences exist between the tasks
that can be automated and tasks that cannot be, with
the work of loan officers being least susceptible to stan­
dardization and automation. However, even in this area,
a growing number of supplementary tasks have been
computerized.2
4

Fixed investment and technology
Between 1967 and 1979, banks’ fixed capital, includ­
ing structures, furniture, and equipment, rose by a fac­
tor of three, while the stock of fixed nonresidential
capital in the private business economy as a whole rose
by a factor of nearly four.2 Price indexes to deflate the
5
banks’ physical capital stock are not available, so no
firm estimate of movements in constant-dollar value can
be offered. When the deflators for the total capital stock
of business are applied to that of the banks, a rise of
about one-third in real terms would result.
About 40 percent of the banks’ spending on fixed
capital went for equipment and furniture during the re­
view period. In 1980, roughly half of the banks’ expen­
ditures for fixed capital other than structures was spent
on computers and computer equipment.2 Fixed capital
6
per employee in commercial banking, at about $16,000
in 1979, ran at three-fifths of the comparable figure for
the business economy.2
7
Computer breakthrough. At the root of equipment
spending has been the transformation of technology by
electronic data processing ( e d p ). While banks progres­
sively mechanized their routine operations throughout
the forties and fifties, the resulting efficiencies improved
but gradually. Some students of the field, in fact, attrib­
uted these efficiencies more to the specialization of labor
and economies of scale in the industry than to mechani­
zation.2 A 1960 study by the Federal Reserve Bank of
8
Philadelphia stated, “Since World War II, banks appar­
ently have expanded operations more by hiring extra
people than by using better equipment.”2 According to
9
the study, the technology used in banks had scarcely
changed during most of the first half of the 20th centu­
ry. The same basic types of cash registers, punched card
tabulators, billing and duplicating machines, and check
signing equipment found in banks in 1914 were still the
mainstay of banking technology at the end of World
War II.
Although computer developments during the fifties
embodied the principle of machine readability, it was
the introduction of magnetic ink character recognition
(micr ) in 1958 that made the breakthrough of electron­
ic data processing in banking possible. The computer
became an indispensable and major factor in improving
banking productivity. Moreover, computer technology

has rapidly spread throughout the industry. The first
bank automation survey conducted by the American
Bankers Association in 1963 showed only 7 percent of
all commercial banks to be users of on-premise or offpremise computers. By 1968, 49 percent were users, and
in 1980, when the latest available survey was conduct­
ed, 97 percent were. The pressures of cost efficiencies,
organizational changes, and competition had reduced
the proportion of surveyed banks without plans to auto­
mate from 84 percent in 1963, to 42 percent in 1968, to
virtually nil in 1980.3 While the larger banks— those
0
with $100 million-plus in deposits—generally maintain
their own computer operations, smaller banks have in­
creasingly used their correspondent relations with the
larger banks to gain access to computers. As of 1980,
26 percent of all banks operated on-premise computers,
while 71 percent used off-premise computers, mostly at
correspondent banks.3 Thus, size of bank, as measured
1
by the dollar value of deposits, does not appear to have
seriously inhibited the diffusion of EDP technology in
the industry.
The computer has had its greatest impact upon the
deposit function, particularly upon check handling. Its
full potential, however, is only beginning to be realized,
inasmuch as optimally most payments transfers could
be processed electronically, that is, without checks. But
only a small proportion of payments is so processed at
present. Each check is, in effect, “a special piece of cur­
rency, created for one transaction only, that has to pass
through complex and repeated identification, verifica­
tion, accounting, and sorting operations before it is re­
tired.” 3 Until the mid-seventies, the enormous and
2
steadily growing volume of checks (estimated at 32 bil­
lion in 1979) was expected to become too expensive to
handle, even by computer. But evolving technology has
expanded the check-processing capacity of computers,
such that they are thought to be able to “handle any
conceivable number” of checks.3
3
The currently most advanced (or “third-generation”)
computer has a built-in reader-sorter processing capaci­
ty of 120,000 checks per hour. Manual reading and
sorting of checks, which for many years has involved
some machine processing such as high-speed readers,
averages 1,200 to 1,400 checks, so that computer use
for this phase of the check-handling process represents
“order of magnitude” reductions in labor requirements.3
4
For other phases of check-handling, comparable pro­
ductivity advances have not been attained, although socalled rejects or exception items, which in earlier years
required laborious interbank correspondence, have come
to be processed with great efficiency thanks to coopera­
tive agreements. According to surveys by the Bank Ad­
ministration Institute, the average labor requirements
for all phases of handling checks were reduced by well
over one-half between 1970 and 1979 among surveyed



banks, mainly because of computerized reading and
sorting of checks and more efficient handling of excep­
tion items.3
5
In loan operations, EBP has been used for information
retrieval, as well as in the administration and bookkeep­
ing operations of such loan categories as installment
loans. Credit information, mortgage servicing, bank
credit card billing, and accounting have also been
among major computer applications. The proportion of
personnel in installment loan operations has tended to
decline, but the available data do not clearly point to
improved productivity in this area of banking. Staff
employed in handling bank credit cards— also a type of
consumer credit— has expanded in recent years.3 Busi­
6
ness loan operations, which require a comparatively
small proportion of bank personnel, have remained rela­
tively labor-intensive—largely owing to their specialized
nature and the need for maintaining close customer
contact. Even here, however, the computer is playing an
increasing role. It is used to provide up-to-date credit
analyses and to serve as a bankruptcy predictor. For
the larger banks, it makes credit information on a
worldwide basis rapidly available. It also facilitates the
collection and arraying of data to meet the requirements
of regulatory authorities, a task that is otherwise highly
labor-intensive.3
7
Computer technology has also contributed to im­
proved productivity in trust departments. It has been
primarily applied to information retrieval for purposes
of controlling individual accounts.3 But it has been in­
8
creasingly used as well in stock trading by trust depart­
ments for customer accounts. With trust departments
holding the largest share of assets in stocks (49 percent
in 1980 by value), such trading accounts for the major
part of their activity. The basis of automated stock
trading has been a numbering system first devised by
the American Bankers Association’s Committee for Se­
curity Identification Procedures in 1968. The use of
committee numbers on stock certificates was mandated
by the Securities and Exchange Commission in 1971.
This and similar systems have tended to standardize
stock identification and have contributed to the transfer
of stock without the physical handling of stock certifi­
cates. These certificates are “immobilized,” that is, they
remain in central depositories. Costly errors and redun­
dant bookkeeping entries have been nearly eliminated
when trust departments have adopted the technology on
which the bankers’ stock transfer system is based.3 Pay­
9
ments and credits involving stock transfers likewise use
the system. Relative to output, trust department per­
sonnel requirements have been evidently reduced as a
result of these and other computer applications.4
0
Electronic Funds Transfer. Potentially the most impor­
tant use of the computer in banking remains electronic
62

ings in labor costs at current volumes of business—a
factor that tends somewhat to retard the diffusion of
the devices.4 According to one authority, 19,000 auto­
5
mated teller machines were in use at the end of 1980,
each averaging about 4,600 transactions per month,
more than 2.5 times the volume 4 years earlier— signi­
fying rapid consumer acceptance of this technology.4
6
The banks have also installed much technologically
advanced equipment other than computers and teller
machines. For example, word-processing equipment is
now being operated in four-fifths of the larger and twofifths of the smaller banks. Optical character recogni­
tion equipment—used, for example, in the processing of
credit card charge slips, checks, and direct bill pay­
ments— has likewise been installed in most larger
banks.4
7

funds transfer ( e f t ). Although the technology for EFT
has existed for nearly two decades, its acceptance by the
public has been comparatively slow. Also, a large part
of the costs of the check collection system and of de­
mand deposit transactions was absorbed by the Federal
Reserve and the banks, rather than passed on to users.
Nevertheless, EFT has been increasingly adopted by the
banks since the mid-seventies. Competition among fi­
nancial institutions, as well as the developing cost ad­
vantages of ED P over conventional transfer activities, are
likely further to speed adoption of EFT technology.4
1
EFT has been increasingly applied in interbank settle­
ments through automated clearinghouses and in basic
kinds of teller operations involving customer services,
such as deposits and withdrawals, direct deposit of pay­
rolls or other recurring payments, direct bill payment,
and transfer of funds from savings accounts to demand
deposit accounts and vice versa. Point-of-sale terminals,
linking merchants with a network of local banks, have
also been spreading, although their acceptance and use
have remained limited.4
2
Automated clearinghouses have spread rather gradu­
ally, although they have not replaced the conventional
clearinghouse process as they handle only paperless
credit and debit entries between banks. Originating in
San Francisco in 1972, automated clearinghouses cur­
rently link an estimated 14,000 financial institutions and
their offices; they process an estimated 300 million items
annually.4 This number represents but a small fraction
3
of the total number of checks drawn on banks other
than the payor’s own bank, but it is expected that auto­
mated clearinghouses will account for a rising propor­
tion of all items in the clearing process. Among reasons
for this expectation have been the success of the direct
deposit of social security payments and of a growing
number of public and private payrolls; the associated
savings in mailing costs; less work incident to replacing
lost checks, and the cost pressures linked to the han­
dling of paper items (which despite the increasing effi­
ciency of the process has been more and more comple­
mented or replaced by e f t ) .44

The growth of branch banking
The number of commercial banking firms barely rose
5 percent between 1967 and 1979. But the total of
banking offices increased 62 percent, mostly reflecting a
doubling in the number of bank branches, and a con­
tinuing shift of offices towards the suburban population
centers of metropolitan areas. The average population
served per bank office declined from nearly 6,000 per­
sons in 1970 to 4,400 in 198Q.4 The decline suggests
8
that banking services became more widely and conven­
iently available to the public. Current-dollar disposable
income per capita nearly tripled during 1967-80 (as did
personal consumption expenditures), and households
generated an expanding volume of banking business,
supporting the spread of branch banking.
Most banks are comparatively small. Those holding
total deposits of up to $50 million represent 79 percent
of all commercial banks, but in 1979 accounted for only
15 percent of total deposits. Smaller banks usually
maintain correspondent relations with the larger banks,
and this relation amounts to a “form of multi-office
banking.”4 Some of the efficiencies or customer utilities
9
associated with large-scale banking are likely, therefore,
to be shared throughout most of the industry.
The larger banks, however, are dominant. The share
of deposits held by the Nation’s largest banks— those
with deposits of $500 million or more— was 62 percent
in 1979. These banks constitute little more than 2 per­
cent of total banks. Moreover, in metropolitan areas,
the two largest banking organizations usually hold be­
tween 55 percent and 67 percent of deposits (the ratio
tends to be lower in unit banking States, higher in
statewide branching States).5 Adoption of computer
0
technology has been shown to be closely associated
with bank size, as well as with holding company affilia­
tion.5
1
As might be expected, banking employment is also
concentrated in the bigger banks. Banks holding $500

Teller machines. Automated teller machines spread rap­
idly in the late seventies. Providing customer access by
means of a magnetic-stripe bank card and unique iden­
tification entered upon a keyboard, the machines receive
deposits and payments and dispense cash. Twenty-four
hour access is a frequent feature, enhancing customer
convenience and reducing waiting lines. Thus, automat­
ed teller machines in effect extend banking hours, al­
though banks also view them as “peaking” equipment,
helping to reduce lobby traffic during peak hours of
business. The machines substitute capital for labor, but
for many medium- and smaller size banks, the relatively
high fixed costs of the equipment are not offset by sav­



63

million or more in deposits employed 56 percent of all
banking personnel in 1979. Banks with less than $100
million in total deposits— 89 percent of all banks —
employed 27 percent of all personnel.5
2
Among changes in the competitive pattern of finan­
cial institutions that have affected banks has been the
spread of NOW (negotiable order of withdrawal) ac­
counts at thrift institutions; their effect on the share of
time deposit accounts at commercial banks, however,
cannot be assayed yet. In some other areas, the role of
commercial banks has been eroded. More efficient cor­
porate cash management, spurred by high interest rates
and advanced information technology, has diminished
the relative importance of demand deposits. Also, com­
mercial banks have evidently been unable to expand
their share of credit cards (15 percent of 600 million
outstanding cards in recent years). Also, business and
consumer credit extended by very large department
store chains, automotive companies, farm equipment
makers, and EDP manufacturers grew in importance un­
til the early seventies, although their share of financial
assets has apparently stabilized since.5
3

have continued to improve, and processing and mailing
costs of checks to rise.
Direct deposit of payrolls and of other recurring pay­
ments, and direct bill payment will likely also expand,
partly owing to the costs of float, which banks must as­
sess as an explicit cost under recent legislation, as well
as because banks must offset the cost of handling
checks against interest on demand deposits (where such
interest is offered). Thus, resistance to EFT is likely to
lessen as costs of processing paper items rise— speeding
its diffusion.
Continued technological advances and the labor sav­
ings expected from them will probably also arise from
intensified competition by nonbank financial institu­
tions. Thus, money market funds have come to compete
with time and saving deposits for both the small and
large investor’s dollar, and this, too, may contribute to
restricting commercial banks’ output growth.5 Also,
6
more than 80 percent of all household and virtually all
business firms had checking accounts in 1977, so that
the expansion of banking services from including addi­
tional households is quite limited. A partially offsetting
factor may be a continued rise in cash withdrawals
from automated teller machines, which are believed to
be smaller and more frequent than withdrawals by
cashing checks.5 The convenience in the use of banking
7
services made possible by the machines may encourage
the banks to adopt product lines similarly appealing to
customer convenience.5
8
With the spread of EFT, and other computerized and
automated transactions, banks’ labor requirements per
unit of output are bound to continue to decline. More­
over, new branch staffing needs should be decreasing,
partly because of the technological developments dis­
cussed, partly because of the already low level of popu­
lation served per branch, and the consequent abatement
in the number of new branches opened. Hence, com­
mercial banks will probably become less important as a
source of added employment in the years ahead— also
indicated by BLS projections to 1990, which imply a
slower rate of banking employment growth than over
the past decade.

Outlook for the industry
The diffusion of EFT is likely to help improve labor
productivity in commercial banks in the years ahead.
During the late seventies, doubts about its widespread
acceptance were expressed in some quarters.5 Resis­
4
tance by consumers to abandoning payment by check
and their fear of loss of control over balances were cited
as two reasons. Regulatory questions concerning the
off-premises installation of automated teller machines
were another. Also, smaller banks were believed to have
opposed EFT because of possible competition from big
money-center banks. These obstacles to the diffusion of
EFT have so far been only partly overcome. However,
cost considerations seem likely to compel its more rapid
adoption. To illustrate, in a study of the benefits of
electronic government payments done in 1977, the Fed­
eral Reserve found the costs of EFT to run nearly twothirds below the costs of processing checks.5 The ratio
5
has lessened since then, for the scale economies of EFT

1 Commercial banks are establishments primarily engaged in accept­
ing deposits from the public and making loans and investments. They
are designated as No. 602 in the Standard Industrial Classification
(SIC) Manual of the Office of Management and Budget. The industry
is part of SIC 60— banking, which also includes Federal Reserve
Banks, mutual savings banks, trust companies not engaged in deposit
banking, and establishments performing functions closely related to
banking. Nonbanking subsidiaries of bank holding companies are not
included; they are separately classified by primary activity. See Federal
Reserve Bulletin, December 1972. Commercial banks account for ap­
proximately 90 percent of the employment of the total SIC 60 group.
A detailed description of banking output and of the procedures
followed in measuring banking productivity, output, and employee-




64

hours, as well as the weighting scheme underlying the output mea­
sure, is available upon request.
2
There is wide agreement among industry observers that scale econ­
omies in banking have declined with the spread of branching — that
is, more resources, including labor inputs, are required per unit of
output. Among definitive studies are Costs in Commercial Banking, by
Frederick W. Bell and Neil B. Murphy (Federal Reserve Bank of Bos­
ton Research Report No. 41, April 1968), and “Economies of Scale
and Marginal Costs in Banking Operations,” by George J. Benston
{The National Banking Review, June 1965), reprinted in that report.
Industry observers confirm that the tendencies analyzed in these
works have persisted.

nors, Federal Reserve System, July 1968). Banks’ adoption and opera­
tion of credit plans of their own has had significant implications for
their output: although credit cards result in consolidation of payments
and, therefore, reduce the number of check transactions, they generate
sales drafts which must be cleared through merchant’s deposit ac­
counts. Thus, they augment “the paperwork burden to the extent that
(they replace) cash in a retail transaction” (p. 63.)
1 David F. Seiders, Mortgage Borrowing Against Equity in Existing
5

3Professor Charles F. Haywood of the College of Business and
Economics, University of Kentucky, interprets the swings in commer­
cial banks’ labor productivity as follows:
. . (At) the beginning of
an upswing, banks have some slack in manpower and can increase
output somewhat without increasing the rate of new hires. At some
point in the upswing, the rate of new hires has to be increased. By the
time these new hires are in place, the upswing in the economy is near
its end and recession soon follows. There may also be some variation
in labor turnover rates related to cyclical variation in the economy
that affects input-output relationships in banks . . . (As) turnover
rates are high in banking, cyclical variation in such rates could have
significant effects on productivity.” Communication to the BLS Office
of Productivity and Technology.
4Among authorities upon whose conception of the banks’ functions
and output the BLS definition is partially based is Professor Donald
Hodgman of the University of Illinois. Hodgman has viewed banking
activity as consisting of a bundle of services, grouped into three cate­
gories: management of the national payments mechanism; intermedia­
tion between borrowers and lenders; and specialized financial services
(of which trust activities are by far the most important ones). See
Donald Hodgman, Commercial Bank Loan and Investment Policy (Urbana, University of Illinois, 1963), p. 165 ff; and John Gorman, Com­
ment, “Real Output and Productivity of Banks,” in Victor R. Fuchs,
ed., Production and Productivity in the Service Industries (New York,
National Bureau of Economic Research, 1969), p. 189 ff.
5See Banking and Monetary Statistics, 1941-1970, Board of Gover­
nors, Federal Reserve System, p. 321 ff., for a detailed explanation of
the turnover rate of demand deposits.
6These and other data on commercial banks’ shares in financial as­
sets or liabilities were calculated from data from Flow o f Funds Ac­
counts (Board of Governors, Federal Reserve System), various recent
issues.
7 “Earliest concern with the payment system was rooted in the fear
that growing check volumes posed a threat to the continued satisfac­
tory performance of the system. Studies sponsored by the Federal Re­
serve System and by several national associations of commercial
banks in the 1960’s placed virtually their entire emphasis on two
areas: measuring the national check volume, the pattern of the flows
of checks into and through the banking system, and check processing
costs; and offering technical and economic feasibility assessments of
electronic alternatives of the time to check clearing and collection sys­
tem. The emphasis throughout was on the use of electronic means to
replace checks, or to reduce check handling, through systems created
and cooperatively operated by groups of commercial banks, with a
key role implied or advocated for the Federal Reserve System.” Ed­
win B. Cox, “Developing an Electronic Funds Transfer System:
Incentives and Obstacles,” The Economics o f a National Electronic
Funds Transfer System, proceedings of a conference held in October
1974 (Federal Reserve Bank of Boston), p. 16.
1A Quantitative Description o f the Check Collection System: Vol. 1, a
report of research findings on the check collection system,
cosponsored by the American Bankers Association, Bank Administra­
tion Institute, and Federal Reserve System (Atlanta, Ga., Federal Reseive Bank, 1981), p. 1.
See also Bryan Higgins, “Velocity— Money’s Second Dimension,”
Monthly Review, Federal Reserve Bank of Kansas City, June 1978,
and George Garvy and Martin R. Blyn, The Velocity o f Money (New
York, Federal Bank of New York, 1970), p. 69.
“New York Stock Exchange, Fact Book 1980, and U.S. Commodi­
ty Futures Trading Commission, Annual Report ( 1980).

Homes: Measurement, Generation, and Implications for Economic Activ­
ity (Board of Governors, Federal Reserve System, 1978), Staff Eco­

nomic Studies 96.
1 See Trust Assets o f Banks and Trust Companies (Board of
6
Governors, Federal Reserve System; Federal Deposit Insurance Cor­
poration; and Office of the Comptroller of the Currency), 1980 and
earlier years.
1 Indicative of the increase in corporate pension and welfare plans
1
is the rise in the number of such plans reported by the U.S. Depart­
ment of Labor. As of January 1, 1970, 157,400 such plans were re­
ported, the number rising to 554,000 by 1977. The bulk of the assets
in which the plan administrators invest consists of stocks and bonds.
See Welfare and Pension Plan Statistics, 1967, 1969, and 1971 (U.S.
Department of Labor, Labor-Management Services Administration),
and information from LMSA.
1 Interview with a banking representative.
8
1 Part-time workers accounted for almost one-sixth of all non9
supervisory office workers in surveyed commercial banks in 1980, up
from one-eighth in 1976, according to Industry Wage Survey: Banking,
February 1980, Bulletin 2099 (Bureau of Labor Statistics), p. 3.
2 Equal Employment Opportunity Commission Summary Statistics,
0
Top 100 Full Service Banks.
2 Technological Change and Manpower Trends in Six Industries, p.
1
51, and Industry Wage Survey: Banking, p. 4.
2 Industry Wage Survey: Banking, p. 4.
2
2 See Banking and Insurance Occupations, Bulletin 2075-7 (Bureau
3
of Labor Statistics).
2 David M. Coit, “Automated Financial Analysis: A New Tool for
4
Commercial Lending,” The Journal o f Commercial Bank Lending,
March 1977.
2 Assets and Liabilities o f all Commercial Banks in the United States,
5
Annual Report for 1980 and Earlier Years (Washington, Federal De­

posit Insurance Corp.).
2 Information on the average annual expenditures per bank for
6
computer equipment, 1980-82, is provided in table 224 of National
Operations/Automation Survey, 1981 (Washington, American Bankers
Association).
2 The prices for computer hardware, as well as for calculating and
7
accounting machinery, widely used by the banks, rose much more
slowly than producer durables prices generally or tended to decline
over part or all of the review period. See Robert B. Archibald and
William S. Reece, “Partial Subindexes of Input Prices: The Case of
Computer Services,” Southern Economic Journal, October 1979, pp.
528-40. The authors show that second generation computers, manu­
factured for large business uses by IBM, dropped in price by 85 per­
cent between 1970 and 1975. Reasons for the drop are discussed by
them. At present, the BLS imputes movements in the value of com­
puter hardware to the office and store machines and equipment group.
2 See Bell and Murphy, Costs in Commercial Banking, discussion in
8
chapter VII, p. 105 ff.
2 “How Banking Tames its Paper Tiger,” Business Review (Federal
9
Reserve Bank of Philadelphia), June 1960.
3 See National Operations/Automation Survey 1981 (Washington,
0
American Bankers Association), p. 7.

' Garvy and Blyn, The Velocity o f Money, p. 43.
2 “Increasing Competition between Financial Institutions,” Eco­
nomic Perspectives (Federal Reserve Bank of Chicago), May/June
1977, p. 23 ff.
1 Term loans rose from 40 percent of total commercial bank loans
3
in 1967 to 44 percent in 1973 and 48 percent in 1978.
1 For some reasons why banks attempt to expand their credit card
4
systems, see “EFT in the United States, Policy Recommendations and
the Public Interest,” The Final Report of the National Commission
on Electronic Fund Transfers (Washington, October 1977), p. 134.
See also Bank Credit-Card and Check-Credit Plans (Board of Gover­




3 Ibid.
1

3 John E. Sheehan, “Higher Productivity Demand Deposits,” in
2
The 1972 National Operations and Automation Conference Proceedings

(Washington, American Bankers Association), p. 363.
3 John S. Reed, executive vice president of Citibank, quoted in
3
“Electronic Banking: A Retreat from the Cashless Society,” Business
Week, Apr. 18, 1977. See also Sanford Rose, “Checkless Banking is

65

17. See also David A. Walker, An Analysis o f Changes in EF TS Activi­
ty Levels, Costs and Structure in the U.S.: 1975 to 1977 (Washington,

Bound to Come,” Fortune, June 1977, p. 118 ff.
3 Information from Bank Administration Institute and Federal Re­
4
serve.

Federal Deposit Insurance Corp.), Working Paper No. 77-3, especial­
ly p. 7.
4 Linda Fenner Zimmer, “ATM Acceptance Grows, Builds Cus­
6
tomer Base for Other EFT Services,” The M agazine o f Bank
Administration, May 1981, p. 31. Cited in Statistical Information on
the Financial Services Industry (Washington, American Bankers Asso­
ciation, 1981), p. 107.
4 American Bankers Association, 1978 Survey, op. cit. On the pro­
7
ductivity effects of such equipment, see also David Cockroft, “New
Office Technology and Employment,” International Labour Review,
November-December 1980, p. 689 ff.
4 Statistical Information on the Financial Services Industry, p. 89.
8
4 Carter H. Golembe, “Growth of Bank Holding Companies,” in
9
Herbert V. Prochnow, ed., The Changing World o f Banking (New
York, Harper & Row, 1974), p. 23.
5 “Recent Changes in the Structure of Commercial Banking,” Fed­
0
eral Reserve Bulletin, March 1970, p. 207.
5 See Charles F. Haywood, “Regulation, Technological Change and
1
Productivity in Commercial Banking,” in Productivity Measurement in
Regulated Industries (New York, Academic Press, 1981), p. 300-01.
5 Based on unpublished data of the Federal Deposit Insurance Cor­
2
poration.
5 Will R. Sparks, Financial Competition and the Public Interest
3
(New York, Citicorp., 1978), p. 23, also pp. 16, 17.
5 Reed, “Electronic Banking.” See also William Ford, The Pay­
4
ments System o f the 1980's, presented at the Second Annual Shared
EFT Systems Conference, Atlanta, Ga., Feb. 5, 1981 (Federal Reserve
Bank of Atlanta).

3 See 1979 Survey of the Check Collection System (Park Ridge, 111.,
5
Bank Administration Institute, 1980).
1 Functional Cost Analysis, 1979 Average Banks. Based on data
6
furnished by 751 participating banks in 12 Federal Reserve districts.
Computer processing of bank credit card transactions has remained
similar to that of checks and therefore is technologically not as ad­
vanced as computer processing of transactions under credit cards is­
sued by the big oil companies, where optical character recognition has
been part of the computer operation. (Conversation with ABA repre­
sentatives.)
’’“Automated Financial Analysis.”
3 Third Trust Operations and Automation Workshop, 1972 Proceed­
8
ings (Washington, American Bankers Association). See also The
Bottom Line: Proceedings, 1976 National Trust Operations and Automa­
tion Workshop, New York, March 21-24, 1976, remarks by William
Schladebeck, p. 216 ff.
3 Third Trust Operations— Proceedings, p. 58.
9
4 H. Russell Morrison, “CUSIP Report— Beyond Apr. 1, 1972,”
0
Third Trust Operations < Proceedings, p. 58.
6
4 See N. Sue Ford, “Electronic Funds Transfer: Revolution Post­
1
poned,” Economic Perspectives (Federal Reserve Bank of Chicago),
November-December 1980, p. 16 ff. Competition between different
types of financial institutions has been fostered by high interest rates
together with NOW (negotiable order of withdrawal) accounts at
thrift institutions, and of share drafts at credit unions. Such instru­
ments have been authorized on a national basis by the Deregulation
and Monetary Control Act of 1980. A detailed analysis of this law
may be found in Economic Perspectives, September-October 1980,
p. 3 ff.

5 Costs, Savings and Benefits o f Electronic Government Paym ents
5

(Unpublished study by the Division of Federal Reserve Bank Opera­
tions, Board of Governors, Federal Reserve system, June 1977).
3 See “The Changing Environment for Banking,” an address by J.
6
Charles Partee, before the American Institute of Certified Public Ac­
countants Annual National Conference on Banking, Capitol Hilton,
Washington, D.C., Dec. 4, 1980. Also, “America’s New Financial
Structures,” Business Week, Nov. 17, 1980, p. 138 ff.; and Constance
Dunham, “The Growth of Money Market Funds,” New England Eco­
nomic Review (Federal Reserve Bank of Boston), September-October
1980, p. 20 ff.
3 On the factors influencing the evolution of EFT and the check
7
payments system, see The Paym ents System o f the 1980's, op. cit.
3 Some nonbank services built into ATM’s are noted in “Diebold’s
8
Shift to Automated Tellers Works,” by Margaret Yao, The Wall
Street Journal, July 15, 1982, p. 45.

4 Ford, “Electronic Funds Transfer,” p. 18.
2
4 Haywood, communication to the BLS. See Philip E. Coldwell,
3
“The ACH in Perspective” (Remarks at the 4th Annual NACHA
Surepay Conference, Houston, Tex., Mar. 13, 1979), p. 3.
4 Ford, “Electronic Funds Transfer,” p. 16. See also Carl M.
4
Gambs, “Automated Clearinghouses— Current Status and Pros­
pects,” Economic Review (Federal Reserve Bank of Kansas City, May
1978), p. 3 ff.
4 ATM’s often “substitute . . . for a more costly full-service brick 5
and-mortar branch.” Haywood, communication to BLS. Another ob­
server has stated that, “The ATM also reduced the need for tellers,
lowering not only the salary cost to the bank, but also of employee
benefits and pension plans.” Ford, “Electronic Funds Transfer”, p.




66

Productivity and new technology
in eating and drinking places
Labor-saving techniques fo r preparing meals,
the rapidly expanding fa st food chains
and a decline in the number o f drinking places
have altered output and hours in the industry

,

R ic h a r d B. C a r n e s a n d H o r s t B r a n d

Productivity in eating and drinking establishments1
rose at an average annual rate of 1.0 percent between
IS1 (when adequate data became available) and
SB
1976, but varied widely over the 18-year span. Output
increased 3.1 percent annually and hours, 2.1 percent.2
(See table 1.) During the same period in the private
economy, productivity advances averaged 2.8 percent a
year.
Factors that have contributed to the advance of
productivity in the food service industry are the
spread of modem management techniques and work
organization, particularly in the rapidly expanding
fast food segment of the industry. Menus have been
simplified and standardized, and menu items are in­
creasingly prepared off premise, reducing on-premise
employee-hour requirements. Layouts of establish­
ments are designed to minimize walking time of per­
sonnel. Technological innovations, such as the mi­
crowave oven, reduce cooking time. Finally, the
decline in the number of single-unit drinking estab­
lishments (usually proprietorships and partnerships)
has resulted in the disappearance of marginal enter­
prises.
Trends,
Between 1958 and 1964, output per hour rose at
an average annual rate of only 0.5 percent, reflecting
gains below the long-term trend in both output and
aggregate hours. Between 1964 and 1968, productiv­
ity increases accelerated, averaging 2.3 percent a

Output and growth factors. The industry’s output
gains reflect an upward trend in real per capita
spending on meals eaten away from home. At $159
per capita in 1976 (constant 1972 dollars), such
spending has risen 24 percent since the mid-1960’s
with all of the rise having occurred since 1973.3The
relation between changes in industry output and
changes in real per capita income is illustrated below:
O u tp u t
1 9 5 8 -7 6
1 9 5 8 -6 3
1 9 6 3 -6 8
1 9 6 8 -7 6

............................................
............................................
............................................
............................................

R e a l p e r c a p ita
in c o m e

3.1
1.6
4 .0
3 .0

2.8
1.5
4 .0
2.3

Relatively slow advances in real per capita income
were associated with modest output growth in the
earlier years; while rapid income increases were
linked with accelerated output rates in the later
years.
The long-term trend in food service output was
also influenced by the more rapid rise in the number

Richard B. Carnes and Horst Brand are economists in the Division of
Iedustry Productivity Studies, Bureau of Labor Statistics.

Reprinted from the
M o n t h l y L a b o r R e v ie w , September 1977.




year, as output grew rapidly and hours advanced
moderately. From 1968 forward, productivity im­
provement again slackened to 0.4 percent a year;
however, output continued to expand vigorously, ac­
companied by relatively large increases in hours.
Year-to-year changes in the trend of labor produc­
tivity deviated significantly from the long-term aver­
age. The largest annual gain, 3.4 percent, occurred in
1970; the largest decline, 2.2 percent, occurred the
following year, when output rose slightly and hours
expanded sharply.

67

of households headed by unattached individuals (97
percent between 1960 and 1975) than in the number
of families (24 percent). Such individuals are more
likely to eat out than families: according to the latest
Bureau of Labor Statistics consumer expenditure
survey, 1-person households, on average, spend 40
percent of their food budgets on meals away from
home, compared with 25 percent for families (the
proportion diminishes as size of family increases).4
Moreover, between 1960 and 1975, real incomes rose
faster for unattached individuals than for families—
56 and 34 percent, in constant dollars.5
Another important factor that bolstered output
gains in food service was the increase in the number
and proportion of wives in outside employment, con­
tributing to family incomes. In 1975, 44 percent of
all wives (husband present) held a paid job, com­
pared with 31 percent in 1960. Real income of such
families climbed 37 percent over that period; real
incomes increased 27 percent for families with wives
not in paid employment. The absolute difference in
income between the two categories, 35 percent in
1975, made a significantly larger absolute difference
in their spending on meals eaten out—51 percent.6
The overall increase in spending for meals' and
snacks eaten away from home was accompanied by
a shift from full-service restaurants to fast-food es­
tablishments. This shift has given rise to greater fre­
quency of eating out and to consumption of lower
priced meals. The share in total industry receipts of
restaurants and lunchrooms declined from 62 per­
cent in 1958 to 50 percent in 1972; over the same
period, the share of refreshment places (which in­
cludes most fast-food units) quadrupled, and stood at
26 percent in 1972. Commercial cafeterias raised
their share from 6 to 8 percent. The remaining share

T a b le 1.
fo o d

of industry receipts is accounted for mostly by drink­
ing establishments which, like restaurants, suffered a
large loss in market penetration over the 1958-76
period.
The average transaction in fast-food establish­
ments is about three-quarters of that in full-service
restaurants.7This does not mean that consumer pref­
erences have shifted to cheaper foods; surveys of
representative menus conducted by Institutions/
Volume Feeding do not indicate significant changes
in the choice of the major classes of breakfast foods,
dinner entrees, or desserts.8 Rather, the evident in­
crease in the number of transactions at fast-food es­
tablishments has been accompanied by fewer services
rendered to consumers (when compared with fullservice restaurants) and by a decline in the variety of
foods offered, reflecting standardized menus.9
Employment doubles and hours moderate. Employ­
ment in eating and drinking places (currently 3.7
million) doubled between 1958 and 1976, rising at an
average annual rate of 3.9 percent. Its growth, like
that of output, was comparatively slow between 1958
and 1963 (1.7 percent annually), but accelerated
from 1964 forward at an annual rate of 4.6 percent.
Total hours of persons engaged in the industry
rose about half as much as employment, with aver­
age weekly hours for nonsupervisory workers declin­
ing from 35.6 in 1958 to 28.0 in 1976. This drop in
weekly hours resulted in part from the expansion of
part-time work. In 1975, 51 percent of all workers in
the industry worked part time, compared with 32
percent in 1962.1 Moreover, the number of proprie­
0
tors and partners dropped, and the working hours of
supervisory personnel declined from an estimated 61
hours in 1958 to 51 hours in 1975.
The occupational composition of food service
workers has not changed significantly since detailed
data first became available in 1972. (See table 2.)
More than half of the employees occupy positions
such as waiters, waiter assistants, counter and foun­
tain workers, or dishwashers. About one-third were
cooks and bartenders and the remainder performed
clerical, or managerial and administrative tasks.1
1
Limited data for earlier years indicate a steady con­
traction in the number of waiters and waitresses, and
an expansion in jobs associated with counter work.
In general, trends in food and equipment technology,
together with organization changes, have increas­
ingly favored the employment of low-skilled persons
in the industry—developments also promoted by ris­
ing labor costs,1 and the difficulty of attracting a
2
stable work force.
Data on the work experience, age, and sex of food
service workers indicate a generally high turnover
rate. The proportion in full-time, year-round jobs, 22

I n d e x e s o f p r o d u c t i v i t y , o u t p u t , a n d h o u r s in

s e r v ic e e s t a b lis h m e n t s , 1 9 5 8 -7 6

[1 9 6 7 = 100 ]

Yoar

Output par hour
of all
parsons

Output

Hours of
all parsons

1958 ...............................
1959 ...............................
1960 ...............................
1961...............................
1982 ...............................
1963 ...............................
1984 ...............................
1985 ...............................
1966 ...............................
1967 ...............................
1968 ...............................
1969 ...............................
1970 ...............................
1971...............................
1972 ...............................
1973 ...............................
1974 ...............................
1975 ...............................
1976 ...............................

91.3
90.3
90.0
90.8
91.8
93.8
93.1
96.0
98.0
100.0
101.9
100.1
103.5
101.2
104.4
106.0
102.8
105.0
103.2

78.8
81.0
81.6
81.5
84.0
86.0
89.8
95.5
99.4
100.0
105.6
106.3
110.4
111.6
118.5
124.6
122.9
127.4
131.9

86.3
89.7
90.7
89.8
91.5
91.7
96.5
99.5
101.4
100.0
103.6
106.2
108.7
110.3
113.5
117.5
119.6
121.3
127.8




!

68

of single- or multi-unit establishments are influenced
by the menu offered, and therefore cannot be used to
indicate changes in labor productivity of specific em­
ployment size classes. However, efficiencies in the
use of capital, materials, and organizational inputs
undoubtedly have been greater in multi-unit than in
single-unit establishments, and this largely accounts
for the more rapid expansion of multi-unit busi­
nesses.
The changes in the structure of the food service
industry were marked by the expansion of fast-food
establishments. According to a Department of Com­
merce survey, there were 43,000 franchised eating
establishments in 1975, representing an estimated 20
percent of all eating and drinking places, and ac­
counting for 25 percent of industry sales.1 A study
3
by the U.S. Senate Select Committee on Small Busi­
ness shows that the number of fast food units nearly
tripled between 1960 and 1971, while the number of
restaurants, other than fast food, declined 9 percent
to 210,000.1 The expansion of fast-food establish­
4
ments has introduced profound systemic changes in
the food industry which lie at the root of recent and
future productivity improvements.
Fast-food operators introduced principles of in­
dustrial engineering in retail food services—includ­
ing work organization and layout—which had previ­
ously been applied mainly by large institutional and
industrial caterers or food contractors.1 These prin­
5
ciples have been implemented throughout numerous
franchised or company-owned outlets. According to
a survey by The Conference Board,1 all or the great
6
majority of fast-food franchisers participating in the
survey distributed operating manuals; operated man­
agement training programs; trained franchisee em­
ployees; selected sites; and designed facilities and
layout. Moreover, many services to franchisees were
rendered on a continuing basis, including counseling
through field personnel; training of new employees;
help with maintaining quality standards; and cen­
tralized purchasing. These organizational features
are more prevalent among company-owned fast-food
chains than among franchised establishments, and
represent key elements in standardizing managerial
practices.1
7

T®
bS© 2. Employment in food service occupations, 1972
and 1976
1972

1976

Occupation
Number
Restaurant, cafeteria, and bar managers . . .
Food service workers.................................
Bartenders.............................................
Waiters and assistants..........................
Cooks....................................................
Dishwashers...........................................
Counter and fountain workers.................
Other (except managerial).....................

Percent

Number

Percent

494
3,263
201
1,263
866
218
307
408

13.2
87.0
5,4
33.6
23.1
5.8
8.2
10.9

505
3,919
261
1,450
1,065
251
421
471

12.1
88.6
5.9
32.8
24.1
5.7
9.5
10.6

SOURCE: BIS Employment and Earnings. Comparable data for years prior to 1972 are not
available.

percent in 1976, was the lowest for any occupational
category reported by the Bureau of Labor Statistics
(except for private household workers). It compared
with 53 percent for all service-producing workers
outside of households, and 54 percent for all occupa­
tional groups. Women accounted for 64 percent of all
workers in the industry, compared with 51 percent
for all services outside households, and 44 percent
for all occupational groups. Women are generally
more likely than men to hold part-time jobs in the
industry.
Furthermore, the average age of the work force
has declined over the past 15 years, suggesting a
decline in the proportion of seasoned, experienced
workers. In 1975, teenagers accounted for 30 percent
of all food service workers, compared with 17 per­
cent for all service industries, and 8 percent for total
nonagricultural payroll employment. Between 1960
and 1970, the median age of food service workers
declined from 42 to 33 years. For the labor force as
a whole, it remained constant at 40 years.
Growth in multi-unit firms
The eating and drinking place industry changed
considerably during the 1958-76 period. The num­
ber of establishments dropped 4 percent, to 359,500,
between 1958 and 1972 (the most recent year for
which data are available). All of the decline occurred
in drinking establishments serving alcoholic bever­
ages. While the number of drinking places dropped
7 percent, eating places rose 10 percent with nearly
all of the rise in multi-unit operations, usually run by
erne firm. Multi-unit establishments almost doubled
ewer the 14-year period; single-unit establishments
grew by less than 2 percent; and owner-operator
units without paid employees dropped by one-third.
No comparable changes occurred for drinking
places, virtually all of which were owner-operated in
both 1958 and 1972.
The impact of these changes on the industry’s
labor productivity cannot be demonstrated. Eating
place sales per employee rose during the period, but
variations from the average by employment size class



Lslbor-savifig Innovations
Productivity gains in the food service industry
have been associated with three kinds of technologi­
cal advances: (1) the off-premise preparation of foods
which permits reduction in on-premise preparation
time and employee-hours, (2) the simplification of
work processess through improvements in materials
handling and cooking devices, and (3) innovations in
food preservation methods and equipment. Food serv­
ice establishments have not adopted these technolo­
gies to the same extent; many of the higher priced
69

expenditures suggests that diffusion of innovated
food service equipment has been rapid for corporate
establishments but slow for others. Overall, capital
expenditures rose 31 percent between 1968 and 1972,
but much less in constant dollars—the same rate of
advance as for the plant and equipment outlays of
U.S. business as a whole. Corporate food service
businesses, however, raised capital expenditures by
67 percent over the period; proprietary firms and
partnerships, nartly because of the decline in their
number, lowered capital spending by 15 percent.
Hours of all persons in the industry rose 7 percent
between 1968 and 1972.2
2
The major improvement in food service equipment
has been the microwave oven. The heat generated in
microwave ovens is distributed uniformly through­
out the product being cooked (rather than conducted
from its surface inward, as in conventional ovens).
Moreover, all the energy produced is absorbed by the
product, rather than by the oven walls and the sur­
rounding air.2 Hence, processing time is greatly re­
3
duced, although microwave ovens are often supple­
mented by auxiliary equipment so that an acceptable
product texture and surface color is obtained.
Forced convection ovens have been rapidly
adopted in the industry. These ovens are reported to
reduce cooking time up to 50 percent by using a
recirculating loop with a built-in fan to reheat the air
within the cooking chamber, thereby increasing the
rate of heat transfer to the product. Forced convec­
tion ovens are being installed in most new operations,
and are replacing free-convection ovens in many ex­
isting facilities.2
4
Fat fryers have been refined for more convenient
operation and better product quality. Processing
control has been improved by more accurate timers
and thermostats, and by automatic basket lifts which
terminate cooking after a preassigned period. Pres­
sure containers, which increase the heat to the prod­
uct and thus speed up processing time, have been
introduced.2
5
Gas burning broilers are still widely used, but
commercial installations are beginning to use infra­
red heat to generate high temperatures and shorten
cooking time. Operations producing large volumes of
processed foods are increasingly using continuousflow broilers which require only unskilled labor once
the temperature and speed of the transfer belt have
been set.2
6

restaurants, for example, capitalize on the culinary
skills of their staff, and use off-premise prepared (or
convenience) foods on only a limited scale.1 The
8
numerous single-unit small diners and refreshment
places that characterize much of the industry are
often slow to modernize their equipment, or unable
to do so altogether. The trend, however, is in the
direction of shortened food preparation time and
higher ratios of equipment to employment.
Food preparation. According to a 1974 survey, 70
percent of all respondents used fresh frozen meats
and 56 percent used meats prepared to some extent
off premises (for example, pre-cut to meet portion
standards). Seafood, fresh frozen or otherwise par­
tially prepared, was served by more than 60 percent
of all respondents, and fruits and vegetables prepared
fully or partially off premises were offered by 40
percent. A significant proportion of respondents also
served baked goods prepared off premise. The great
majority using frozen or other partially prepared
foods served them regularly, and not merely as sup­
plements to conventionally prepared foods.1 How­
9
ever, much of the food served is still prepared on the
premises. For example, roughly 40 percent of all
standard meat dishes—that is, fried chicken, meat
balls, roast beef, steaks—are still prepared by restau­
rant. staff.
Food service establishments have been substitut­
ing off-premise for on-premise prepared foods in
order to reduce labor costs and to control the por­
tions served. Almost three-fifths of the respondents
in a 1972 survey gave these two reasons for serving
convenience foods.2 Other respondents cited the
0
broadened menu, as well as reduction in costs per
portion made possible by convenience foods. (Al­
though a substantial proportion of off-premise pre­
pared food originates in central kitchens or commis­
saries classified in the industry, some originates in
food processing industries. See the appendix for a
discussion of the effects on the productivity meas­
ure.)
The use of foods prepared off premise facilitates
large-scale operations. Units with 25 employees or
more are more likely to offer such foods than smaller
firms. Frozen entrees, for example, were served by
more than two-fifths of the larger establishments,
compared with one-third or fewer of the smaller
ones. Frozen baked goods and vegetables showed the
same pattern.2 Also, the use of such foods has im­
1
proved the uniformity of food quality, saved on in­
vestment in inventory, and has enabled the industry
to reduce the level of needed culinary skills—partly
in response to the shortage of qualified cooks and
chefs..

Food preservation. Important developments have oc­
curred in the quick freezing of fresh foods and in the
efficient thawing of frozen foods. Minor but signifi­
cant changes have also been taking place in other
phases of food preservation. Whether or not food
service establishments operate their own food proc­

Food processing. The trend in the industry’s capital



70

essing and preservation equipment, these changes
tend to reduce on-premise preparation time, improve
ths quality, and expand the variety of foods served.
The development of thawing equipment has been
spurred by concern with the nutritional and chemi­
cal deterioration of foods allowed to unfreeze for
long periods and by the larger size of frozen food
packages used in the industry. Microwave thawing
systems can temper frozen foods in a few minutes.2
7
Thus, reduction in on-premise preparation time is
sustained when efficient thawing systems are used.
Changes in food preservation methods, other than
freezing, have been modest in their impact on the
food service industry. Dehydration, the most widely
used preservation method, underwent no significant
evolution during the period (except for freeze-drying
of coffee).2 In canning, however, the aseptic process
8
was introduced and spread rapidly. The process,
which involves packing sterilized food in sterilized
containers in a sterile environment, eliminates the
change in flavor, texture, and appearance that usu­
ally results from thermal treatment of products for
canning.2 While the use of all canned foods cuts the
9
time spent in on-premise preparation (in comparison
with cooking from the raw), the introduction of aseptic ally canned foods enhances the acceptability and
extends the variety of foods.
Some improvements ahead
Productivity in the food service industry should
continue to improve. The adoption of labor-saving
equipment and off-premise prepared foods is likely to
be spurred by the expansion of corporate establish­
ments with their focus on efficient management. The
continued decline in the number of smaller marginal
firms, while perhaps a loss in terms of customer con­
venience, will nonetheless help raise industry pro­
ductivity.
Developments in food processing and preservation
technology will probably make for more widespread
oil-premise preparation of food, especially insofar as
such developments improve quality and help
broaden menu choices. Irradiation or radiation-pas­
teurization may become acceptable in preserving
foods high in moisture content and therefore liable to
rapid bacterial decomposition (for example, fish,
fruits, and vegetables); freeze-drying may spread to
products such as eggs; and aseptic canning is likely
to spread.
Completely integrated food service systems, with
precisely timed and mechanized transfer operations,

may increasingly mark the spread of contract feeding
(they are not likely to prove feasible in smaller retail
operations). In such systems, also called a “cooking
street,” two persons operate five pieces of equipment
—a steam cooker, a water cooker, a deep fat fryer,
a grill, and a broiler. All pieces are removable with­
out tools, and there is a minimum of complex mech­
anisms. These “cooking streets” result in a very high
ratio of meals served per employee;3 however,
0
menus are necessarily restricted and there is little if
any floor service.
Standard menus and simple equipment have, in
part, been dictated by persistent shortages of skilled
kitchen personnel, and the resulting need for equip­
ment that can be operated with minimum training by
unskilled persons of whom a high turnover rate is
expected. In addition, customer self-service has
spread, to some extent, to full-service restaurants
with buffet offering. In fast food shops, customers
often accept the job of clearing their tables. Such
self-service tends to reduce the industry’s reliance on
low-skilled labor.
Over the long term, the supply of low-skilled
workers is expected to contract, assuming full or
near full employment is attained. Based on that as­
sumption, recent projections indicate a relatively
small rise in the number of low-skilled or unskilled
workers to the end of the decade; and a decline in the
first half of the 1980’s.3 Incipient labor shortages
1
would compel the industry to upgrade its work force
and to develop career progression systems.3 At the
2
same time, the industry will very likely continue to
substitute capital for labor, possibly at a stepped-up
rate.
Output growth in the food service industry hinges,
of course, on continued gains in real family and per
capita income. The expansion in the proportion of
working wives should continue to raise the demand
for food eaten away from home.
Some productivity advances may arise from cer­
tain changes in patterns of eating out. The traditional
concept of three meals a day—tending to bunch
labor inputs at peak periods—has in part given way
to and in part been supplemented by a greater fre­
quency of consuming snacks or “mini-meals.” To the
extent this pattern prevails, more efficient utilization
of labor and capital would be attained, but food out­
lets would have to operate longer hours and therefore
would have to generate higher output volume to en­
sure productivity gains.

1 Eating and drinking establishments include restaurants, lunch coun­
ters, refreshments stands, cafeterias, and other facilities selling food or

drink (including alcoholic beverages) for on-premise consumption. Eat­
ing facilities in department stores, hotels, and motels are excluded, unless




7!

leased to outside operators. The industry is classified as Eating and
Drinking Places (code 58) in the Office of Management and Budget’s

1 For some applications of industrial engineering to food services, see
5
Raymond Pedderson et al, I n c r e a s in g P r o d u c tiv ity in F o o d S e r v ic e (Chi­
cago, Institutions/V olum e Feeding Management, 1973), 206 pp. E cono­
mies through centralized purchasing appear in large measure to have
been achieved through distributors in the wholesale industries. See
Charles Sirey, Jr. “Food Service Logistics: Roadsigns in the W ilderness,”
I n s t itu tio n s /V o lu m e F e e d in g , December 1970, beginning on page 53.

1 9 7 2 S t a n d a r d I n d u s tr i a l C la s s ific a tio n M a n u a l.
2 The average annual rates of change are based on the linear least
squares trend o f the logarithms of the index number. Extension of the
indexes will appear in the annual BLS bulletin, P r o d u c tiv ity I n d e x e s f o r
S e le c te d I n d u s trie s .

1 E. Patrick McGuire, F r a n c h is e d D is tr ib u tio n (N ew York, The Con­
6
ference Board, 1971).

3 Corinne LeBovit, “The Changing Pattern of Eating O ut,” N a tio n a l
F o o d S itu a tio n , No. 144, May 1973, p. 31. Data for recent years were

derived from data com piled by the U.S. Department o f Commerce.

1 The M cD onald’s operations manual is a 385-page book detailed
7
down to the most minute facets of running the stand and its machffiery.
See Charles G. Burck, “Franchising’s Troubled Dream W orld,” F o rtu n e ,
March 1970, p. 116.

4 C o n s u m e r E x p e n d itu r e S u r v e y S eries: D ia r y D a ta 1 9 7 2 , Report
448-1 (Bureau o f Labor Statistics, 1975).
! M ost o f the 15-year real incom e gain of primary' individuals occurred
from 1965 forward— 33 percent, compared with 16 percent for real
family incomes. Gains in the average income of primary individuals are
in part attributable to the rise in the average social security benefit which,
for w idows and widowers, more than tripled between 1960 and 1975. The
total number eligible more than doubled. All but $6 of the $124 monthly
increase in the average benefit occurred from 1965 forward.

1 Marshall C. Warfel, “Convenience Foods— What is the Score?” T h e
8
C o r n e ll H .R .A . Q u a r te r ly , May 1971, beginning on page 33.

1 See I n s t itu tio n s /V o lu m e F ee d in g , December 1974.
9
2 Ibid., September 1972.
0
2 Ibid., December 1974.
1

6 S e le c te d F a m ily C h a r a c te r is tic s a n d A v e r a g e W e e k ly E x p e n d itu r e s b y
I n c o m e C la s s e s o f F a m ily I n c o m e B e fo r e T axes, Consumer Expenditure
Diary Survey (Bureau o f Labor Statistics, 1976).

2 The long-term growth rate of capital in the food service industry
2
(eating and drinking places) has been estimated at 3.4 percent annually
for 1929-63, nearly twice that for total retail trade. The growth o f capital
per em ployee-hour has been estimated at 0.6 percent annually for the
same period, which compares with 0.7 percent for total retailing. See
David Schwartzman, T h e D e c lin e o f S e r v ic e in R e t a i l T r a d e (Pullman,
W ashington, State University, 1971), p. 67.

7 Corinne LeBovit, “The Changing Pattern,” p. 30.
8 See I n s t itu tio n s /V o lu m e F e e d in g (Chicago, Cahners Publishing Co.),
April 1975.
9 See Theodore Levitt, “Production Line Approach to Service,” H a r ­
v a r d B u s in e s s R e v ie w , September-October 1972, beginning on page 41.

2 The conversion of electrical into radiation energy remains relatively
3
inefficient, although one manufacturer reportedly has claimed that his
brand of microwave oven converts 72 percent of electrical into radiation
energy. The average conversion range is estimated at 30 to 50 percent
for all brands. See Frank W. Schmidt and Stephen Bartlett, “Food
Processing and Preparation Equipment as It Shapes the Future of
Food Service,” in Thom as F. Powers, ed. T h e F u tu r e o f F o o d S e r vic e ,
p. 94.

A lso Thom as F. Powers, “Food Service in 1985,” T h e C o r n e ll H . R . A.
Q u a r te r ly , May 1976, p. 47.
1 Unpublished data on work experience of the population in 1962 and
0
1975, Bureau o f Labor Statistics.
1 “The food services have been drastically reducing the number of
1
skilled workers by simplifying the operation so that it can be run by
quickly trained, low-paid, high-turnover em ployees.” See Daniel M.
Seifer, “The Service Industries: Automation, Minimum Wages, Unem ­
ployment,” B u lle tin o f B u s in e ss R e s e a rc h , The Ohio State University,
Vol. 46, No. 8; G. E. Livingston, “Changes in the Food Service Indus­
try,” T h e C o r n e ll H . R . A. Q u a r te r ly , May 1974, p. 15; and “Young
Women Who Work, An Interview With Myra W olfgang,” Irving Howe,
ed., T h e W o r ld o f T h e B lu e - C o lla r W o r k e r (New York, Dissent Publish­
ing Co., 1972), p. 26.

2 Ibid., p. 93.
4
2 Ibid., p. 100.
5
2 Ibid., p. 104.
6
2 Ibid., pp. 89-90.
7
2 Ibid., p. 149.
8
2 Ibid., p. 191.
9
3 See the section on com pletely integrated systems in “Food Process­
0
ing and Preparation Equipment,” T h e F u tu r e o f F o o d S e rvic e , beginning
on page 109. A lso see “Health Services,” T e c h n o lo g ic a l C h a n g e a n d
M a n p o w e r T r e n d s in S ix I n d u s tr ie s , Bulletin 1817 (Bureau o f Labor
Statistics 1974) p. 58.

1 See Thomas F. Powers. “Labor Supply, Payroll Costs and
2
Changes.” T h e C o r n e ll H . R . A . Q u a r te r ly , May 1974, beginning on page
5.
1 Andrew Kostecka. F ra n c h isin g in th e E c o n o m y , 1974-76 (W ashing­
1
ton, D.C., U.S. Department of Commerce, 1976), p. 7. See also Philip
B. Dwoskin, “Fast Food Franchises: Market potentials for agriculture
products in foreign and domestic markets,” M a r k e tin g a n d T r a n s p o r ta ­
tio n S itu a tio n , No. 196, February 1975, beginning on page 20.
1
4

3 See Harold Wool, “Future labor supply for lower level occupa­
1
tions,” M o n th ly L a b o r R e v ie w , March 1976, pp. 27-28.

Quoted in Thomas F. Powers, ed., T h e F u tu r e o f F o o d S e r v ic e : A

B a s is f o r P la n n in g (University Park, Pa., The Pennsylvania State Univer­

sity, Food Service and Housing Administration, 1974), pp. 35 and
41.




72

3 The Employment and Training Administration o f the U.S. Depart­
2
ment of Labor has sponsored several studies on em ployment and career
progression in the food service industry. For example, see Gary L.
Hotchkin, D e v e lo p m e n t o f C a r e e r P ro g ressio n S y s te m s f o r E m p lo y e e s in
th e F o o d S e r v ic e I n d u s tr y (Chicago, National Restaurant Association,
1975).

in farm machinery manufacturin
Productivity gains, aided by new technology,
especially computers, but moderated
by cyclical downturns, averaged 2.6 percent
a year over the 1958-80 period
A

rthur

S.

H

erm an a n d

Jo h n

W.

F e r r is

Productivity, as measured by output per employee hour,
in farm machinery manufacturing1 was about the same
as the average for all manufacturing industries over the
1958-80 period. Growth was aided by numerically con­
trolled machine tools, automatic welding, computerized
ma nufacturing, industrial robots, and computerized au­
tomatic warehouses, but was partially offset by sharp
declines in demand. Almost every decline in productivi­
ty during the period studied can be associated with a
drop in output, which, in turn, usually coincides with
downturns in the economy. During the 22-year period,
productivity in the farm machinery industry grew at a
rate of 2.6 percent a year, compared with 2.7 percent
per year for all manufacturing industries; 1.9 percent for
construction machinery, an industry which uses similar
manufacturing techniques; and 3.2 percent for motor
vehicles, another similar industry.

Output, productivity follow farm income
Productivity growth in the farm machinery industry
can be divided into three distinct periods. From 195865, productivity grew at an annual rate of 1.7 percent;
from 1965-74, it accelerated to a 3.3-percent rate; and
from 1974-80, slowed to 0.2 percent. (See table 1.) The
higher rate of gain during the 1965-74 period can be as­
sociated with years of very high output, fueled by dra­
matic increases in farm income.
Arthur S. Herman is an econom ist and John W. Ferris is a statistician
in the Division of Industry Productivity Studies, Bureau of Labor Sta­
tistics.

Reprinted from the
M o n t h l y L a b o r R e v i e w , O ctober 1982.




Productivity changes in the farm machinery industry
are closely tied to output changes over the short term.
Demand for farm machinery is based on a number of
interrelated factors. A major factor is the overall state
of the economy. However, an even more directly related
factor is farm income. Changes in the output of farm
machinery closely parallel changes in farm income.
When farm income is up, farmers tend to purchase new
equipment. Among the determinants of income are crop
size, both actual and anticipated in the near future, and
farm prices. Crop size is, of course, affected by a num­
ber of variables, including the weather, farm prices,
government policies, and the worldwide food supply.
Other important factors affecting the production of
farm machinery are farmers’ costs, such as for loans,
new machinery, land, fertilizers, and pesticides, as well
as age and condition of existing equipment and imports
and exports of farm equipment.
When income is low and prospects appear poor,
farmers tend to make do by repairing, rather than re­
placing, existing equipment. Conversely, when income is
growing and prospects for further expansion of profits
appear good, they tend to purchase new, more produc­
tive equipment. Demand for machinery increases signifi­
cantly during these expansive periods, as does produc­
tivity.
The impact of the numerous variables affecting
demand changes rapidly over time; therefore, output of
farm machinery shows wide swings. Productivity, how­
ever, moves in a less volatile manner. For example, out­
put grew by 6.3 percent between 1958 and 1959, but

then dropped precipitously in 1960, a recession year,
falling 18.3 percent. Concomitantly, productivity had
no growth in 1959 and dropped sharply, by 7.1 percent,
in 1960. In 1966, output increased substantially, up 19.4
percent, then declined for 4 consecutive years, one of
which was the recession year of 1970. Following output,
productivity also grew substantially in 1966, up 6.2 per­
cent, and then dropped sharply, averaging 0.8 percent
from 1967 to 1970.
The early 1970’s were a period of high output
growth, with gains of 16.5 percent in 1972, 21.3 percent
in 1973, and 14.3 percent in 1974. This strong growth
can be attributed to a sharp increase in farm income re­
sulting, in part, from large exports of farm products, in­
cluding sales of grain to Russia. Productivity recorded
its largest advances during this period, with increases of
8.9 percent in 1971, 9.3 percent in 1972, 5.2 percent in
1973, and 3.6 percent in 1974.
In the more recent period— 1980, a recession year—
output dropped 15.1 percent, as farm income declined
precipitously. In turn, productivity declined 6.7 percent.
A factor affecting output over the long term is the
continuously increasing size of farms. The average farm
in the United States has shown a significant increase in
size, growing about 40 percent in acreage over the peri­
od studied.2 This created a need for an increase in the
physical dimensions and horsepower of farm machinery.
To cope with the growing acreage, farmers purchased
larger, more powerful equipment, rather than increasing
their labor force. For example, the average horsepower
( p t o ) rating of tractors was 106 in 1980, compared with
67 in 1958. Demand for farm equipment has also been
enhanced by such equipment as 4-wheel drive tractors,
which allow farming in previously marginal areas, and
such amenities as air conditioning and stereo radio and
cassette equipment in the cabs of the larger units.
Demand for larger, more productive farm machinery
has been one factor leading to the industry’s long-term
growth rate in output of 4.2 percent, somewhat higher
than the 3.8 percent for the total manufacturing sector.
Highly advanced farm equipment is one of many rea­
sons that productivity has been significantly higher in
the farm sector than in the nonfarm sector.

Table 1. Output per employee hour and related indexes
in the farm machinery equipment industry, 1958-80
[1977=100]
Output per hour
Year

1958 . ..
1959 . ..
1960 . ..

65.1
65.1
60.5

64.9
63.4
61.3

65.5
70.3
58.6

49.4
52.5
42.9

75.9
80.7
70.9

76.1
82.8
70.0

75.4
74.7
73.2

1961
1962
1963
1964
1965

...
...
. ..
...
. ..

62.9
65.1
66.6
70.2
72.2

61.3
65.1
64.3
66.9
68.6

67.7
64.8
74.3
82.0
84.8

45.7
48.8
53.7
60.1
64.0

72.7
75.0
80.6
85.6
88.6

74.5
75.0
83.5
89.9
93.3

67.5
75.3
72.3
73.3
75.5

1966
1967
1968
1969
1970

...
. ..
...
...
...

76.7
76.8
76.7
73.8
75.7

72.3
73.3
75.0
73.2
75.2

92.7
88.8
82.1
75.9
77.3

76.4
73.6
70.8
65.8
65.1

99.6
95.8
92.3
89.1
86.0

105.6
100.4
94.4
89.9
86.6

82.4
82.9
86.2
86.7
84.2

1971
1972
1973
1974
1975

.. .
...
...
...
...

82.4
90.1
94.8
98.2
97.7

83.0
87.0
90.7
92.6
95.3

81.0
99.9
109.2
118.3
105.2

66.2
77.1
93.5
106.9
100.0

80.3
85.6
98.6
108.9
102.4

79.8
88.6
103.1
115.4
104.9

81.7
77.2
85.6
90.4
95.1

1976
1977
1978
1979
1980

...
.. .
.. .
.. .
...

101.1
100.0
100.8
103.2
96.3

100.5
100.0
100.1
101.7
99.6

103.1
100.0
103.1
108.0
88.1

98.9
100.0
95.6
114.7
97.4

97.8
100.0
94.8
111.1
101.1

98.4
100.0
95.5
112.8
97.8

95.9
100.0
92.7
106.2
110.6

Average annual rates of change (percent)1

1958-80
1958-65
1965-74
1974-80

2.6
1.7
3.3
0.2

2.7
1.0
3.4
1.2

2.4
3.9
2.9
-2 .9

4.2
3.9
3.7
-0.1

1.5
2.2
0.5
-0.3

1.5
2.9
0.4
-1.4

1.8
( 2)
0.8
2.9

' Based on the least squares trend of the logarithms of the index numbers.
•2Rate of change is less than 0.05 percent.

employees per establishment has remained fairly con­
stant, dropping slightly from 74 in 1958 to 70 in 1977
(the average for all manufacturing industries was 53).
The industry has a few very large firms with numer­
ous establishments making a variety of equipment—
tractors, combines, and other harvesting equipment,
crop sprayers, plows, harrows, planters, cultivators, hay
balers, and fertilizing equipment. These firms are highly
integrated and manufacture many of the parts that are
assembled into the final products, including both gaso­
line and diesel engines, as well as replacement parts for
the older units in operation. The large firms generally
produce the larger equipment, such as grain harvesting
combines, 4-wheel drive tractors, and accessories. There
are numerous medium and small firms in the industry.
They usually specialize in a particular line or type of
equipment, such as milking, poultry, or irrigation equip­
ment. Many of them serve local markets for highly spe­
cialized equipment. The smaller firms also make lawn
and garden equipment, such as walk-behind lawnmowers and snowblowers.
Farm machinery manufacturers are concentrated in
the Farm Belt, with most plants in mid western States—
Illinois, Wisconsin, Minnesota, Iowa, Nebraska, and

Plants located in Farm Belt
The farm machinery manufacturing industry has
paralleled the growth of agriculture in the United
States. Some of the larger firms can trace their origins
to the development of horse drawn harvesting equip­
ment in the early 1800’s. Therefore, farm machinery
manufacturing is a mature industry, producing a variety
of equipment for both U.S. markets and export.
There were 2,148 establishments in the farm machin­
ery industry as of 1977, a significant increase over the
1,949 establishments reported in 1958. The number of




Employee hours

Production Nonpro­
All
Production Nonpro­ Output
All
employees workers duction
employees workers duction
workers
workers

74

an assembly operation which uses welding and fastening
with air powered tools. Farm machinery is usually fin­
ished by painting, either in the parts stage or as a com­
pleted unit.
Because of the complex nature of many of the
products, the varied manufacturing operations involved
in producing units, and the fact that farm machinery
manufacturing is a mature industry with many old
plants, there are numerous areas that are subject to
technological change. The larger companies usually
make most of the parts they assemble into the final
product. Therefore, the technological innovations they
employ cover a range of manufacturing operations and
have resulted in significant labor savings.
During the 1960’s, capital expenditures per employee
for new plant and equipment were consistently below
the average for all manufacturing industries. However,
because of sustained demand for farm equipment in the
early 1970’s which strained the industry’s capacity,5
firms began to increase their capital expenditures for
new plant and equipment. By 1975, capital expenditures
per employee had almost tripled, compared to the level
in 1970. This resulted in the installation of advanced
manufacturing equipment and large scale plant modern­
ization and probably was one of the factors leading to a
higher rate of productivity increase during the 1970’s
than during the earlier decade.
Computers are among the widespread innovations
with significant impact upon the industry. They are
used for many functions, including inventory control,
data collection, tracking progress of semi-completed
products, design, and for numerous accounting and oth­
er business purposes. In recent years, computers have
been more directly used for manufacturing operations
on the factory floor.
Numerically controlled machine tools are used exten­
sively by major companies in the manufacture of the
parts used in assembling farm machinery. A recent in­
novation is computerized numerically controlled ma­
chine tools, which are more versatile than standard
equipment because they can be programmed for chang­
es by the operator rather than from tapes. One unit in­
stalled in a large firm is a completely computercontrolled gear case transfer line, using numerically
controlled machine tools, where parts automatically go
through 87 machining operations.6
One plant is experimenting with a change in machine
tool layout, from the traditional setup consisting of
banks of individual machines designed for a single oper­
ation to cells of machine tools based on workflow. This
new layout requires high volume, but has cut bottle­
necks in production and has resulted in operating effi­
ciencies.
Automatic welding has replaced manual welding in a
number of installations. In addition, industrial robots

Kansas. Texas and California also have a large number
of plants.
The largest export market for U.S. manufacturers is
Canada. In turn, Canada provides the largest amount of
imports of farm machinery into the United States.

Employment and faoinrs rapidly adjusted
Over the 1958-80 period, the number and hours of
production workers and nonproduction workers in the
farm machinery industry have grown at similar rates.
Production workers increased at an average annual rate
of 1.7 percent and their hours grew 1.5 percent.
Nonproduction workers grew at rate of 1.7 percent, and
their hours increased at a rate of 1.8 percent.
Year-to-year changes in employment and hours in
this industry tend to move in a similar but less volatile
pattern than changes in output. This indicates that the
industry can adjust its hours and employment fairly
rapidly to changing demand. For example, when de­
mand is falling overtime usually is cut, the number of
shifts worked are reduced, the normal summer shut­
downs may be extended, and workers may be laid off.
The extent of the adjustments in hours due to chang­
es in demand is influenced by the occupational makeup
of the work force. In the farm machinery industry, the
largest occupational group is operatives, most of whom
are assemblers. Welders, precision machine operators,
punch and stamp machine operators, and transportation
operators also are important. These employees, along
with laborers (mainly freight handlers) are most affected
by reductions in demand. The industry also employs a
large group of craftworkers— machinists, mechanics,
tool and die makers, and blue-collar supervisors.3
Craftworkers are least affected by declines in produc­
tion; because of their skill levels, employers are reluc­
tant to lay them off for fear that they may not be
available when demand picks up.

Technology aids productivity
Technological change varies greatly among plants in
the farm machinery industry. The more advanced highly
sophisticated equipment is used, for the most part, by
larger firms engaged in mass production of various
products. Slower changes are undertaken by the smaller
firms which make short runs of highly specialized prod­
ucts and generally have limited capital.4
The level of complexity of farm machinery manufac­
turing differs greatly depending on the product, which
can range from a simple plow pulled by a tractor to a
complex self-propelled grain harvesting combine. How­
ever, there are factors common to most farm equipment
manufacturing: most of the components are made of
iron and steel; they are shaped by such processes as
casting, cutting, stamping, punching, boring, and ma­
chining; and they are joined to form the final product in



75

is run by a single operator. The assembly lines are set
up so that fasteners and other small parts are fed di­
rectly to the assemblers at the correct height for their
use. This plant’s design significantly cuts parts invento­
ry, reduces handling, increases manufacturing efficiency,
and results in overall labor savings.
Besides robotic painting, which is just being intro­
duced in the industry, there are a number of other inno­
vations that increase painting efficiency. One system,
electrostatic painting, has been used for a number of
years. In this process, electrically charged parts move
through an automatic paint spray booth, with the paint
mist attracted to the charged part. Another innovation
is electric dip paint lines, in which charged parts are
dipped into a paint-filled tank from which paint is pre­
cipitated out on the part. These systems have resulted
in savings in both paint and labor.
While the advanced innovations are most readily
adapted by the larger multiline companies, smaller firms
in the industry tend to introduce new technology more
slowly. Many of the latter specialize in a particular
product, such as pipeline milking units or self-propelled
irrigation systems. Although these units are usually pro­
duced from common components (pipes, tanks, spray
guns, and pumps), they are generally assembled to fit a
particular farmer’s need. Because of the semicustom na­
ture of production used by these smaller firms, it is dif­
ficult to adapt much of the available new technology
which is designed for volume production. In addition,
many of the smaller firms are located in rural areas near
the farms they serve and do not have the access to the
capital markets as do the major companies.

are being introduced for welding functions, resulting in
more versatile automatic welding operations.
Significant efforts have been made to increase efficien­
cy in materials handling and warehousing functions.
These functions are very important because of the nu­
merous parts that must be moved, the many operations
that must be carried out, and the large size of the facto­
ries involved in the manufacture of the more complex
farm machines. A number of plants have installed
computerized automatic warehousing and materials han­
dling systems. In one plant, such a system is used for
the materials receiving warehouse. The system is located
in a special high rise building attached to the single
story plant. Materials are shipped in using the plant’s
containers, logged on the computer, and moved auto­
matically to a preassigned location. When needed, they
are called for by the computer, which automatically
sends a remote controlled sideloader for them, and are
sent via conveyor to the location requesting them. This
warehouse is run by a single computer operator. The in­
stallation of this system resulted in substantial labor
savings, while doubling warehouse capacity, because the
previously used equipment required numerous forklift
operators.
Sideloaders are an important innovation in the indus­
try, even though they require operators. They are
narrower and higher than the conventional forklifts
which they replace, allowing for increased storage space
and versatility in the warehouse. Sideloaders are in­
creasingly being used in semi-automatic computerized
high rise warehousing systems installed in a number of
plants.
An example of the most advanced technology for as­
sembly line manufacture in the industry is a recently
built tractor plant designed specifically for computer
control.7 This plant is unique in that almost all phases
of its operations are computer controlled or directed.
The plant has high rise computerized automatic ware­
houses. The parts to be assembled are programmed to
move in the correct sequence to produce a finished trac­
tor via conveyor through the various assembly lines.
This is a major advance over the system where parts are
made in advance and stored until needed, boxes of parts
are moved to the assembly line via forklift trucks, and
assemblers pick the correct parts out of the boxes to as­
semble the final product. The new plant uses industrial
robots for welding and painting. The robotic painting
machines are programmed to move their spray guns to
paint the correct part of the tractor chassis as it moves
by on the conveyor line. This differs from conventional
automatic spray painting equipment, which uses fixed
spray guns, in that it more closely approximates a hu­
man spray painter. Almost all welds for the frame of
the tractor cabs made at this plant are done on an
electronically controlled automatic framing buck which



Future trends uncertain
Changes in output and productivity in the farm ma­
chinery industry are expected to continue to reflect
changes in farm income. In the near future, the outlook
for farm income is uncertain. It has been falling since
1979; and currently, there are pressures on farm prices
that are expected to slash farm profits. In addition, such
factors as high interest rates and high fertilizer and pes­
ticide costs are also expected to reduce farm income.
The export market is uncertain, and farm prices are
down. This situation could result in a continuation of
the recent negative pressure on demand for farm ma­
chinery. In addition, technological changes in the near
future may be affected by the financial difficulties of a
number of the major companies in the industry, which
are expected to limit capital expenditures for new plant
and equipment.
Over the long term, modernization of plant and
equipment is expected to continue in the farm machin­
ery industry, with particular emphasis on labor savings
and cost reduction. These changes will be fueled by
possible competition with Japan in the market for larger
76

and increasing use of industrial robots for welding,
painting, and other high volume, difficult operations.
Computers will increasingly be used for manufacturing
operations and in design functions.

farm equipment, which is presently dominated by U.S.
concerns. Japan currently holds a large share of the
U.S. market for small tractors.8 The future will see
growing installation of automatic welding equipment

2 S ta tis tic a l A b s tr a c t o f th e U n ite d S ta tes , 1 9 8 0 (U.S. Department of
1
Average annual rates of change are based on the linear least
Commerce, 1980), p. 686.
squares trends of the logarithms of the index numbers. The farm ma­
chinery and equipment industry is- designated industry 352 in the
3 1 9 7 0 C e n su s o f P o p u la tio n , O c cu p a tio n b y I n d u s tr y , Vol. PC(2)-7C
S ta n d a r d I n d u s tr ia l C la ss ific a tio n M a n u a l, 1 9 7 2 E d itio n , issued by the
(U.S. Department of Commerce, 1972), pp. 281-88.
Office of Management and Budget. The industry comprises establish­
4 Based on discussions with industry experts.
ments primarily engaged in the manufacture of farm machinery and
5 U.S. I n d u s tr ia l O u tlo o k , 1 9 7 4 (U.S. Department of Commerce,
equipment, and garden tractors and lawn and garden equipment. A
1973), p. 301.
technical note describing the indexes is available from the Office of
6 J o h n D e e r e H a r v e s te r W o rk s (Deere and Company, 1979), p. 10.
Productivity and Technology, Bureau of Labor Statistics, Washington,
1 J o h n D e e r e T r a c to r W o rk s (Deere and Company, 1980), pp. 6-18.
D.C. 20212. The indexes for this industry will be updated and includ­
8 U.S. I n d u s tr ia l O u tlo o k , 1 9 8 1 (U.S. Department of Commerce,
ed in the Bureau of Labor Statistics’ annual bulletin, P r o d u c tiv ity
1980), p. 260.
M e a s u r e s f o r S e le c te d I n d u s trie s .




77

Productivity trends
for intercity bus carriers
During 1954-79, modest advances
in technology, and more package
and charter service, were offset
by declining passenger demand and
reduced bus speeds, resulting in
a 0,4-percent rise in productivity
R

ic h a r d

B. C a r n e s

Lower speeds have increased the labor time needed to
drive a given distance, and have reduced productivity.
However, lower speeds have also cut fuel costs. Al­
though total transportation travel might be expected to
decline because of higher fuel costs, the relative fuel effi­
ciency of buses enhance future demand for this mode of
transportation, especially for shorter distance travel.
Productivity movements were uneven over the 195479 period, ranging from a 9.4 percent increase in 1962
to a decline of 11.9 percent in 1975. Generally, these
changes have been in response to cyclical swings in in­
dustry output. There were three distinct trend periods.
During 1954-60, output per hour rose at a 1.2-percent
average annual rate. Output declined at an average
yearly rate of 1.3 percent and hours dropped more
sharply, by 2.6 percent. From 1960 to 1966, demand
for bus service increased 4.7 percent annually, but em­
ployee hours increased at only a 1.3 percent average an­
nual rate. The more efficient utilization of equipment
and facilities, which resulted from this higher demand,
raised productivity at a 3.6 percent annual rate during
those 6 years. Load factors and average length of haul
both increased appreciably. Load factor is the percent­
age of capacity actually utilized.
In the third period, 1966-79, all of the measures
turned down. Productivity and output fell at an annual
rate of 1.4 and 2.5 percent, respectively, while employee
hours dropped 1.1 percent. Output fell in all years ex-

During 1954-1979, output per employee-hour in the
class I bus industry rose an average of 0.4 percent a
year, a rate significantly below those of other segments
of the transportation industry.1Class I bus carriers pro­
vide intercity service and may also provide local or
charter service. Not included are those public and pri­
vate transit systems that provide urban mass transpor­
tation service and do not come under Interstate
Commerce Commission (icc) reporting requirements.*
2
The 0.4-percent growth in productivity resulted from
a small average annual increase in industry output of
0.1 percent combined with an average annual decline in
employee hours of 0.3 percent. (See table 1.) By com­
parison, other transportation industries for which mea­
sures are available showed productivity increases over
the same period that equaled or exceeded overall pro­
ductivity growth for the private nonfarm business sector
of the economy. For example, productivity in air trans­
portation, an industry which competes for public pas­
senger traffic, rose 6.3 percent, compared with 2.1
percent for the private nonfarm business sector. (See ta­
ble 2.)
Bus operations have suffered from the recent energy
shortages. Longer running times between cities have re­
sulted from the 55-mile-per-hour national speed limit.3
Richard B. Carnes is an econom ist in the Office of Productivity and
Technology, Bureau of Labor Statistics.

Reprinted from the
M o n t h l y L a b o r R e v i e w , M ay 1981.




78

cept 1967, 1974, and 1979. Since 1974, the beginning of
the energy crisis and the year of the 55-mile-per-hour
speed limit, productivity trends have been mixed, as ta­
ble 1 indicates. There were sharp rises in 1974 and
1977, and a small gain in 1979. These were offset by a
serious drop in 1975, and smaller declines in 1976 and
1978. More travelers rode buses in 1974 when fuel for
private passenger cars became scarce. But when gasoline
once again became plentiful in 1975, even at higher
prices, bus travel declined drastically. Again in 1979,
gas shortages in the second quarter helped boost indus­
try output by 6.1 percent for the year and productivity
by 0.4 percent.

Table 2. Productivity comparison, private nonfarm
business and selected transportation industries, 1954-79
Average annual rate of change
Industry

Private nonfarm business ..............
Transportation sector ................
Petroleum pipelines ’ ..............
A transportation1 ................
ir
Class 1railroads ....................
Intercity trucking 1 ..................
Class 1bus carriers1 ..............

P r o d u c t iv it y a n d r e la t e d in d e x e s f o r c la s s I b u s

[1967 - 100]
Output per
employee-hour

Output

Employeehours

.................................
.................................
.................................
.................................
.................................

77.4
80.4
81.2
81.6
81.9

80.5
79.0
78.0
78.3
74.0

104.0
98.3
96.1
96.0
90.3

1959 .................................
1960 .................................
1961.................................
1962 .................................
1963 .................................

84.6
83.7
85.3
93.3
94.6

74.0
75.4
77.1
86.2
86.6

87.5
90.1
90.4
92.4
91.5

1964 .................................
1965 .................................
1966 .................................
1967..................................
1968 .................................

95.7
101.2
103.4
100.0
98.6

90.2
95.0
99.2
100.0
97.5

94.3
93.9
95.9
100.0
98.9

1969 .................................
1970 .................................
1971.................................
1972 .................................
1973 .................................

95.7
93.4
91.3
93.0
92.5

94.2
92.5
86.9
83.3
79.8

98.4
99.0
95.2
89.6
86.3

1974 .................................
1975 .................................
1976 .................................
1977 .................................
1978 .................................
1979' ...............................

95.9
84.5
81.7
87.1
86.8
87.2

86.5
78.0
75.2
74.7
73.7
78.2

90.2
92.3
92.1
85.8
84.9
89.7

1954
1955
1956
1957
1956

:

]

|

Average annual rates of change

1954-79 ..........................
1974-79 ..........................

0.4

- .9

0.1
-1.9

-0.3
-1.0

1Prelim
inary.




1.9
2.7
7.5
6.3
4.9
2.4
.4

3.7
2.9
5.6
11.0
1.2
5.6
.1

1.7
.2
-1.8
4.5
-3.5
3.1
- .3

bile travel represents the primary source of competition
to the industry, followed by air and then train service.
Expansion of charter bus and package express service
has helped to offset passenger declines. (See table 3.)
Intercity bus operations have the potential to provide
service over a wide area because of the national high­
way network. Nonstop intercity buses can operate at
speeds similar to those of an autombile. And, over
shorter distances buses generally provide lower cost ser­
vice than air or rail travel.4 Most demand comes from
short-haul passengers even though the average length of
trip for intercity service has more than doubled from 62
miles in 1954 to 130 miles in 1979.5
When intercity bus service began in the early 1900’s
it was characterized by a large number of local and re­
gional carriers. Startup costs were modest and there
was rapid growth. By the 1930’s, the industry had
evolved into its present form, with fewer bus companies
and with national systems operating over longer dis­
tances. These national networks were thought to facili­
tate through-service for passengers and improve bus
and terminal utilization. During World War II, industry
o u tp u t in c re a se d ra p id ly d u e to ra tio n in g of a u to p a rts
and gasoline. Load factors during this period reached
nearly 80 percent. Passenger-miles peaked in 1952 and
did not reach that level again until 1967. Since 1954,
few new intercity bus carrier operations have been au­
thorized by the ICC. Presently, Greyhound and Trailways dominate the market.6
The bus industry is subject to both Federal and State
regulation. There are restrictions on the entry of new
firms, fares, route requirements, and service levels.
Competition along joutes is limited. Federal regulation
has encouraged merger activity of carriers into larger
national companies. Recently there has been an effort
on the part of the ICC to liberalize entry controls and to
provide greater carrier rate making autonomy. General
deregulation of the industry, however, has not been for­
mally introduced.
The sources of revenue for bus carriers have changed
substantially since 1954 as table 3 indicates. Intercity
and local passenger revenue has declined in relative

The class I regulated bus industry comprises 43
intercity and 13 local carriers certified by the ICC. In
1978, these companies operated about 9,700 buses and
had 34,000 employees. During that year, they moved
237 million passengers, and generated $961 million in
passenger revenue and $175 million in freight revenue.
For most of the 15,000 communities served by
intercity bus carriers, there is no other form of public
transportation. Despite this, the bus passenger market
has declined during the period of this study. Automo­

Year

Output

]

Employeehours

1O
utput per employee.

Industry profile

T a b l e 1.
c a r r ie r s

Output per
employee-hour

79

terms while charter and package express services have
shown significant growth. Charter service has expanded
due to the increase in group travel and tourism, while
package express service has benefited from the large dis­
tribution network provided by intercity buses.
The private automobile has been a major factor in
the slow growth of intercity bus travel. The doubling of
new car registrations since 1955 and the use of these
cars for both personal and business trips impacted bus
travel, and is expected to be the primary source of bus
industry competition in the foreseeable future. Autos
accounted for 89 percent of all intercity passenger-miles
in 1954, and for 83 percent in 1979. Passenger-miles
flown during this period increased their relative share of
the market from 3 to 15 percent while both bus and rail
passenger-miles declined.7

of drivers in the industry, by increasing the proportion of
administrative and service workers. Since 1954, workers
paid on a daily basis, mostly supervisory personnel, as
opposed to hourly wage employees, have increased from
8 percent of total employment to 10 percent. In the
intercity portion of the regulated bus industry, women
represent 12 percent of the work force, up from 10 per­
cent in 1960. By contrast, women make up 40 percent of
the work force in the total private nonfarm sector.
Changes in technology associated with the bus indus­
try have been characterized by a gradual trend toward
innovation, fuel efficiency, and greater passenger com­
fort. Diesel-powered buses, in primary use since the ear­
ly 1950’s, have undergone steady advances in
performance and reductions in maintenance require­
ments. Current-model intercity buses have a seating ca­
pacity of 47 passengers and have space for large
amounts of baggage and cargo. Typically, buses are 8
feet wide and 40 feet long, and weigh 13 tons. Including
resale after use by class I carriers, useful bus life is over
20 years and mileage may exceed 3 million.8 The aver­
age number of seats for the bus fleet in 1955 was 39.1
and increased 10 percent to 43.1 by 1978. However, the
seating capacity utilized during this period has remained
at about 47 percent, and load factors have changed lit­
tle since 1954, which helps explain the low rise in pro­
ductivity in the industry.
From 1950 to 1973, average bus speeds increased
from 50 to 60 miles per hour because of improved high­
ways and urban beltways. But the introduction of the
national speed limit in 1974 reduced average speeds to
less than 55 miles per hour,9 and has also slowed pro­

Employment and influences on productivity
Employment in the class I regulated bus industry de­
clined from 39,000 in 1954 to an estimated 35,300 in
1979. Employment dropped steadily in the 1950’s, then
advanced irregularly through 1967, and thereafter gen­
erally declined again to the present level. Recent excep­
tions to the downward trend were in 1974-75 and again
in 1979. Energy shortages resulting from the Organiza­
tion of Petroleum Exporting Countries oil embargo
boosted both employment and passenger service in
1974, the year that also marked the introduction of the
55-mile-per-hour national speed limit. Employment
needs increased partially as a result of the decline in the
number of bus miles per driver. Again in 1979, fuel
shortages reversed the downward trends in both em­
ployment and p a s s e n g e r service.
Since 1954, there has been a change in the composi­
tion of employment. The number of equipment mainte­
nance and garage personnel has declined from 22 to 17
percent of the work force because of reduced service re­
quirements. Station workers, however, have increased
from 11 to 19 percent of total employment, reflecting
the greater demand for package express traffic. Drivers
have accounted for about half of industry employment
since 1954. However, more fully utilized and larger ca­
pacity buses may, in the future, reduce the percentage

d u c tiv ity g ro w th .

The growth in package express and charter services,
however, has aided productivity. Delivering package ex­
press while engaging in regularly scheduled passenger
service has resulted in more efficient use of vehicle and
driver time. Charter services have also offered signifi­
cant economies of scale for bus companies. Charters
typically have a 50»percent greater load factor and
100-percent longer average trip length than regular
route carriers. This form of passenger service also pro­
vides economies in baggage handling, ticketing, and
scheduling terminal facilities.
Reduced investment has hurt industry productivity.
Since 1954, investment in plant and equipment by
intercity bus carriers has declined. Buses, which present­
ly cost about $135,000 each, account for about 80 per­
cent of industry capital expenditures. Annual constant
dollar investment dropped from $78 million in 1954 to
$56 million in 1974, the latest year for which data are
available. Similarly, the constant dollar stock of plant
and equipment fell 18 percent, while capital investment
per worker declined more than 20 percent. In contrast,
gross constant dollar investment in the transportation

Table 3. Revenue distribution for class I bus carriers and
percent of total service, 19S4 and 1978
1054
Service

1978

Revenue in
millions

Percent

Revenue in
millions

Percent

Total..................

$467

100

$1137

100

Passenger:
Intercity ....................
Locai.......................
C
harter ....................

306
112
33

66
7

678
73
211

60
6
19

Freight.........................

16

3

175

15




24

80

Factors are emerging which are both favorable and
unfavorable to demand and productivity growth in the
bus industry. Energy and demographic variables are
likely to be positive factors while negative public image
and low capital investment may retard growth.
Restructuring the industry has been suggested as a way
to increase capacity utilization and spur productivity.
With current low rates of bus utilization, increased
demand would likely result in higher load factors and
enhance productivity. Several projections of growth in
the bus industry for the next decade have been made.
The Federal Energy Administration (now part of the
Department of Energy) estimates a 25 percent growth
in passenger-miles over the next decade. This projection
is not altered substantially even when based on different
fuel availability assumptions. The Department of Trans­
portation ( d o t ) makes a similar growth projection but
notes the negative effect of rising income levels and shift
from longer-haul bus travel. DOT sees potential for
greater demand through improved service and regulato­
ry reform. A third projection estimates a more optimis­
tic 40-percent growth based on assumptions of fuel
shortages and restricted auto use. In contrast to these
three optimistic scenarios the ICC concludes that regular
route traffic will continue to experience flattened de­
mand and market share loss.1
1
In a period of energy shortage, bus operations are
likely to increase because of the comparative fuel effi­
ciency of this mode of transportation. This was demon­
strated both during World War II and in 1974 when
fuel shortages existed. Given energy priorities, buses
would make inroads into the use of the private automo­
bile. Presently, diesel turbocharged engines are being in­
troduced into service because of their potential for fuel
savings and reduced emissions. Gas turbine buses now
being used experimentally are able to run on non-petro­
leum based fuels and may aid future productivity

growth because of their increased reliability.1
2
Fuel shortages would likely create more reliance on
the use of buses for lower density routes to and from
small towns and rural areas. Higher utilization of
existing capacity in the industry would boost labor pro­
ductivity. However, a recent DOT study projects that
over the next two or three decades the passenger auto­
mobile will continue in its dominant transportation role
because of its flexibility and tailored service.1
3
Demographic changes may also help to increase the
demand for bus service, raising both load factors and
productivity. The trends toward population dispersion,
smaller households, and an older population are all fac­
tors which favor increased use of intercity bus service.
Population dispersion reduces the availability of other
forms of transportation; private cars are more cost effi­
cient for larger families; and many older persons prefer
the relative comfort and safety of bus travel.
However, a history of low productivity growth, lack
of demand, and reduced profits may impair the ability
of the industry to attract needed capital and enhance
future performance. The ICC sees a need for changes in
policy to insure a balanced transportation network.
Such changes would include bus and engine design
studies, similar to those conducted for air transporation
and other forms of mass transit, to find ways to in­
crease productivity. Improvements in the quality and
location of bus terminals and facilities have also been
recommended.1 Because the price differential between
4
long distance air fares and bus fares has narrowed over
the years, some analysts argue that bus carriers should
drop coast-to-coast service and concentrate in shorthaul markets of 100 to 200 miles. Such a system could
enlarge the number of daily departures and increase bus
utilization from its current average of 7 hours a day to
16 hours.1 Further advances in productivity are possible
5
through improvements in intermodal linkages. Con­
struction of municipal transportation terminals to serve
as connectors for bus, train, and plane service could im­
prove productivity for all of these forms of transporta­
tion.

' This study is based on statistics reported to the Interstate
Commerce Commission for all class I motor carriers of passengers.
Class I carriers are those that have 3-year average annual revenues of
more that $3 million. This portion of the bus industry, as defined in
the 1972 Standard Industrial Classification (SIC) manual, makes up a
small part of SIC 4111 (local and suburban transit), and a more sub­
stantial part of both SIC 4131 (intercity and rural highway passenger
transportation) and SIC 414 (passenger transportation charter ser­
vice). Based on their major source of revenue, class I carriers have
been divided by the ICC into local or intercity service. Local service
is defined as transportation performed within a city or town, includ­
ing service for the contiguous suburban area. Intercity service includes
all transportation performed beyond the limits set for local service.
Either of these carrier types may also engage in intercity, local, or
charter operations.

2The output measure underlying the productivity series for the bus
industry has been constructed using data on passenger-miles, passen­
gers, and express freight service, combined with appropriate weights
relating to labor importance. A technical note describing the methods
used in the construction of the index is available upon request.
1Lawrence Leist, Intercity Bus Service: Frequency and Running
Time, Report No. WP-220-04-20 (Washington, U.S. Department of
Transportation, 1975).
4 Transportation and the Future (Washington, U.S. Department of
Transportation, 1975), p. 35.
5Derived by dividing revenue passenger miles by revenue passengers.
6 The Intercity Bus Industry: A Preliminary Study (Washington, In­
terstate Commerce Commission, 1978), pp. 2-3.
7 Transportation facts and Trends (Washington, Transportation

sector as a whole increased more than 150 percent,
while gross stocks of capital increased 35 percent.1
0

Outlook




Association of America, 1980), p. 18.
"America's Most Fuel Efficient Passenger Transportation Service
(Washington, American Bus Association, 1979), p. 5.
’ The Intercity Bus Industry, p. 26.
See Capital Stock Estimates for Input-Output Industries: Methods
and Data, Bulletin 2034 (Bureau of Labor Statistics, 1979).




" The Intercity Bus Industry, pp. 106-08.
I: America's Most Fuel Efficient, p. 5.
" Transportation and the Future, p. 111.
1 The Intercity Bus Industry, pp. 121-27.
4
'■ Rush Loving, Jr., “The Bus Lines are on the Road to Nowhere,”
Fortune, Dec. 31, 1978, pp. 58-64.

Laundry and cleaning services
pressed to post productivity gains
Increases in 1958-76 averaged 1.6 percent,
as hours dropped twice as fa st as output;
wash and wear fabrics and home appliances
have displaced m any traditional laundries
R ic h a r d B. C a r n e s
celerated to 2.8 percent. However, both output and
hours decreased. Output fell at a 1.4-percent yearly rate;
hours fell more rapidly—about 4.0 percent per year.
Textile manufacturers introduced wash and wear fabrics
during these years—a development which reduced both
industry demand and industry labor requirements.
M any significant technological changes also contributed
to the reduction in hours in the late 1960’s.
D uring the m ore recent period, 1 9 7 0 -7 6 , p rod u c­
tivity grow th averaged on ly 0.7 percent per year. A
very rapid increase betw een 1 9 7 0 -7 3 w as partially
offset by a 1 9 7 3 -7 6 decline. T h e below -average pro­
d uctivity grow th since 1970 w as accom p an ied by
substantial d eclines in both ou tp u t and hours; 5.3
percent and 6.0 percent, per year, respectively. O u t­
put declines in the private business sector in 1970,
1974, and 1975 had a negative im pact on the laundry
and cleaning services industry. C onstant dollar per­
sonal consum ption expenditures for laundry and
cleaning services are estim ated to have d eclin ed 30
percent from 1970 to 1976.

H om e laundering— d evoid o f yesteryear’s drudgery
by the use o f m odern m ach in es and easy care fabrics
— has increased significantly in recent tim es, slack en ­
ing the dem and for traditional, labor-intensive, fam ­
ily laundry and drycleaning services. T his trend, in
conjun ction w ith the grow th o f m ore capital-in ten­
sive op eration s such as industrial launderers and
linen suppliers, has resulted in a large drop in the
hours w orked by all persons in the in d u stry .1 B e­
tw een 1958 and 1976, hours dropped at an average
annual rate o f 3.1 p ercen t.* But. because the in dus­
2
try’s ou tp u t d eclin ed at about on ly h a lf that rate, 1.6
percent per year, ou tp u t per hour “grew ” at an aver­
age annual rate o f 1.6 percent over the 19-year pe­
riod. O ver the sam e period, p roductivity in the n on ­
farm bu sin ess sector rose at an annual rate o f 2.2
percent. O n the w h ole, the cleaning in d u stry’s m o d ­
erate p rod u ctivity perform ance has been bu oyed by
gradual ad van ces in tech n o lo g y and the introduction
o f easy-to-clean apparel.
A lth o u g h the ann u al productivity gain in the
cleanin g industry averaged 1.6 percent from 1958 to
1976, grow th w ith in th e period varied m arkedly.
(See table 1.) During the 1958-65 period, for example,
the average annual increase in productivity was 0.9 per­
cent. Output grew at a 1.5-percent annual rate, while
hours increased at a rate o f 0.6 percent per year. From
1965 to 1970, the rate o f productivity growth ac­

Cleaning industry composition
T he chan ge in the dem and for various types o f
laundry and cleaning services has changed the in d u s­
try’s structure. (See table 2.) T hus, alth ou gh overall
output dropped at an average annual rate o f 1.6 per­
cent in 1 9 5 8 -7 6 , diverse m ovem en ts occurred at the
subindustry level. T h e focus o f service has shifted
from personal clean in g services to com m ercial and
industrial custom ers.

Richard B. Carnes is an economist in the Division of Industry Productiv­
ity Studies, Bureau of Labor Statistics.

Reprinted from the
M onthly L abor Review, February 1978.




83

D esp ite the 1 9 5 8 -7 6 increase in real per capita
in com e in the U n ited States (57 percent), real per
capita expenditures for fam ily laundry and cleaning
services declin ed by 50 percent. The sharp decline in
the ou tp u t o f fam ily laundry establishm ents has been
a direct result o f increased hom e laundering as w ell
as easy-to-care-for fabrics. N o w , fam ilies consider
w ashers and dryers to be necessary appliances. T he
output o f the h o u seh o ld laundry equipm ent industry
is estim ated to have increased nearly 70 percent be­
tw een 1958 and 1976. In 1975, m ore than 4 m illion
hom e w ash in g m ach in es w ere sold, increasing m ar­
ket penetration to 70 percent, from 53 percent in
I9 6 0 .3 T h e in trod u ction o f perm anent press fabrics in
the m id -1 9 6 0 ’s alon g w ith m ore recent m arketing o f
knit garm ents has further reduced the dem and for
fam ily laundry services. C onsequently, d eclines in
output have occurred for non sp ecialized pow er lau n ­
dries, n on p ow er laundries, and drycleaning plants.
In addition, there have been declines in ou tp ut for
garm ent pressing estab lish m en ts, rug clean in g and
repairing plants, and diaper services. F rom 1958 to
1972, the pow er and n on p ow er laundries declined
from 32 percent o f industry sales to 14 percent. D ryclean in g plants lost 3 percent o f the m arket, and
garm ent pressing sales dropped from 7 to 5 percent
over th is period.
B etw een 1958 and 1972 (the latest year for w hich
detailed data are available), sales by industrial lau n ­
dry and linen supply establishm ents— w h ich provide
both rental and clean in g services to businesses and
institu tion s— increased from 19 percent to 34 per­
cent o f total industry sales. In 1972, sales per estab­
lish m en t for industrial launderers and linen suppliers

T a b l© 1.

Table 2. The distribution of receipts and employment,
laundry and cleaning services, 1958 and 1972
Percent distribution
Industry segment

83.9
87.1
82.7

86.1
89.7
85.1

86.9
86.9
86.6
92.7
87.3

88.6
88.9
92.4
97.0
93.6

101.9
102.3
106.7
104.6
107.2

1966 .......................................
1967 .......................................
1968 .......................................
1969 .......................................
1970 .......................................
1971........................................

92.8
100.0
103.4
101.7
99.3
102.9

98.6
100.0
98.2
94.4
87.6
80.3

106.3
100.0
95.0
92.8
88.2
78.0

1972 ........................................
1973 ........................................
1974 .......................................
1975 ........................................
1976p........................................

107.0
109.6
107.3
104.1
105.2

79.7
75.7
69.3
65.4
63.2

74.5
69.1
64.6
62.8
60.1

^

100.0

100.0

100.0

33.6
29.9

41.0
37.7

24.8
13.1

30.0
20 7

2.1
1.6
38.8
37.0
1.8
20.7
12.6
6.8
1.3

1.9
1.4
39.5
38.3
1.2
14.5
8.8
4.8
.9

10.7
1.0
35.9
34.5
1.4
34.7
18.2
15.3
1.2

8.7
.6
43.4
42.0
1.4
23.0
15.2
6.8
1.0

6.9

4.9

4.6

3.6

fivefold increase since 1958. Currently there are
40,000 self-service laundry and drycleaning stores in
the United States.4
Limited capital requirem ents have encouraged the
growth of franchising in laundry and cleaning serv­
ices. The available data on franchised laundry and
drycleaning services show that average investment
and startup funds totaled $71,000 in 1974. In that

p = prelim
inary.




100.0

w ere 12 tim es greater than th e industry average o f
$60,000. B ecause o f the sim ilarity in types o f item s
to be cleaned, these establishm ents have introduced
m any standardized p rocessing techniques. T hese
new m ass-production m eth od s, in conjunction w ith
m ore m odern equipm ent, have greatly increased the
efficiency o f these typ ically large-scale laundries.
Industrial laundry establishm ents perform w ork
on a con tract basis, sp ecializin g in personalized w ork
garm ents, w iping cloth s, m ats, and dust control
item s. G arm ents are identified so they can be re­
turned after servicing to the sam e custom er. In m any
cases, businesses rent these item s from the laundry,
w h ich then provides service on a regular schedule.
L inen suppliers w ork on a sim ilar rental-contract
basis, except that laundered item s are not usually
personalized. L inen services include the rental o f ap­
parel, tow els, table and bed linen, uniform s, and a
variety o f other cloth item s. C ustom ers served by
industrial launderers and linen suppliers in clu d e fac­
tories, restaurants, hospitals, lod gin g places, and
sch ools.
O ther establishm ents that experienced sales
grow th from 1958 to 1972 w ere coin-operated lau n ­
dry and drycleaners. B ecause few er new apparel re­
quire professional cleaning, th e grow th in com m er­
cial coin-operated laundry and drycleaning facilities
has been encouraged. C oin-operated services a c­
cou n ted for 11 percent o f industry sales in 1972— a

102.6
103.0
102.9

1961.......................................
1962 .......................................
1963 .......................................
1964 .......................................
1965 .......................................

Paid
employees

SO R E US. D
UC:
epartm of C m Basedon1967 StandardIn u C
ent om erce.
d strial lassificationM u
an al.
D are for those establishm w payrolls.
ata
ents ith

Hours of all
persons

1958 .......................................
1959 .......................................
1960 .......................................

Receipts

L n ry services.....................
au d
Pow laundries.................
er
C
oin-operated lau d and
n ry
drycleaning.....................
L
aundries, except power___
D
rycleaning services...............
D
rycleaning plants..............
R cleaning and repairing. . .
ug
R services.......................
ental
L en supply.....................
in
In u launderers............
d strial
D service...................
iaper
G ent pressing, alteration and
arm
repair................................

[1967 = 100]
Output

Paid
employees

T
otal....................................

In d e x e s o f p r o d u c t iv it y , o u tp u t, a n d h o u r s o f

Productivity
(output per hour)

1972

Receipts

a ll p e r s o n s , la u n d r y a n d c le a n in g s e r v ic e s , 1 9 5 8 -7 6

Year

1958

84

new technology. In addition, improvements in fab­
rics, detergent products, and drycleaning solvents
have helped to reduce washing and drying times as
well as finishing requirements.
Although data on new capital expenditures are not
available for the laundry and cleaning services indus­
try, some insights into equipment purchases have
been obtained from data published for manufactured
commercial laundry equipment.9 In 1972, the com­
mercial laundry equipment industry had sales of
$219 million—more than double the 1958 sales level
of $107 million. (During this period, the Wholesale
Price Index for laundry equipment showed no in­
crease.) Laundry equipment accounted for 75 per­
cent of 1972 sales; drycleaning equipment accounted
for the remaining 25 percent. The distribution of
these sales has not changed since 1958. However, the
type of equipment sold has changed, reflecting im­
proved fabrics and newer, laborsaving, cleaning and
drying methods. In 1958, laundry pressing equip­
ment represented 16 percent of sales, but by 1972,
had declined to just 2 percent of sales. Similarly,
flatwork ironers have also declined in relative impor­
tance as “no iron” fabrics became prevalent during
the period of this study. Combination washer and
water extractors have advanced from 5 percent of
equipment sales in 1958 to 38 percent in 1972.
An innovation which has diffused rapidly through
the industry is steam tunnel finishing. After cleaning,
conveyors transport hangered garments through a
steam bath which relaxes the fibers and restores the
garments to their original shape. They are then dried
by high-velocity hot air. This equipment, when used
on synthetic fabrics or blends, significantly reduces
the amount of hand-operated pressing that was
previously required.
The use of synthetic fabrics has also led to a grad­
ual phasing out of older types of industrial and com­
mercial laundry equipment. Synthetic fabrics reduce
the machine cycle and process times because they are
lightweight and heat resistant. Lightly soiled gar­
ments, such as office apparel, are increasingly being
drycleaned or are being cleaned by dual-phase wash­
ing and drycleaning equipment.
Labor time has also been reduced by the introduc­
tion of fully automated, large-scale washers and dry­
ers. Washer-extractors can be self-loaded, while the
washing, extracting, and conditioning operations can
be programmed for different load conditions to per­
form automatically. These automated controls have
reduced operating personnel to as few as one person
for this stage of processing. Labor requirements for
large amounts of flatwork finishing have been signifi­
cantly reduced by the use of more fully automated
flatwork spreaders and high-speed folders and stack­
ers. Electronically controlled folders can process up

year, there were 3,700 franchise establishments with
sales of $237 million—about 5 percent of total indus­
try receipts. Three companies accounted for about 85
percent of both franchise establishments and sales/
To enhance sales, many franchise establishments
specialize in convenient 1-day drycleaning and shirt
service.
Shrieking labor input
Changes in employment among the subindustries
have mirrored the changes in industry organization.
From 1958 to 1976, the number of persons engaged
in laundry and cleaning services fell from 688,000 to
445,000—an average annual industry decline of 2.3
percent. As with output, the largest employment
losses occurred in power and nonpower laundries,
drycleaning plants, and garment services. Coinoperated laundry and drycleaning establishments
along with industrial laundry and linen supply estab­
lishments experienced employment growth.
Over the 19-year period, the industry average
weekly hours for nonsupervisory workers declined
from 38.7 to 35.9; average weekly hours for manag­
ers and the self-employed dropped from 52.9 to 46.2.
As a result of the overall decreases in employment
and average weekly hours, total hours fell at an aver­
age annual rate of 3.1 percent from 1958 to 1976.
The occupational composition within the laundry
and cleaning service industry has also been changing.
Since 1970, when detailed data first became availa­
ble,6 industry employment has been decreasing for
most occupations. Only white-collar occupations
have posted employment gains in recent years. Tra­
ditional blue-collar workers—craft and service
workers, operatives, and laborers—lost a significant
share of industry employment, dropping from 64
percent in 1970 to 58 percent in 1974. Improvements
in fabric finishes have reduced demand for pressing
operatives, stock handlers, laundry and drycleaning
operatives, and craftworkers such as tailors, station­
ary engineers, and Vehicle mechanics. Demand for
route drivers has also fallen for family laundries as
their sales have declined. Clerical workers, manag­
ers, salesworkers, and professional and technical
workers have had corresponding increases in their
relative share of employment, growing from 36 to 42
percent over the same 5-year period.7
Cleaning technology
Increased mechanization and new cleaning meth­
ods for synthetic fabrics are the primary areas of
technological change in the industry/ Significant ex­
penditures for larger, more efficient washers and dry­
ers, “steam tunnels,” ultrasonic drycleaning ma­
chines, high speed sorters and folders, and
mechanical transfer devices reflect the demand for



85

to 1,500 items per hour with various automatic width
and fold combinations possible. Manual material
handling for small-piece flatwork has been reduced
by the use of air jets to propel items from extractors
to ironers or mechanized folders.1
0
Gains in productivity have resulted from the intro­
duction of integrated systems for material handling.
Conveyors are often used for loading washing wheels
and to move cleaned loads to in-line extractors, tum­
bler-dryers, and to the finishing department. In in­
dustrial laundries, these material handling systems
are further used to automatically route personalized
garments through the steam tunnels to the proper
assembly area. Interest currently centers on the de­
velopment of a hangered garment system which can
be automatically scanned and sorted.1
1
Looking down the line
The decline in family laundries and the growth of
high-volume, rental-cleaning services is expected to
continue. Productivity should advance over the next
decade as technological innovations continue to be
introduced and diffused throughout the industry. In
many cases, further improvements in textiles will
result in longer lasting garments that will be even
easier to clean. Polyester fabrics, for example, are
being improved so they can be washed at lower tem­
peratures. Lower temperatures and improved fabrics
will reduce detergent requirements and shrinkage as
well as improve color retention.

1The study covers paid and unpaid persons (including paid employees,
partners, proprietors, and unpaid family members) working in laundry
and cleaning service establishments. The industry is designated as laun­
dry, cleaning, and garment services, SIC 721, in the 1972 Standard
Industrial Classification (SIC) Manual. SIC 721 is composed of 9 subin­
dustries: 7211, Power Laundries Family and Commercial; 7212, Gar­
ment Pressing and Agents for Laundries and Dry Cleaners; 7213, Linen
Supply; 7214, Diaper Service; 7215, Coin-operated Laundries and Drycleaning; 7216, Drycleaning Plants Except Rug Cleaning; 7217, Carpet
and Upholstery Cleaning; 7218, Industrial Launderers; 7219, Laundry
and Garment Services not elsewhere classified.
2 All average annual rates of change are based on the linear least
squares trend of the logarithms of the index numbers. Updated indexes
will appear in the annual BLS Bulletin, Productivity Indexes for Selected
Industries.

Because comprehensive delivery routes already
exist, many laundry and cleaning establishments are
expected to increase sales by broadening the product
lines and services they offer. Rental service firms, for
example, are now marketing such items as disposable
rest room products and air fresheners.
More stringent environmental air and water stand­
ards have acted to reduce gains in productivity as
plants divert labor and capital from the production
process to meet these new demands. The impact of
collecting and processing waste products and meet­
ing fabric flammability standards has affected plant
operating methods and equipment needs. In some
cases, customers must be screened to insure that
their garments will not cause unacceptable levels of
water and air pollution. This has led to industry
requirements for sophisticated ventilation and filtra­
tion equipment along with hot water and heat recov­
ery systems. As new technology is developed to meet
these environmental and health needs, future trends
in productivity and output could be enhanced. For
example, “sniffer” units which help reduce solvent
emissions into the air can also be used to reclaim and
recycle usable solvent, thus increasing operating effi­
ciency. Demand in some segments of the industry
could increase as establishments which currently
launder their own textiles turn to commercial laun­
derers and cleaners who have developed the equip­
ment and expertise needed to meet stricter environ­
mental standards.

4 Information provided by the National Automatic Laundry and
Cleaning Council.
5 Andrew Kostecka, Franchising in the Economy, 1974-76 (Washing­
ton, Department of Commerce, 1976), p. 71.
6 Bureau of Labor Statistics, unpublished data for 1970-85, National
Industry Occupational Matrix.
7 Ibid.
8 For an earlier study of technology in this industry, see Mary L.
Vickery, “New technology in laundry and cleaning services," Monthly
Labor Review, February 1972, pp. 54-59.
Bureau of the Census, Census of Manufacturers, 1972 Industry
Series: Service Industry Machines and Machine Shops, MC72 (2)-35G
1975
1 Vickery, New technology, p. 56.
0

3See “ 1976 Statistical and Marketing Report,” Merchandising, March
1976.




1 The Year 75-76 in Review,” Industrial Launderer, November
1
1976, p. 7.

Cyclical behavior of productivity
in the machine tool industry
Productivity growth was slow during 1958-80,
partly because o f the industry's tendency
to retain skilled workers during cyclical downturns;
computers and other electronic equipment aided production,
but diffusion o f such innovations has been slow
Jo h n D

uke and

H o rst B r a n d

Output per employee-hour in the machine tool industry
rose at an average annual rate of 1.1 percent over the
1958-80 period—significantly below the 2.8-percent
rate for manufacturing.1 A combination of factors
slowed productivity in the machine tool industry, in­
cluding the tendency of machine tool firms to keep
highly skilled workers on the payroll, even when output
fell during cyclical slowdowns, and the slackened de­
mand for capital goods after the mid-sixties. However,
the slowdown was moderated by technological advances
in the manufacture of machine tools, as well as by high
rates of productivity improvement in periods of cyclical
recovery.
Until 1966, productivity in the machine tool industry
rose at a high annual rate, but thereafter the rate de­
clined for several years. Its subsequent recovery re­
mained incomplete— the high levels of the mid-sixties
were not reattained. The recovery was again interrupted
by a slump in 1974; it resumed in 1977, continuing to
1979, but even then productivity did not top its 1966
peak. (See table 1.) The cyclical behavior of productivi­
ty in the industry and in manufacturing is shown in the
following tabulation (average annual changes in per­
cent):

M a c h in e to o ls

M a n u fa c tu rin g

Upswings:
1958- 59 ...................
1961-66 ......................
1971-74 ......................
1976-80 ......................

23.1
5.6
7.8
2.4

4.8
4.4
2.9
0.9

Downswings:
1959- 61 ...................
1966-71 ......................
1974-76 ......................

-2 .0
-4.2
-5 .2

1.7
2.0
3.7

Productivity in both the metal cutting and metal
forming segments of the industry paralleled the cyclical
patterns shown above, although amplitudes differed.
Productivity improvement averaged 1.5 percent annual­
ly in metal cutting (which accounts for three-fourths of
total industry employment), and 0.1 percent in metal
forming. Upswings in productivity were more pro­
nounced in metal cutting than in metal forming; down­
swings were more pronounced in metal forming. In
metal cutting, productivity dropped in 8 of the 22 years
examined (table 2); in metal forming, in 12 (table 3).
The drops were only in part associated with general
business cycles; they occurred in years of economic ex­
pansion as well as during contractions.
O utput recovery slow in the seven ties

John Duke and Horst Brand are economists in the Office of Produc­
tivity and Technology, Bureau of Labor Statistics.
Reprinted from the
M onthly L abor Review, November 1981.




The machine tool industry manufactures cutting tools
for boring, drilling, gear cutting, grinding, and milling

machines and lathes, as well as forming tools such as
punching, shearing, bending, and forming presses. These
tools are usually shipped as units, that is, as single-pur­
pose machines, but their basic features may also be
combined into “machining centers.” The machine tools
may be equipped with manual controls or with pro­
grammed numerical controls which require little labor
by users. Machine tools are not mass produced, al­
though they may make mass production processes pos­
sible in user industries. Rather, the parts and
components of a finished machine tool are usually made
in relatively small batches, and require comparatively
large amounts of labor.
Output in the machine tool industry rose at an aver­
age annual rate of 1.6 percent between 1958 and 1980,
compared with 3.8 percent for manufacturing. Underly­
ing the long-term trend were cyclical swings of consid­
erable amplitude. The metal cutting and metal forming
segments of the industry traced similar cyclical patterns.
(See table 4.)
The following tabulation shows the cyclical behavior
of output in the machine tool industry and in manufac­
turing, 1958-80 (average annual changes in percent):
M a c h in e tools

Upswings:
1958-59
1961-66
1971-74
1976-80

Table 1. Productivity and related indexes for the machine
tool industry, 1958-80
[1977 = 100]

Year

Output

1958 ...........
1959 ...........
1960 ...........

71.5
88.0
84.7

63.0
79.2
82.8

88.1
90.0
97.8

Employee-hours

1961
1962
1963
1964
1965

...........
...........
...........
...........
...........

84.5
88.5
90.1
99.9
101.4

77.4
88.0
93.0
112.3
125.3

91.6
99.4
103.2
112.4
123.6

1966
1967
1968
1969
1970

...........
...........
...........
...........
...........

111.7
101.8
97.9
100.1
91.7

156.1
149.9
137.6
137.8
112.0

139.8
147.3
140.5
137.7
122.1

1971
1972
1973
1974
1975

...........
...........
...........
...........
...........

87.9
98.0
107.3
109.4
103.0

81.4
91.2
116.3
127.4
109.1

92.6
93.1
108.4
116.5
105.9

1976
1977
1978
1979
1980

...........
...........
...........
...........
...........

98.4
100.0
102.6
107.0
106.9

93.9
100.0
111.8
125.9
129.1

95.4
100.0
109.0
117.7
120.8

Average annual rates of change (in percent)

1958-80 .......
1975-80 .......

1.1
1.3

1.6
5.4

0.5
4.0

M a n u fa c tu rin g

.................
.................
.................
.................

25.7
14.6
17.2
9.1

11.7
8.2
5.9
2.9

Downswings:
1959-61 .................
1966-71 .................
1974-76 .................

-1.1
-11.1
-14.1

0.2

1.0
0.9

Recoveries in machine tool output during the seven­
ties were less vigorous than they had been in the 1958—
59 and 1961-66 upswings. Slumps were deep. Longterm factors contributing to the comparative weakening
of output included the volatility in the demand for pro­
ducers’ durable equipment. Following 12 percent annual
increases in the 1961-66 period, growth in demand for
producers durable equipment contracted to 2 percent a
year for 1966-71. Demand rebounded at an 11-percent
annual rate in the early seventies, declined by 3 percent
annually over the 1974-75 period, then recovered to a
10-percent annual growth rate in 1976-79. Even so, the
long-term growth in the demand for producers’ durable
equipment slackened in the seventies (compared with
the demand in the sixties) from an average annual
growth rate of 8.1 percent in 1958-68 to 4.8 percent in
1968-79. However, the levels of the sixties were consis­
tently exceeded subsequently—contrary to the situation
in machine tool output and productivity. Thus, the rela­
tion between producers’ durable output and machine
tool output clearly weakened.



Output per employee-hour

88

During the seventies, a number of metalworking in­
dustries representing key markets for machine tools reg­
istered comparatively slower growth or actual declines
in output. For example, production of motor vehicles
after the mid-sixties rose at pnly about one-half the rate
for 1959-66. Similarly, output growth of construction
machinery contracted. Steel output, which had ad­
vanced at more than 5 percent a year until 1966, be­
came stagnant thereafter, then fell, as did output of
electric motors and generators, nonferrous metals,
household appliances, and household furniture.2
Furthermore, expenditures for machine tools dropped
as a proportion of total equipment expenditures by
manufacturing firms. In the sixties, such expenditures
accounted for 11 percent of the total, in the seventies,
for only 9 percent. Moreover, imports increasingly
displaced domestic machine tools. In the sixties and up
to 1973, machine tool imports averaged well under 10
percent of total U.S. machine tool units purchased;
thereafter, the volume of machine tool imports soared,
and by 1978, they accounted for 21 percent of total
units purchased.3In contrast, exports did not rise mark­
edly relative to output—exports represented 8 percent
of machine tool units purchased in the sixties and about
10 percent in the seventies.
Still another factor underlying slackened output of
machine tools has been the rapid rise in their productive
capacity. (This factor will be explained more fully later
in this article.) A study of more than 350 companies

showed that reduced machining time for numerically
controlled (or programmed) machine tools ranged from
35 percent to 50 percent.4 According to the American
Machinist's periodic inventories of metalworking equip­
ment, the “population” of machine tools in use did not
change significantly between 1963 and 1976-78, but the
output of the metalworking industries using them gener­
ally increased, indicating rising productive capabilities
of the machine tools, particularly those equipped with
numerical controls.5 Some engineering authorities main­
tain that numerically controlled machine tools permit
“drastically reduced” handling time because they elimi­
nate the separate operations of transferring and
clamping and unclamping.6
The relative importance of all categories of machine
tools lessened during 1958-80, except lathes, drill­
ing machines, and machining centers. (Machining cen­
ters combine the separate operations of boring, drilling,
and milling units.) Most of the shift toward machining
centers occurred after 1968, when the diffusion of nu­
merical control, an essential component of machining
centers, began to accelerate. In 1978, the number of ma­
chining centers shipped was half again as high as in
1968. During that decade, the number of numerically
controlled metal cutting machine tools shipped more
than doubled and the number of metal forming machine
tools rose by 14 percent.
The diffusion of numerically controlled machine tools
has remained limited, however. According to the Amer­

Table 3. Productivity and related indexes for metal
forming, 1958-80
(1977 = 100]

Year

1958 ...........
1959 ...........
1960 ...........

Output

67.6
83.2
81.5

58.1
74.2
81.0

...........
...........
...........
...........
...........

80.0
83.2
84.3
94.9
98.7

72.7
83.0
88.4
109.2
124.8

...........
...........
...........
...........
...........

107.8
98.0
95.7
97.5
89.5

154.7
150.6
139.8
139.0
107.2

...........
...........
...........
...........
...........

85.5
94.8
105.5
108.9
102.9

75.2
83.9
108.6
122.3
107.5
92.5

...........
...........
...........
...........
...........

97.3
100.0
103.6
109.7
111.2

100.0
113.7
130.6
138.3

100.0
109.7
119.0
124.4

1.5
2.3




1.9
7.2

...........
...........
...........
...........
...........

123.1
112.7
103.9
107.0
98.5

160.8
147.9
131.7
134.3
126.2

130.6
131.2
126.8
125.5
128.1

...........
...........
..........
..........
...........

95.7
107.5
114.1
111.9
104.0

99.6
112.1
139.2
142.5
114.1

104.1
104.3
122.0
127.4
109.7

...........
...........
...........
...........
...........

101.7
100.0
99.9
100.4
95.2

98.1
100.0
107.2
114.8
106.1

96.5
100.0
107.3
114.3
111.5

0.1
-1.4

0.7
0.5

0.6
1.9

E m ploym ent concentrated in m etal cutting

In 1980, employment in the machine tool industry
numbered about 108,000 persons, with about one-quar­
ter of them in metal forming establishments. Employeehours rose quite slowly over the 1958-80 period (0.5
percent, compared with 1 percent in manufacturing)
but, like productivity and output, were characterized by
pronounced cyclical swings. The cyclical volatility of
employee-hours in the machine tool industry, compared

Average annual rates of change (in percent)
1958-80 . . . .
1975-80 . . . .

93.8
98.7
99.2
105.8
116.3

Employee-hours

ican Machinist's 1976-78 inventory of metalworking
equipment, only 2 percent of the machine tools in the
United States were numerically controlled, and only 7
percent of machine tools 10 years old or less were nu­
merically controlled.7
The output capacity of metal forming machine tools,
like that of metal cutting tools, significantly increased
during 1958-80, tending to retard demand and, hence,
output growth. For example, the size of presses used in
the automotive and appliance industries— which ac­
count for the lion’s share of the demand for presses—
has increased such that, in the past 15 years, it tended
to be four times greater than that in the preceeding 35
years.8 Changes of dies, which used to require 30 to 40
minutes, now take only 90 seconds— hence, long pro­
duction runs are no longer needed to justify die chang­
es.9 Numerical controls have been applied to operations
such as bending— now tube benders perform more than
30 types of bends.1
0

88.0
88.5
102.9
112.3
104.5
95.1

1976
1977
1978
1979
1980

91.9
104.3
107.5
122.2
127.1

Output

Average annual rates of change (in percent)

143.5
153.6
146.1
142.5
119.8

1971
1972
1973
1974
1975

98.0
105.7
108.4
115.5
109.3

1958-80 .......
1975-80 .......

90.9
99.7
104.9
115.1
126.4

1966
1967
1968
1969
1970

...........
...........
...........
...........
...........

1976
1977
1978
1979
1980

85.9
89.2
99.4

1961
1962
1963
1964
1965

93.8
92.5
94.1

1971
1972
1973
1974
1975

Employee-hours

1958 ...........
1959 ...........
1960 ...........

78.4
95.1
88.9

1966
1967
1968
1969
1970

[1977 = 100]

Output per employee-hour

83.6
102.8
94.5

1961
1962
1963
1964
1965

Table 2. Productivity and related indexes for metal
cutting, 1958-80
Year

Output per employee-hour

0.5
4.8

89

with manufacturing, is illustrated in the following tabu­
lation (average annual change in percent):
M achine tools

M an ufactu ring

...............
...............
...............
...............

2.2
8.5
8.8
6.6

6.6
3.6
2.9
1.9

D ow nsw ings:
1959-61 ...............
1966-71 ...............
1 974-76 ...............

0.9
-7 .3
-9 .5

0.9 percent per year versus 0.3 percent. There were 43
percent more nonproduction workers in 1980 than in
1958, and 38 percent more production workers, al­
though employment of both groups was below the 1967
peak. In metal cutting, the proportion of nonproduction
workers remained above 30 percent of total employment
during the period, reflecting the continued importance
of engineers, designers, and other leading personnel.
The proportion of women also rose, from 9 to 13 per­
cent of total employment, but was still far below the
manufacturing average of 31 percent.
In metal forming, the number of production workers
showed no change on average; in contrast, non­
production workers rose 2.6 percent—from 31 percent
of total employment in 1958 to 34 percent in 1980. Oc­
cupational data are not available for the machine tool
industry, but are available for the metal working ma­
chinery group of industries, of which the machine tool
industry accounts for about 30 percent of employment.
The occupational mix in the machine tool industry is
unlikely to differ very much from that in metalworking.
In 1978, metalworking machinery had an unusually
high percentage of craft and kindred workers—nearly
one-third of its employment, compared with just under
one-fifth for manufacturing. As might be expected, the
proportion of metal craftworkers and machinists consid­
erably exceeded the manufacturing average. Operatives
accounted for a smaller proportion of employment in
metalworking than in manufacturing (33 percent versus
43 percent), although the proportion of semiskilled
workers in metalworking was nearly three times higher
(15 percent versus 6 percent). As for professional and
technical workers, the employment differences were
small between the metalworking and all manufacturing
industries— 9 percent versus 10 percent—and this was
true for other white-collar categories. However, from
1970 forward, the rise in the number of professional
and technical workers was almost three times greater in
metalworking than in manufacturing—-14 percent ver­
sus 5 percent— reflecting the growing relative impor­
tance of electronic technicians and computer and
numerical control specialists and programmers.

-1 .4
-1 .0
-2 .6

U psw ings:
1 9 58-59
1 9 61-66
197 1 -7 4
1 976-80

Although recoveries in employee-hours in the seven­
ties were about as strong as in the sixties, the levels of
the mid-sixties were not reached. In 1980, employeehours were one-fifth below those of the sixties. Employ­
ment was less affected by cyclical swings and was 17
percent lower in 1980 than in 1967, the peak year of the
22-year period. The metal cutting and metal forming
segments of the industry displayed comparable cyclical
patterns in employee-hours. (See table 4.)
The cyclical declines in output and, hence, in employ­
ee-hours, probably aggravated the industry’s perennial
shortages of skilled help when business picked up again.
In part, these shortages were met through overtime
work. Following are relatives of overtime hours in the
metal cutting and metal forming segments of the ma­
chine tool industry (overtime hours in manufacturing =
100):
M etal cutting:
1958 .................
1959 .................
1960 .................
1 9 6 1 .................
1962 .................
1963 .................
1964 .................
1965 .................
1966 .................
1967 .................
1968 .................

.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.

60
122
144
113
143
157
181
175
203
206
131

1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980

...........
...........
...........
...........
...........
...........
...........
...........
...........
...........
...........
...........

...........
...........
...........
...........
...........
...........
...........
...........
...........
...........
...........
...........

150
110
55
117
168
191
138
103
154
178
188
211

M etal forming:
1972 .................
1973 .................
1974 .................
1975 .................

.
.
.
.

. 134
. 189
. 206
. 154

1976
1977
1978
1979
1980

...........
...........
...........
...........
...........

...........
...........
...........
...........
...........

129
140
172
191
161

T echnology diffused gradually

A number of important innovations have been
adopted in the manufacture of metal cutting and metal
forming machine tools, but diffusion among machine
tool producers has been slow— slower than among in­
dustries which apply the innovations in mass produc­
tion. As will be documented, this slowness is related to
the predominance of small firms which produce small
batches of frequently complex machinery and compo­
nents. The machine tool industry is labor-intensive, rel­
ative to most manufacturing industries, as indicated by
the high ratio of payroll to value added. Over the 1958—
77 period, this ratio averaged 58 percent for metal cut-

In only 2 years (1958 and 1971) of the review period
did overtime in metal cutting fall below the manufactur­
ing average. In all other years it was above, and often
was half again to twice as high. Metal forming (for
which pertinent data exist only since 1972) showed the
same overtime pattern.
The number of nonproduction workers in metal cut­
ting rose more rapidly than that of production workers,



90

Table 4.

Cyclical behavior of productivity in the machine tool industry and its components, 1958-80

[Average annual rates in percent]
i Output per employee-hour
Period

1958-80 .................

Machine
tools

Meta!
cutting

Output

Metal
forming

Machine
tools

Metal
cutting

Employee hours
Metal
forming

Machine,
tools

Metal
cutting

Metal
forming

1
.1

1.5

0.1

1.6

1.9

0.7

0.5

0.5

0.6

......................
......................
......................
......................

23.1
5.6
7.8
2.4

23.1
6.3
8.7
3.7

23.0
3.8
5.4
-1.3

25.7
14.6
17.2
9.1

27.7
16.1
18.7
11.3

21.3
10.6
13.8
3.0

2.2
8.5
8.8
6.6

3.8
9.2
9.2
7.4

-1.4
6.5
7.9
4.3

Downswings:
1959-61 ......................
1966-71 ......................
1974-76 ......................

-2.0
-4.2
-5.2

-1.9
-4.0
-5.5

-2.4
-4.6
-4.7

-1.1
-11.1
-14.1

-1.0
-12.4
-13.0

-1.7
-7.8
-17.0

0.9
-7.3
-9.5

0.9
-8.8
-8.0

0.7
-3.4
-13.0

Upswings:
1958-59
1961-66
1971-74
1976-80

ting establishments, and 60 percent for metal forming
establishments, compared with 52 percent for non­
electrical machinery, 52 percent for transportation
equipment, and 47 percent for all manufacturing. The
mass production techniques made possible by machine
tools generally cannot be used in building them, al­
though significant improvements in small-batch produc­
tion processes have resulted from some basic techno­
logical advances.
By far the most important development in machine
tool technology has been the evolution of numerical
control. In fact, numerical control has reshaped ma­
chine tool technology, and continues to transform it.
Essentially, numerical control made multifunction ma­
chine tools possible (exemplified by the machining cen­
ter, discussed earlier). According to Iron Age, numerical
control tools have been decisive in achieving “the criti­
cal balance . . . . in machine construction and rigidity,
horsepower, speed and feed ranges, standard tooling
and management control over the machine cycle and
operation.” " Numerical control was first applied in the
manufacture of machine tools in the mid-fifties, but cer­
tain innovations were required to lower its cost and,
thus, spur adoption by the smaller machine tool firms.
Although these innovations have occurred, their impact
on productivity was retarded by the severe cyclical
downturns in the industry’s business in the early and
middle seventies.
Numerical control is a method whereby metal cutting
(and to some extent metal forming) machine tools are
controlled by instructions which are programmed and
then punched on a tape. Information from the tape is
converted into instructions which position the tools
with respect to the workpiece; no templates, drill jigs,
or stops are used and manual operation is not neces­
sary. (The operator can service more than one numeri­
cally controlled machine tool.) A feedback mechanism
adjusts (or stops) the tool’s movement if programmed
distance does not adhere to commanded tolerance, and
stops it when the process is completed.1
2
Numerical control has always required drives which



91

would ensure that performance followed command. Hy­
draulic servomechanisms are still used for this purpose.
In the late sixties, however, silicon-controlled rectifiers
(which are solid-state devices) were introduced; these,
together with improvements in the control motor, made
possible much higher degrees of accuracy in machining
work. Also, tool life was extended as gear transmission,
hand wheels, and clutches were eliminated.1 Perhaps
3
most important, the substitution of transistors, and lat­
er of integrated circuits, for electric relays reduced the
number of control components by up to 90 percent, and
the amount of wiring by up to 80 percent.1 These devel­
4
opments slashed costs, and also allowed less highly
trained personnel to program the machines. Thus, im­
proved control mechanisms gave impetus to the dif­
fusion of numerical controls.
Numerical controls accelerated the consolidation of
machine tool production—as well as the production of
metalworking equipment— into machining centers. Ma­
chining centers are basically milling machines which
also drill, ream, bore, tap, and so forth. In machining
centers, complex shapes may be made by mounting cut­
ting tools of varying sizes and power configurations on
a single spindle. The cutting tools then are automatical­
ly changed by transfer arms, which also store the tool.
These automatic tool changes take only a few seconds;
formerly several minutes of an operator’s time were re­
quired.1 Machining centers also eliminate the need to
5
design, build, and store the jigs and fixtures needed by
single-purpose machines.1
6
Single-purpose machines also have been much im­
proved by numerical controls. For example, numerically
controlled boring machines have reduced downtime for
loading and unloading by up to 30 percent.1 Numerical
7
control applied to grinding machines often halves lay­
out time; programmable electronic wheel feed and wheel
retraction have been developed which reduce labor time
and enhance precision. The design of hobs for gear cut­
ting has been subjected to computer calculation, saving
cutting time.1
8
Cutting tool materials have become harder, permit­

gures for all manufacturing; for metal forming machine
tools, the ratio was 40 percent. Fixed assets per worker
in metal cutting and metal forming were 77 percent and
81 percent of the manufacturing average in 1976. More­
over, the long-term growth in the industry’s expendi­
tures for new plant and equipment, expressed in
constant dollars, averaged 2.7 percent annually between
1958 and 1978— compared with 4.6 percent for all
manufacturing industries. However, these long-term
trend indicators obscure significant cyclical changes.
Following are average annual rates of change in expen­
ditures (in constant dollars) for new plant and equip­
ment in the machine tool industry and in all industries,
1958-78:2
4

ting increased cutting speeds (albeit at the cost of re­
quiring heavier, more powerful machines). Tungsten
carbide which replaced high-speed steel in 1929 was in
turn supplanted by ceramic materials and polycrystal­
line diamond-tipped tools. Until 1900, cutting speeds
ran up to 25 feet per minute; high-speed steel tools av­
eraged 90 feet per minute; tunsten carbide, 150 feet per
minute; ceramic materials, 650 feet per minute; and
polycrystalline diamond-tipped tools can cut several
thousand feet per minute. Meanwhile, the older cutting
materials have been improved— for example, steel tools
are hardened by cobalt and continue to be widely used.
Naturally, the high speeds enlarge the machine tool’s
output capacity.1
9
Metal cutting tools predominate over the use of metal
forming tools in the manufacture of either type of ma­
chine tool. Thus, technological improvements in metal
forming tools and increases in their output capacity
have, of course, greatly benefited those who use the
tools intensively, but have only marginally affected pro­
ductivity of those who produce the tools.2
0
Computers are used in tandem with or incorporated
into numerically controlled machine tools where reli­
ability or control is crucial (as in the machining of
frames for aircrafts), or where minimizing of downtime
is essential. The recent trend has been toward relatively
small computers interfacing with individual machines,
rather than a single computer controlling a number of
machines.2 The computer has also been used in produc­
1
tion management, as well as in the design of machine
tools, significantly reducing labor requirements of engi­
neering and drafting personnel. Conventionally, engi­
neers and aides graphed the design for a machine tool
on drawing boards, according to a customer’s specifica­
tions; corrections usually required redrawing of all or
most of the design to preserve proportionalities. Now,
computers do the corrected redrawing, cutting the time
required for such corrections. This so-called interactive
graphics has permitted a 4-fold increase in the design­
er’s productivity. Memory storage of given designs fur­
ther aids productivity.2
2

M achine tools

2.3
29.5
28.0
16.7

7.3
10.2
7.3
10.5

D ownswings:
1959- 61 ...................
1966-71 ......................
1974—
76 ......................

-9.5
-1 8 .0
-9 .6

2.4
1.4
-3.8

Cyclical patterns in the real value of the industry’s
capital outlays parallel those for productivity, output,
and employee-hours. Even though capital outlays were
strong during the upswings of the seventies, they did
not reattain the levels of the sixties. In the 1976-78 up­
swing, the outlays were nearly one-third below those of
the mid-sixties, while outlays for all industries were
nearly a third higher.
The machine tool industry’s low levels of expendi­
tures for plant and equipment are reflected in the
relatively high average age of its equipment. According
to the American Machinist, 23 percent of the industry’s
machine tools were less than 10 years old in 1976-78,
compared with 31 percent for all metalworking indus­
tries; 37 percent were 10 to 19 years old, compared
with 35 percent for all metalworking and 40 percent
were more than 20 years, compared with 34 percent.
The American Machinist's periodic inventories suggest
that user industries tend to delay replacement of aging
machine tools. On average, only 31 percent of machine
tools in service in all metalworking industries were less
than 10 years old in 1976-78, compared with 36 per­
cent in 1968 and in 1963; 34 percent were more than 20
years old in 1976-78, compared with 23 percent in 1968
and 21 percent in 1963.2
5
The rising average age of machine tools may have
been offset to some degree by the high proportion of
parts and rebuilt machine tools shipped by toolmakers.
Parts for metal cutting tools and rebuilt machine tools
accounted for 19 percent of total shipments in 1976,

Relatively old capital stock
The machine tool industry, although vital for the
expansion and modernization of industrial machinery,
has spent relatively little for its own plant and equip­
ment. During the review period, the long-term growth
in such spending was significantly below that for all in­
dustries. One of the results has been that the average
age of equipment in the machine tool industry is well
above that in all other metalworking industries.2
3
According to 1977 census data, plant and equipment
expenditures per employee in metal cutting machine
tools represented only 52 percent of the comparable fi­



A ll industries

Upswings:
1958- 59 ...................
1961-66 ......................
1971-74 ......................
1976-78 ......................

92

are anticipated from automotive and aircraft manufac­
turers, and from manufacturers in other metalworking
industries requiring more “flexible” technology for
small-batch production.2
6
For the next several years, the automotive industry
will be retooling for the production of smaller, more en­
ergy-efficient vehicles, at an estimated cost of $60 bil­
lion. Undoubtedly, this will strain machine tool
manufacturing capacity. However, in the long run, de­
mand for machine tools from the automotive industry is
likely to slacken because of the prospective reduction in
the number of automobile models.2 Similarly, the air­
7
craft industry may replace about one-half of its 6,000
commercial air carriers, some of which were placed in
service 20 years earlier. New configurations of air
frames will be needed which conform with mandated re­
quirements to reduce noise levels and fuel consumption.
Therefore, the aircraft industry will need more costeffective machine tools.2 Metalworking firms generally
8
have become concerned with more efficient production
of small batches of parts and components; their interest
in automated batch manufacturing systems is likely to
intensify. In such systems, electronically-controlled as­
semblages of machine tools are linked by material-han­
dling equipment so as to convert a system of discrete
parts manufacturing into one of continuous (or nearly
continuous) processing.2 Automatic-batch manufactur­
9
ing systems have been increasingly used in the construc­
tion machinery industry.3
0
The building of craftworkers skills into the machine
began when Eli Whitney constructed musket-making
machines in the early 19th century.3 The need “to build
1
the skill in the machine” arose partly from the perennial
shortage of craftworkers (which often resulted in un­
skilled workers operating complex equipment) and part­
ly from the increased precision demanded of machine
tools. Quite possibly, the diffusion of numerically con­
trolled machine tools will accelerate the trend “to build
the skill in the machine” in the eighties. As noted in the
discussion on occupational patterns, this trend has af­
fected the machine tool industry less than most other
industries. This occupational pattern has been projected
to persist: in 1990, the Bureau of Labor Statistics proj­
ects that 31 percent of metalworking machinery indus­
try employees will be skilled craftworkers (only slighly
below the 1980 proportion), compared with 20 percent
for all manufacturing. Thus, the Bureau’s projections
implicitly assume that skill needs in the metalworking
industry will change little; and that in the machine tool
industry, it will continue to be difficult, at times even
infeasible, to build the skill of craftworkers into ma­
chine tools.
Nevertheless, the diffusion of numerically controlled
machine tools will probably accelerate under the Spur of

compared with 14 percent in the late sixties. Parts for
metal forming tools and rebuilt machinery constituted
33 percent of shipments in 1976, compared with 20 per­
cent in the late sixties. The proportion rises in periods
of slack business, but the rise may, in part, indicate
intensified efforts to retrofit and upgrade aging machine
tools, in lieu of purchasing new machines.
However, the high average age of equipment in the
machine tool industry may have been partially offset
through the replacement of worn-out parts, or by the
rebuilding of machines along more up-to-date lines.
Furthermore, the industry has an above-average propor­
tion of numerically controlled machine tools— nearly 4
percent of its tools are numerically controlled, com­
pared with 2 percent for all metalworking industries.
Because numerically controlled machine tools are gener­
ally under 15 years old, they probably represent at least
6 percent of the industry equipment that has been in
service less than 20 years, and surely a much larger pro­
portion of its total output capacity.
Industry structure, The structure of the machine tool in­
dustry does not differ much from that of manufacturing
as a whole. In 1972, the latest year for which data are
available, the four largest of the nearly 900 companies
making metal cutting machine tools accounted for 25
percent of the industry’s total employment, 22 percent
of its value of shipments, and 30 percent of its capital
expenditures. In metal forming, concentration was
slightly less. The 50 largest metal cutting companies,
representing 10 percent of all establishments in the in­
dustry, accounted for three-quarters of employment,
value of shipments, and capital expenditures. Trends in
value added per employee by employment size class of
establishment suggest that productivity has risen at a
somewhat higher rate in establishments with 100 or
more employees than in smaller establishments.
A ccelera ted dem and m ay aid diffusion

Industry observers generally expect that demand for
machine tools will remain strong. Whether this means
that skilled labor shortages will persist is arguable.
Skilled workers who have been laid off because of slow
business in key metalworking industries such as auto­
mobiles may be available. But, because average hourly
wages in these industries are often higher than those in
machine tools, it may be difficult for the machine tool
industry to attract such workers. Hence, incentives for
technological advances in the machine tool establish­
ments may remain fairly strong. Therefore, unless the
machine tool industry also suffers from slow business,
productivity should improve at somewhat higher rates
than the long-term rates reported here.
Continued high levels of demand for machine tools



93

strong demand (which justifies the investment) and re­
current labor shortages. Also, as new generations of
managers, engineers, and technicians enter the industry,
numerical control and other computer-related methods
will be more widely applied. The costs of these systems
are likely to fall; hence, they will become more widely
diffused.3
2
Although some manufacturing industries use un­
manned machining systems,3 demand is likely to be
3
small for them. It would not be feasible financially for
the machine tool industry to use such complex systems

— downtime being very expensive.3 Thus, the “un­
4
manned factory” cannot be envisioned for the machine
tool industry; its manufacture by this industry, however,
can be.
It is said that automotive engine plants rely heavily
on the machine tool industry for advances in their pro­
duction equipment.3 In turn, the machine tool industry
5
increasingly relies on electronics and the computer for
its technological advances. Electronics and computers
will likely be dominant in machine tool production in
the years ahead.

1The machine tool industry consists of machine tools, metal cutting
types (SIC 3541) and machine tools, metal forming types (SIC 3542)
as designated in the Office of Management and Budget’s 1972 Stan­
dard Industrial Classification Manual.
2In this article, metalworking industries conform with those includ­
ed in the American Machinist inventory of metalworking equipment
and include the furniture industry (SIC 25); primary metals industry
(SIC 33); fabricated metal products industry (SIC 34); nonelectrical
and electrical machinery industries (SIC 35 and 36); transportation
equipment industry (SIC 37); precision instrument industry (SIC 38);
and miscellaneous manufactures industry (SIC 39).
According to BLS data, average annual rates of change of the out­
put of some major metalworking or other major machine tool using
industries moved as follows:

1 McGraw-Hill Encyclopedia o f Science and Technology, vol. 13,
2
(McGraw Hill, 1977), p. 692.
1 Iron Age, Aug. 8, 1976, p. 166.
,1
1 Machine Design.
4
1 Iron Age, Aug. 8, 1976, p. 200.
5
1 Iron Age, Aug. 8. 1976, p. 174.
6
1 Iron Age, Aug. 8, 1976, p. 189.
7
1 Iron Age, Aug. 8, 1976, p. 256.
8
1 L.T.C. Rolt, A Short History o f Machine Tools (Cambridge, The
9
M.I.T. Press, 1965), p. 223. American Machinist, Sept. 2, 1974.
2 Information obtained from an industry representative.
0
2 Agis Salpukas, “Computerizing machine tools,” The New York
1
Times, June 5, 1980, p. D2.
2 Information obtained from industry representatives.
2

Industry and sic
number
1959-66
Household furniture, 251 . .
Steel, 331 .............................
Copper rolling and drawing, 3351 ...........................
Aluminum rolling and
drawing, 3353,4,5.............
Metal cans, 3411...................
Construction machinery,
3531 ...................................
Electric motors and generators, 3621 ...................
Household appliances,
3631,2,3,9 ........................
Motor vehicles, 371 ...........
11973-78

Output
1966-73

4.8
5.1

21American Machinist, December 1978.
2 Expenditure data for the machine tool industry available to 1978
4
only. Data are from the 1980 Economic Report o f the President. D e­
flators are for private nonresidential fixed investment.

1973-79

4.5
0.7

1.9
-0.8

0.1

-0.5

9.4
2.8

6.7
3.9

10.7

2 Manufacturing Technology— A Changing Challenge to Improve
6
Productivity (W ashington, General Accounting Office, 1976).

5.8

3.3

1.5

7.4

-1.2

1.4

7.6
9.1

2.8
4.8

2.7
4.8

2 “A paucity of new models means layoffs and toolmaking plant
7
closings, while continual changes, such as those that occurred during
the mid-sixties, signal exciting mechanical challenges, full work force
utilization, and extended overtime premiums . . .” H.E. Arnett and
D .N . Smith, The Tool and Die Industry, Problems and Prospects (Ann
Arbor, The University of Michigan Graduate School of Business A d ­
ministration, 1975), p. 18. Estimated retooling cost from Facts and
Figures 1980, (Detroit, M otor Vehicle Manufacturers Association
1980), p. 5.

7.2

-

2 Of the leading industrial countries, the United States has the
5
smallest percentage of machine tools in service less than 10 years.
Even so, the actual number of such tools in the United States was
803,000 in 1976-78, nearly half again as many as in Germany and Ja­
pan. (See American Machinist, December 1978.)

2.2

3Metalworking Machinery, Current Industrial Reports, Series MQ
35-W (U.S. Department of Commerce, various issues).
4 Donald N. Smith and Larry Evans, Management Standards fo r
Computers and Numerical Controls (University of Michigan, 1977).
5 “Fewer, more productive machines: The 12th American Machinist
Inventory of Metalworking Equipment, 1976-78.” American Machin­
ist, December 1978, pp. 133-43.
6 L. Mackay, and R. Leonard, “NC and Conventional Manufactur­
ing Systems— A General Comparison.” Proceedings o f the 18th
International Machine Tool Design and Research Conference (London,
The Macmillan Press, 1978), pp. 651 ff.
1American Machinist, December 1978.
8 “The Machine Tools That Are Building America,” Iron Age, Aug.
8, 1976, p. 269.
9Iron Age, Aug. 8, 1976, p. 271.
1 Iron Age, Aug. 8, 1976, p. 274.
0

2 Iron Age, Mar. 17, 1980, p. 37.
8
2 Iron Age, Nov. 20, 1978, p. 75 ff.
9
3 John Duke, “Construction machinery industry posts slow rise in
0
productivity,” Monthly Labor Review, July 1980, pp. 33-36.
3 A Short History o f Machine Tools. See especially pp. 147— and
1
48,
223. See also David F. Noble, “Social Choice in Machine Design: The
Case of Automatically Controlled Machine T ools,” in Andrew Zimbalist, ed., Case Studies on the Labor Process (N ew York, M onthly Re­
view Press, 1979), pp. 18-50.
3 A. Harvey Belitsky, “M etalworking Machinery,” in Technology
2
and Labor in Five Industries (Bureau of Labor Statistics, forthcoming).
3 Iron Age, Dec. 17, 1979.
3
3 American Machinist, December 1979, p. 82.
4
3 William J. Abernathy, The Productivity Dilemma, Roadblock to
5
Innovation in the Automobile Industry (Baltimore, The Johns Hopkins
Press, 1978), p. 61.

" Iron Age, Aug. 8, 1976, p. 165.




94

Nonwool yam mills experience
slow gains in productivity
During 1958-80, new equipment and techniques
aided productivity growth; although the 2.3-percent rate
o f increase was less than fo r manufacturing as a whole,
it accelerated during the last 8 years o f the period
Ja m e s D . Y o r k

As measured by output per employee hour, productivity
in the nonwool yarn mill industry increased at an aver­
age of 2.3 percent during 1958-80, somewhat below the
2.8-percent rate for all manufacturing.1 (See table 1.)
Output increased at an average annual rate of 4.5 per­
cent while employee hours advanced at a rate of 2.1
percent. For the most recent period, 1973-80, produc­
tivity has risen at a faster annual rate— averaging 3.0
percent. Improved preparatory and spinning equipment
have contributed to these gains.
Growth varied over the period of study. From 1958
to 1965, productivity increased every year, rising at an
average annual rate of 5.2 percent. The largest jump oc­
curred in 1961 with a rise of 9.3 percent. The 5.2-percent average gain in productivity reflected an average
annual growth of 6.7 percent in output and 1.5 percent
in employee hours. Since 1965, productivity gains have
slowed considerably. During 1965-73, output per em­
ployee hour grew at an average annual rate of only 1.2
percent. Output increased at a 4.6-percent rate—just
slightly faster than that of 3.4 percent for employee
hours. Productivity movements displayed much year-toyear fluctuation during this time. There were increases
in only 5 of 9 years, with the largest— 7.1 percent— oc­
curring in 1971.
In contrast to productivity movements for most in­
dustries, the growth in this industry accelerated during*
James D . York is an econom ist in the Division of Labor Force
Studies, Bureau of Labor Statistics.

Reprinted from the
M onthly L abor Review, March 1982.




95

1973-80, rising at an average annual rate of 3.0 percent.
Output grew at a rate of 2.6 percent, while employee
hours declined at a rate of 0.4 percent. Recessionary
conditions in the economy in 1974 and 1975 had a
strong impact on the trends in output and employee
hours. In 1974, the yarn industry began sharp reduc­
tions in employee hours, as output fell 3.5 percent. The
more than proportional drop in employee hours of 7.9
percent led to a 4.8-percent rise in productivity. In
1975, output posted a further decline of 4.2 percent. In
the face of this continuing deterioration in output, em­
ployee hours experienced their largest single-year de­
cline in the entire 1958-80 period, 15.7 percent. The
resulting productivity increase, 13.6 percent, was the
largest of the two-decade period.

Employment and plant size
Total employment in the spun yarn industry in­
creased by more than 28 percent between 1958 and
1980, rising at an average annual rate of 2.1 percent.
There were 67,800 employees in 1958, but by 1980 the
total had risen to 86,900. However, the increase in em­
ployment was not steady; cyclical patterns were evident
throughout the period, which were related to trends in
the overall economy.
The establishments which produce yarn vary in size
but, generally, are rather large. According to the 1977
Census of Manufactures, nearly 40 percent of all estab­
lishments employ 100 to 249 employees and these ac­
count for more than 30 percent of industry value of

population changes, housing starts, changes in lifestyle
or consumer tastes, and general economic conditions.
Non wool yarn is purchased by many different manu­
facturers. Broad woven fabric mills are major users of
spun yarn. These mills produce goods made from cot­
ton, synthetic fibers, and silk, such as sheets, pillow­
cases, draperies, and towels. The firms which use
synthetic fibers and silk are the largest purchasers of
spun yam. From 1963 to 1977, purchases by non wool
spinning mills increased nearly 90 percent, but those by
broad woven cotton mills changed very little.2
Mills which produce knit fabric also account for a
large proportion of total spun yam purchases. These
mills knit tubular or flat fabric and dye or finish knit
fabric; their output increased rapidly from 1963 for­
ward. This increase in output translated into rising yarn
purchases by these mills. It is estimated that between
1963 and 1967, purchases of spun yam by both circular
and warp knit fabric mills increased by approximately
136 percent.
Other types of knitting mills also use spun yarn, in­
cluding knit outerwear and underwear mills and hosiery
mills. The first type manufactures products such as
suits, slacks, shirts, neckties, and skirts. Although com­
plete information is not available for all years, estimates
indicate that purchases of spun yam by knit outerwear
mills decreased during the 1963-77 period.
Exports have historically accounted for a negligible
portion of the total market. In 1979 and 1980, although
exports increased rapidly, they only accounted for ap­
proximately 2 percent of yam shipments. Imports have
had little impact on the domestic market, making up
less than 1 percent of apparent consumption in recent
years.3

T a b le 1. P r o d u c tiv ity a n d r e ia te d in d e x e s f o r n o n w o o l
y a m m ills , 1958-BO
[1977=100]

j
Output per
employee hour

Output

Employee hours

Employees

1958 .............
1959 .............
1960 .............

59.5
62.3
65.3

39.4
44.8
42.5

66.2
71.9
65.1

72.9
72.9
68.6

1961
1962
1963
1964
1965

.............
.............
.............
.............
.............

71.4
74.7
76.3
80.6
84.6

45.1
48.7
50.3
57.4
66.5

63.2
65.2
65.9
71.2
78.6

65.6
66.3
66.3
69.1
73.8

1966
1967
1968
1969
1970

.............
.............
.............
.............
.............

81.8
77.5
80.2
84.5
84.3

66.7
61.7
68.2
74.4
70.9

81.5
79.6
85.0
88.0
84.1

78.0
80.1
83.3
89.3
86.3

1971
1972
1973
1974
1975

.............
.............
.............
.............
.............

90.3
91.0
85.0
89.1
101.2

82.0
91.2
88.9
85.8
82.2

90.8
100.2
104.6
96.3
81.2

89.2
96.4
102.7
101.2
87.7

1976
1977
1978
1979
1980

.............
.............
.............
.............
.............

93.5
100.0
104.2
103.9
106.1

89.7
100.0
103.8
99.0
97.5

95.9
100.0
99.6
95.3
91.9

97.7
100.0
100.3
95.8
93.4

Year

!

Average annual rates of change (in percent)
1958-80 .......
1975-80 .......

2.3
1.7

4.5
3.4

2.1
1.7

2.1
0.7

shipments. Of the 456 establishments in the industry,
almost 20 percent employ 250 to 499 employees and
also account for more than 30 percent of total value of
shipments. Only about 7 percent of all establishments
employ 500 to 999 employees but these produce more
than 20 percent of industry value of shipments.
Production workers have always represented a high
proportion of total industry employment and that pro­
portion has changed very little over time. In 1958, they
accounted for slightly more than 94 percent of total em­
ployment and in 1980 their share was still about 92 per­
cent. The proportion of female employees has increased
in recent years, rising from approximately 44 percent of
the work force in 1972 to 46 percent in 1980.
Average hourly earnings in the spun yarn industry
have remained well below those of all manufacturing. In
1972, the first year for which such data are available,
average hourly earnings were $2.53, significantly less
than the $3.82 for all manufacturing. By 1980, average
hourly earnings in the industry had risen about 89 per­
cent to $4.78. However, this was still well below the av­
erage for all manufacturing, which was up to $7.27.

Advances in technology
The production of spun yam involves a number of dif­
ferent operations. Improvements in technology have
taken place at different stages of the production process
and have contributed to the industry’s overall growth.
Much of the advance has resulted from gradual im­
provements in the equipment over time.
The raw material arrives at the mill as bales. The
adoption of automatic bale opening and blending equip­
ment, which eliminates the need for manual perfor­
mance of this operation, has led to greater efficiency in
this initial stage of the production process.
Likewise, improvements have occurred in the carding
operation. In it, the fibers of the raw material are made
parallel to each other and most of the foreign matter is
removed. The fibers emerge in a form known as sliver.
Formerly, the yam entered a picker which formed it
into a roll before being fed into the carding machine.
However, the introduction of the automatic chute feed
which permits the blended fibers to be fed directly into

Diverse industry markets
Spun yam is used for the manufacture of the great
majority of textile products; household items which
contain yarn include carpets and mgs, bedspreads,
draperies, and towels. Its demand can be influenced by



96

rapid productivity growth. During 1958-66, the rate of
increase in output per employee hour was 4.6 percent.
Both productivity and capital expenditures posted in­
creases in all but one year of the 1958-66 subperiod.
The tremendous growth in capital expenditures dur­
ing this time caused the ratio of capital expenditures per
employee to go up far more rapidly than for all manu­
facturing. In 1958, capital spending per employee was
only $229 in the spun yarn industry, compared with
$619 for all manufacturing. However, by 1966, capital
spending per employee in the industry had risen to
$1,368, compared with $1,112 for all manufacturing.
From 1966-79, the trends in both capital expendi­
tures and productivity were quite different from the pre­
ceding years. The rate of increase in capital expend­
itures declined to 4.9 percent. There were even
substantial decreases in a number of years. Productivity
growth likewise experienced a slowdown, dropping to
an average annual rate of increase of 2.2 percent. As in
the case of capital expenditures, there were declines in
some years. The decline in the growth rate of capital ex­
penditures caused the rate of increase in the ratio of
capital expenditures per employee to fall behind that of
all manufacturing. From 1966-79, capital spending per
employee in the spun yarn industry advanced at an av­
erage annual rate of only 3.0 percent, compared with a
rate of 9.2 percent for all manufacturing. Consequently,
capital spending per employee was only $1,607 in 1979,
compared with $3,118 for all manufacturing.

the carding machine, has eliminated the need for a picker. The carding machinery itself has increased in speed,
further contributing to productivity gains.
The drawing and roving operations follow the
carding process. The drawing operation makes the sliv­
ers more uniform. In the roving process, a twist is
imparted to the sliver by the roving frame. This results
in greater strength. The product that emerges is known
as roving and is wound onto bobbins which are taken
up when full. The adoption of larger bobbins has re­
duced the amount of tending necessary because they do
not have to be removed as often. Improved roving
equipment has been introduced which is faster and
eliminates the need to remove the flyers (which insert
twist into the fibers) for doffing (removal of the bob­
bins).
After the roving operation, the roving bobbin pro­
ceeds to the spinning operation. The spinning drafts
(draws out) the fibers to size the yam and puts the de­
sired twist into it, providing necessary strength. Yarn is
spun onto bobbins; the use of larger ones on spinning
machines has reduced the frequency with which bobbins
have to be removed. The introduction of automatic doff­
ing equipment— equipment which removes the full bob­
bins and replaces them with empty ones—has also im­
proved productivity. Increased operating speed of the
spinning equipment itself has also contributed to pro­
ductivity gains.
After spinning, the yarn is taken to the winding de­
partment. Here, winding machines remove the yarn
from the spinning bobbins and wind it onto cones for
direct customer shipment or onto tubes for dyeing. Au­
tomation in winding equipment has taken place and has
increased productivity growth immensely.4
A number of plants have introduced open-end spin­
ning, which has also aided productivity. This combines
into one process the three separate operations of roving,
spinning, and winding, thus eliminating the need for a
separate roving and winding operation. Open-end spin­
ning can wind the yarn onto a package rather than a
spinning bobbin; thus, a separate winding operation to
transfer the yarn from the spinning bobbin to a cone is
no longer needed.5

s h o u l d c o n t i n u e to increase as im­
provements in production equipment take place and as
more manufacturers take advantage of these. Some of
the newer equipment, which embodies more advanced
technologies than past models, is capable of producing
better quality yam with fewer imperfections and weak
spots. This top quality is increasingly demanded by cus­
tomers as they adopt higher speed weaving and knitting
machinery.6 This is accelerating the adoption of more
modern production equipment by nonwool yarn mills.
Some industry officials expect labor market condi­
tions to provide added stimulus for use of improved
production equipment.7 Relocation of the manufacturing
operations of many industries into major textile produc­
ing areas exerts additional pressure on existing labor
and wages. This, in turn, is impelling more yam manu­
facturers to utilize the equipment and techniques which
provide the greatest levels of output per employee hour.
Open-end spinning will continue to contribute to pro­
ductivity gains as it becomes more widely adopted. This
form of spinning has a particularly favorable effect on
production efficiency because, as noted earlier, it com­
bines the separate operations of roving, spinning, and
winding into a single process.

P r o d u c t iv it y

C apitol spending

Rises in labor productivity are frequently linked to
increases in capital formation. Data available through
1979 indicate that, over 1958-79 as a whole, currentdollar new capital expenditures rose at an average annu­
al rate of 9.4 percent. However, the advances were not
uniform throughout the period, and the most rapid ones
took place in the earlier years. From 1958 to 1966, new
capital expenditures increased at a 22.1-percent rate.
This acceleration in capital spending coincided with



97

‘ See McAllister Isaacs III, “Winding a 138 Percent Boost in Oper­
ator Pounds,” Textile World, February 1980, pp. 79-82.
5 See Brenda V. Lloyd, “Meeting the Challenges of Modernization,”
Textile Industries, September 1979, pp. 114-17. Also, see McAllister
Isaacs III, “Avondale Open-End Cuts Labor, Ups Output,” Textile
World, May 1980, pp. 63-66.
6W. Bud Newcomb, “U.S. Sales-Yam Firms Are Poised For Fu­
ture Growth,” Textile World, September 1980, p. 203.
7 See Douglas A. Bowen, “Linn-Corriher: Yam Making Pioneer,”
Textile Industries, March 1980, p. 50. Also, see Joseph L. Lanier Jr.,
“Plants and Equipment,” America’ Textiles, June 1976, p. 21.
s

1The non wool yam mill industry consists of establishments primar­
ily engaged in spinning yam wholly or chiefly by weight of cotton,
synthetic fibers, or silk. It is designated as industry 2281 in the 1972
Standard Industrial Classification (SIC) Manual. All average annual
rates of change are based on the linear least squares trend of the loga­
rithms of the index numbers. Extension of the indexes will appear in
the annual BLS Bulletin, Productivity Measures fo r Selected Industries.
2The discussion of yam purchases from nonwool yam mills by the
consuming industries is based on constant-dollar estimates.

3 U.S. Industrial Outlook (U.S. Department of Commerce, 1981),
p. 402.




98

The office furniture industry:
patterns in productivity
Product proliferation and short production runs
limited the use o f laborsaving equipment
in office furniture establishments; as a result,
productivity grew only moderately during 1958-80
J. E d w

in

H

enneberger

ing at much higher rates in the wood component (7.2
percent and 5.5 percent) than in metal (4.6 percent and
2.8 percent).
The metal office furniture industry, which experienced
five output downturns between 1958 and 1980, was,
nevertheless, able to maintain productivity growth in all
but 2 of these years. This suggests that the industry’s
work force is flexible and can be rapidly reduced if in­
dustry sales are declining. However, the wood office fur­
niture industry was never able to maintain positive
productivity during the six declines in output from 1958
to 1980. The more highly skilled work force, utilizing
craftworkers, in the wood segment may be more diffi­
cult to periodically layoff and rehire.

Productivity growth (as measured by output per em­
ployee hour) in the office furniture industry1 has been
low, in large part because of relatively short production
runs engendered by product proliferation. Between 1958
and 1980, the industry posted an average annual pro­
ductivity gain of 1.8 percent, substantially below the 2.8percent rate for all manufacturing industries. The gain
resulted from growth in output of 5.5 percent, annually,
and employee hours of 3.6 percent.
In many industries, declines or small gains in output
are associated with reduced or even negative growth in
productivity. This seems to be true of the office furni­
ture industry as a whole. (See table 1.) Thus, in the 9
years in which output either declined or grew at a less
than average rate, productivity either fell or grew at a
less than average rate in 5 of these years.
The trend in productivity for the overall office furni­
ture industry must be viewed in light of the underlying
trend movements of the two component industries—
wood office furniture and metal office furniture. Metal
furniture is the dominant industry in the office furniture
group, employing about two-thirds of the 53,000 work­
ers and accounting for roughly the same percent of
shipments. Although both industries exhibited nearly
the same growth in productivity between 1958 and 1980
(1.7 percent for wood furniture and 1.8 percent for met­
al furniture), the growth in output and employee hours
was more diverse, with both output and hours grow-*

Productivity trends have varied
The industry’s long-term productivity growth can be
divided into three periods (table 1). From 1958 to 1966,
productivity grew at a rate of 3.6 percent annually.
Slowing dramatically, productivity growth advanced by
only 0.1 percent per year during the middle time span
— 1966 to 1975. However, from 1975 to 1980, the rate
of advance increased to 5.1 percent per year.
Recession-induced falloffs were particularly acute
from 1966 to 1975. During the 1970 recession, industry
output dropped 17 percent while employee hours were
reduced by 6.6 percent. Consequently, productivity in
1970 fell by more than 11 percent. During the 1974-75
recession, output declined 5.3 percent in 1974 and 17.7
percent in 1975 while productivity posted its largest
falloff in 1974 ( - 8 .3 percent). More recently, produc­
tivity exhibited positive growth during the short reces-

J. Edwin Henneberger is an econom ist in the Division of Industry
Productivity Studies, Bureau of Labor Statistics.

Reprinted from the
M o n t h l y L a b o r R e v i e w , December 1982.




99

industry’s output growth. Some of these factors have in­
cluded the amount of available office space, growth of
the white-collar work force, replacement demand, and
the introduction of new products.
The most important factor influencing the long-term
growth of office furniture undoubtedly has been the
growth of the white-collar or office work force. Between
1958 and 1980, white-collar workers have grown from
about 27 to nearly 53 million. Currently, officeworkers
account for slightly more than one-half of the total
employed work force.2This translates into a 2.9-percent
average annual increase. Available office space also is a
determinant of office furniture demand. The amount of
public and private detached office space doubled be­
tween 1958 and 1980.3
As the stock of existing office furniture grows, the de­
mand for replacement of womout or obsolete equip­
ment grows also. The data suggest that in recent years
roughly one-third of office furniture production has
been consumed by the replacement market.4
The introduction of new products also stimulates in­
creased demand for office furniture. In the past, office
furniture usually consisted of desks, chairs, tables, and
storage equipment, sold as individual pieces. Now,
modular or systems furniture is sold as complete inte­
grated packages that include movable partitions, storage
components, and service modules. Advantages claimed

Table 1. Productivity and related indexes for the office
furniture industry, 1958-80
[1977 = 100]

Year

Output per
employee
hour

Output

All
employee
hours

Employees

1958 ..........................
1959 ..........................
1960 ..........................

64.0
69.8
70.4

33.1
37.5
39.4

51.7
53.7
56.0

51.8
52.9
54.7

1981..........................
1982 ..........................
1983 ..........................
1964 ..........................
1985 ..........................

72.5
74.4
75.9
82.1
84.2

38.4
42.1
45.6
50.8
57.5

53.0
56.6
60.1
61.9
68.3

51.8
55.8
58.7
60.0
64.9

1966 ..........................
1387 ..........................
1988 ..........................
1989 ..........................
1970 ..........................

86.7
86.5
85.2
88.0
78.2

67.9
69.7
70.9
81.4
67.6

78.3
80.6
83.2
92.5
86.4

74.7
78.2
78.7
88.9
82.7

1971..........................
1972 ..........................
1973 ..........................
1974 ..........................
1975 ..........................

83.9
91.8
90.6
83.1
85.5

64.8
82.7
87.5
82.9
68.2

77.2
90.1
96.6
99.8
79.8

74.9
87.3
94.4
98.9
81.8

89.7
100.0
100.1

75.8
100.0
108.1

107.3
108.9

125.9

84.5
100.0
108.0
112.9
115.6

85.6
100.0
107.8
110.9
118.4

1976
1977
1978
1979
1980

..........................
........................
..........................
..........................
..........................

121.1

Average annual rates of change

1958-80
1958-66
1988-75
1975-80

....................
....................
....................
....................

1.8
3.6

0.1
5.1

5.5

8.4
1.4
13.9

3.6
4.6
1.4
8.3

3.8
4.1
2.0
8.0

sion in 1980. However, this gain in productivity (1.5
percent) was somewhat less than the industry’s long­
term growth (1.8 percent per year).
Among the component industries, the same midterm
pattern of productivity slowdown is evident. (See table
2.) From 1958 to 1966, productivity advanced in both
industries at about 3.6 percent per year. But from 1966
to 1975, productivity fell at an annual rate of 1.1 per­
cent in the wood component while advancing by only
0.5 percent per year in the metal furniture industry.
Rebounding from the recession-marked middle period,
productivity advanced sharply from 1975 to 1980 in the
wood and metal industries— 7.2 and 3.8 percent, re­
spectively. Output in this recovery period was up sharp­
ly in both industries, paced by the nearly 22-percent
average annual growth in wood furniture. Lagging
somewhat behind wood furniture, the output of metal
furniture increased by about 10 percent per year during
this later period, as market share was lost to the more
natural look and feel of wood.

.

Tab!® 2
P r o d u c tiv ity index© © f o r th@ o f f ic e fu rn itu r®
a n d t w o c o m p o n e n t, 1©5@-80
[1977=100]
All office
furniture




64.0
69.8
70.4

67.1
69.5
68.0

64.5
71.6
72.7

1961
1962
1963
1964
1965

.......................................
.......................................
.......................................
.......................................
.......................................

72.5
74.4
75.9
82.1
84.2

70.5
69.9
80.4
84.5
82.8

74.7
77.9
75.9
82.9
86.3

1966
1967
1968
1989
1970

.......................................
.......................................
.......................................
.......................................
.......................................

86.7
86.5
85.2
88.0
78.2

85.9
88.1
87.7
91.9
83.9

88.3
87.6
86.2
88.5
78.0

1971
1972
1973
1974
1975

.......................................
.......................................
.......................................
.......................................
.......................................

83.9
91.8
90.6
83.1
85.5

81.2
84.5
78.5
83.0
80.5

86.4
96.7
97.9
84.5
88.9

1976
1977
1978
1979
1980

Between 1958 and 1980, output of the office furniture
industry grew at an average annual rate of 5.5 percent
per year, substantially above the 3.8-percent average
rate for all manufacturing industries. A number of fac­
tors have shaped the demand for office furniture and the

Metal
furniture

1958 .......................................
1959 .......................................
1960 .......................................

Gffle© furniture demand growing

Wood
furniture

.......................................
.......................................
.......................................
.......................................
.......................................

89.7
100.0
100.1
107.3
.108.9

81.9
100.0
100.7
110.7
109.2

94.8
100.0
99.9
104.8
108.6

Average annual rates of change
1958-80
1958-66
1966-75
1975-80

ICO

.................................
.................................
.................................
.................................

1.8
3.6
0.1
5.1

1.7
3.5
-1.1
7.2

1.8
3.6
0.5
3.8

turing office furniture increased from 289 in 1958 to 486
in 1977—most of this growth occurring in the wood
furniture segment. At the same time, the proportion of
industry shipments accounted for by the four largest
companies in each industry increased modestly.
Between 1975 and 1980, the average annual growth
in capital expenditures per employee was lower for the
office furniture industry than for all manufacturing. For
example, from 1958 to 1975, capital expenditures per
employee grew at an annual rate of 6.3 percent in office
furniture, while the all manufacturing rate over the
same time period was 7.5 percent. Productivity growth
over this period was also lower in the office furniture in­
dustry than in all manufacturing. From 1975 to 1980,
however, capital expenditures per employee accelerated
to 29.6 percent per year, compared with a rate of 11.1
percent for all manufacturing. Productivity from 1975
to 1980 increased sharply also, growing at a rate of 5.1
percent. The level of expenditures per employee, howev­
er, has been substantially less than all manufacturing.
In 1980, the office furniture industry expended roughly
$2,900 per employee for new capital equipment while
the all manufacturing average was almost $3,700.

for systems furniture include design flexibility, more ef­
ficient use of floor space, low rearrangement costs, and
built-in electrical outlets. In recent years, systems furni­
ture has outpaced the growth of conventional office fur­
niture. Currently, systems furniture accounts for about
20 percent of the total office furniture market. Comput­
ers and word processors, which require support furnish­
ings, have also resulted in increased demand for office
furniture.

Industry employment more than doubles
The number of employees in the office furniture in­
dustry increased from 23,000 in 1958 to about 53,000 in
1980. Sustained expansion of the work force during the
1960’s accounted for much of this growth.
While the overall employment growth for the indus­
try was 3.8 percent per year from 1958 to 1980, em­
ployment trends varied among the subindustries. The
work force in the wood office furniture industry
expanded at an average of 6.0 percent per year. The
metal furniture industry grew at less than half of that—
2.8 percent per year.
Compared with other manufacturing industries, office
furniture production is relatively labor intensive. About
10 percent more production worker hours are needed to
generate $1 of added value in office furniture than in all
manufacturing. Among the component industries, wood
office furniture is the most labor intensive.
Production workers accounted for 79 percent of total
industry employment in 1980, down slightly from the 81
percent reported in 1958. About 25 percent of the indus­
try’s workers in 1980 were women, slightly less than the
31 percent level for all manufacturing. Average hourly
earnings of production workers— $5.92 in 1980— were
somewhat below that of the all manufacturing rate of
$7.27. Over the long term, employee turnover has been
slightly below that of the all manufacturing rate.

Manufacturing innovations limited
Typically, production in the office furniture industry
takes place at mechanized work stations with workpiece
transfer accomplished by conveyor line, forklift truck,
or handcart. The wood furniture industry employs gen­
eral purpose woodworking machinery such as saws,
planers, glue presses, and sanders. Basic operations in
the metal furniture industry include metal cutting,
stamping, welding, and tubeforming. With minor differ­
ences, both industries have common operations such as
painting and upholstering. Obviously, many of the pro­
cesses used for manufacturing wood furniture bear little
resemblance to those used for metal furniture. However,
even within the component industries, variations in
equipment and processes are evident. This is particular­
ly true of wood furniture. Some of the finer grades are
produced almost entirely by hand, while the less expen­
sive grades are produced in assembly line fashion.
Product proliferation is a problem within the office
furniture industry, and this has hindered the introduc­
tion of special purpose and highly efficient machinery
and equipment. While the household furniture industry
finds it relatively easy to drop product lines and styles,
office furniture companies must maintain the capacity to
produce old as well as new product lines. This problem
is particularly acute in the more expensive wood office
furniture lines. Reorders of wood furniture must match
style as well as wood grain pattern and color (which
may not be the same as when the pieces were new).
Therefore, the potential number of product types,
styles, and colors, coupled with the bulkiness of furni­

Imdissfry establishment size Increasing
Although office furniture production is geographically
dispersed throughout the United States, there is a large
concentration of firms in Ohio, Indiana, Illinois, Michi­
gan, and Wisconsin, with many plants clustered in and
around Grand Rapids, Mich. Until World War II, the
Grand Rapids area had been a major center for house­
hold furniture. After the war, the household furniture
industry dispersed, and commercial and office furniture
manufacturers moved in to fill the void.
From 1958 to 1977, the number of establishments in
the industry has been growing. In the wood segment,
the number of establishments more than doubled, while
in metal furniture, the number increased by only 25 per­
cent. For the industry as a whole, average employment
size per establishment increased by about 12 percent.
During the same period, companies primarily manufac­



101

ture, discourage factories from accumulating large in­
ventories of finished goods. Most office furniture, per­
haps as much as 90 percent, is for order rather than
inventory. Office furniture dealers do not stock large in­
ventories either; rather, an accumulation of customer
orders is periodically sent to the factory. This results in
short production runs of individual items.
This diminished ability to control production runs may
be one of the reasons productivity growth in the office
furniture industry has been less than that of the house­
hold furniture industry.5 The office furniture industry
must remain even more flexible in terms of production
capabilities than household furniture manufacturers,
many of whom are also troubled by short, inefficient
production runs and difficulty in incorporating highly
specialized and efficient equipment. Nevertheless, some
notable advances in the technology of manufacturing
office furniture have been introduced.
In the wood office furniture industry, one of the most
pronounced trends in innovation has been increased use
of particleboard. While the primary impetus for the
expanded usage of particleboard has been its lower cost
in relation to the cost of solid lumber, the industry has
focused considerable attention on new technologies to
handle the material. A wide variety of surface laminates
and films and application techniques have eliminated
several time-consuming production and assembly opera­
tions. Groove-folding, a technique whereby V-shaped
grooves are cut in the particleboard substrate, but not
through the flexible surface material, produces seamless
furniture edges which are held in place by the continu­
ous outer wrap.6
Although somewhat hampered by increased petro­
chemical prices in recent years, the use of plastic
materials has simplified construction and added
strength to furniture components, and can also produce
mar-resistant surfaces. Reconstituted wood veneer, an­
other advance in materials, has uniform thickness,
grain, and quality and can be evenly stained. Its use
eliminates the need for the labor intensive procedure of
manually grading, selecting, and removing defects from
natural veneers.
In addition to new materials, notable advances have
occurred in woodworking machinery. Abrasive planing,
introduced in the early 1960’s, combines heavy stock re­
moval with direct dimensioning at the sanding machine.7
Machines which glue and trim veneer strips to the edges
of particleboard can eliminate the complicated set of
clamps and pressure bands which formerly had to be
locked in place until the glue dried.
In the metal office furniture industry, machines have
recently been installed that automatically position and
cut shapes into the large flat metal blanks that later will
be fashioned into desks, file cabinets, and so forth. This
equipment is more efficient because it does not require



102

moving the workpiece to a separate machine for each
cut. Also, setup time is considerably reduced.
Savings in the time needed to produce tubular shapes
have been accomplished by new tubeforming and cut­
ting equipment. Tubemaking, which starts from flat
coiled steel, has been speeded up by the use of automat­
ic welders which join the ends of the coils so that the
tubeforming equipment need not be shut down while
coils are being changed.
Metallic inert gas (mig ) welding has largely supplant­
ed most other forms of welding. Its advantage is that
the parts being joined do not have to be as thoroughly
cleaned as with brazing. Although robot welders are
not common, automatic welding is. Once travel and an­
gle of the welding arm have been adjusted, a worker is
required only to load and unload workpieces onto and
from the equipment.
Although not designed specifically for the metal office
furniture industry, automated parts inventory storage
and retrieval systems are being used by several plants in
the industry. Operating under the control of a computer
which “explodes” or breaks down orders for the re­
quired number of finished pieces of furniture into the
necessary parts demand, robot crawlers and unmanned
forklift trucks retrieve and deliver the parts to various
pickup stations where they are transferred to the assem­
bly line in the correct sequence for manufacture.
Upholstering, an operation which is similar in both
wood and metal office furniture, is a particularly labor
intensive operation and requires a skilled work force.
Although still used in many plants, manual pattern lay­
out and fabric cutting have in some cases been phased
out, superseded by diecutting of fabric. Computer-controlled cutting equipment, which combines high speed
with accuracy and eliminates manual pattern layout, is
also available.8 Steam tables, installed at upholsterers’
work stations, expand the cut fabric workpiece. Once
removed from the steam and stapled around the foam
rubber cushion, the fabric shrinks back to its normal
size and becomes taut. Airpowered plunger tables, used
to compress the fabric-covered foam shape, have made
button insertion and tiedown operations easier.
Electrostatic finishing, used widely by the metal fur­
niture industry, can be used successfully on wooden fur­
niture,9 resulting in increased labor productivity in the
finishing area and a substantial reduction in material
and maintenance costs. Automatic electrostatic spray
lines allow closer spacing of pieces to be painted and,
thus, greater efficiency. With these automatic lines, col­
or changeover is automatic and can be done in 30 sec­
onds rather than the 2 minutes previously required on
the nonautomatic electrostatic lines. Electrodeposition
lines, which are powdered coatings in a medium of ei­
ther air or water, are particularly efficient with respect
to labor, materials, and solvent emissions.

Likewise, both the metal and wood office furniture in­
dustries have shared the advances made in portable,
handheld power fastening tools, resulting in added
worker efficiency through more power, greater capacity,
and less weight and maintenance. Productivity has also
been enhanced by improved workflow layout, compu­
terized recordkeeping, and new materials such as quick­
setting glues and improved finishes.

Recent trends may continue
If continued, the industry’s capital spending surge of
the last few years may provide the plant and equipment
necessary to maintain the recent above average growth in
productivity. However, the current economic downturn
may have a negative effect on demand and productivity.
Although the full consequences of the current eco­
nomic downturn cannot be foreseen, it is worth noting
that previous recessions have had only limited ef­

' The office furniture industry is classified as SIC 252 in the 1972
Standard Industrial Classification Manual and its 1977 supplement, is­
sued by the U.S. Office of Management and Budget. The subindustries
within the office furniture group include establishments that are pri­
marily engaged in manufacturing furniture commonly used in offices—
wood (SIC 2521) and metal (SIC 2522).
2Employment and Training Report o f the President, 1981 Report
(The White House, 1981), pp. 148— see also table 3, p. 73, of the
49;
April 1982 issue of the Monthly Labor Review.
3 See P. W. Daniels, ed., Spatial Patterns of Office Growth and Loca­
tion (New York, John Wiley & Sons, Inc., 1979), pp. 67-69.
4 “Equipment Purchases Planned by Readers in 1980,” The Office,
January 1980, p. 26.
5See J. Edwin Henneberger, “Productivity Growth Below Average
in the Household Furniture Industry,” Monthly Labor Review, Nov­




103

fects on the growth of the white-collar work force, one
of the key factors in the output growth of the office fur­
niture industry. In fact, even though there have been
four recessions since 1958, the total white-collar work
force has never declined. With the forecasted continued
expansion in the white-collar work force,1 demand for
0
the industry’s products should continue to increase and
may, therefore, present the industry with opportunities
to expand productivity. Also, the industry’s output
should be further bolstered if the growth of systems fur­
niture continues.
While the “paperless office” is not as yet a reality,1
1
over the long term, the increasing sophistication of elec­
tronic office equipment may result in officeworkers
becoming more productive. This, in turn, can influence
output of the office furniture industry by dampening
growth in the white-collar work force and affecting de­
mand and productivity in the office furniture industry.

ember 1978, pp. 23-29.
6 Darrell Ward, “Groove Folding for Contract and Contemporary,”
Woodworking and Furniture Digest, June 1981, pp. 42-45.
7 ---- , “Abrasive Planing Challenges Your Knife Cutting Tech­
niques,” Hitchcock's Wood Working Digest, November 1963, pp. 2932.
8 Robert Michael, “New Techniques of Computerized Fabric Cut­
ting,” Furniture Methods and Materials, June 1971, pp. 12-15.
9 Richard D. Rea, “Electrostatic Disks Win,” Woodworking and
Furniture Digest, April 1982, pp. 22-25.
1 Economic Projections to 1990, Bulletin 2121 (Bureau of Labor Sta­
0
tistics, 1982), pp. 34-47.
" See Paul Lieber, “Office Automation: The Job Threat that Never
Happened,” The Office, May 1980, p. 158.

Productivity in the
pump md compressor industry
During 1958-80, the industry experienced
long-term advances, reflecting improvements
in metalworking machinery and computer aid;
but since 1965 productivity has decelerated,
being especially slow from 1973 forward

,

H o rst B r a n d

and

Cl y d e H uffstutler

Output per employee hour in pump and compressor
manufacturing rose at an average annual rate of 2.1 per­
cent between 1958 and 1980—compared with a rate of
2.6 percent for manufacturing as a whole.1 Output in­
creased 4.7 percent a year, employee hours 2.6 percent.
Among the sources of the industry’s long-term produc­
tivity advance were improvements in metalworking ma­
chinery, which lies at the core of the production
processes for pumps and compressors, and computer
technologies, which were increasingly applied to engi­
neering design.
The labor productivity trend for the industry was
marked by strong advances during the early part of the
period (from 1958 to 1965), followed by deceleration
during 1965-73, and a further slowing thereafter. As
the tabulation shows, using average annual rates of
change in percent, the trend pattern paralleled manufac­
turing:

1958-80 ..................
1958-65 .............
1965-73 .............
1973-80 .............

Pumps and
compressors
2.1
3.4
2.1
1.0

By 1980, the level of labor productivity in the industry
had risen 55 percent from 1958, as against 78 percent
for all manufacturing.
The long-term productivity trend, in addition to evi­
dencing divergent medium-term movements, was punc­
tuated by sharp year-to-year swings. These swings were
generally related to the business cycle, although they
show no uniform pattern. Thus, labor productivity fell
steeply in 1960 (3.2 percent), 1975 (5.6 percent), and
1980 (2.6 percent). In these years, output either grew
more slowly than employee hours (1960), or fell more
rapidly (1975), or fell while hours rose (1980). Yet, in
1961, 1967, and 1971, years when the economy slowed,
significant increases in productivity occurred (3.9 per­
cent, 1.7 percent, and 1.5 percent)— which, however,
stemmed from drops in employee hours exceeding drops
in output.
Years of recovery or boom in which productivity
soared to more than twice its long-term rate, displayed
a more uniform pattern of change in output and em­
ployee hours. In 1959 and 1976, gains in productivity
were linked with large output increases but slight em­
ployee hour declines.
Separate data for pumps and pumping equipment,
and for air and gas compressors, are available only
from 1972 forward. Average annual rates of change in
labor productivity for the two separate industries com­
pare as follows for the 1972-80 span:

Manufacturing
2.6
2.7
2.4
1.8*

Horst Brand and Clyde Huffstutler are economists in the Division of
Industry Productivity Studies, Bureau of Labor Statistics.

Reprinted from the
Monthly L abor Reivew, December 1982.




104

Percent
Pumps and compressors ...................................
Pumps and pumping equipment .................
Air and gas compressors ..............................

1.2
1.2
1.1

All manufacturing ............................................

1.9

Reflecting contrasting trends in output and employee
hours, labor productivity movements in the pump and
pumping equipment segment were considerably less vol­
atile than in compressor manufacturing. The former
attained a productivity level in 1979 that exceeded 1973
by 7 percent (both years registered cyclical peaks); the
latter failed to reattain its 1973 high.

Output increases
Pumps and compressors are used throughout manu­
facturing and many nonmanufacturing industries, as
well as agriculture. Pumps are the second most com­
mon machine in use after the electric motor.2 Compres­
sors generate compressed air, which may be regarded as
a form of energy ranking in breadth of use only below
electricity, gas, and water, in addition to being indis­
pensable in the transportation of gas.3
Between 1958 and 1980, output of pumps and com­
pressors rose 175 percent, or at an average annual rate
of 4.7 percent. Manufacturing output grew at a rate of
3.8 percent over the period. Like the long-term trend in
the industry’s labor productivity, the long-term trend in
output rose less after 1965 than earlier, as the following
tabulation indicates by showing average annual rates of
change in percent:
Pumps and
compressors
1958-80 ............
1958-65 ___
1965-73
1973-80

Manufacturing

4.7
6.5
2.5
4.2

3.8
5.9
3.0
2.5

Output of pumps and compressors reached a peak in­
dex level of 115 (1977=100) in 1979, from which it re­
ceded slightly in 1980. The dip was caused by a decline
in compressor manufacturing, which had climbed 51
percent between 1973 and 1979. Pump and pumping
equipment output had risen 22 percent between those 2
years of cyclical highs.
Of the total output of pumps and compressors, the
former accounted for about two-thirds, according to the
1977 Census of Manufactures, the latter for the remain­
ing one-third. Industrial pumps represented more than
half of the output of pumps and pumping equipment
(other than accessories). Hydraulic fluid power pumps,
oil well and oilfield pumps, and other pumps and equip­
ment installed in appliances, fire engines, and structures,
made up the remaining output. Parts and attachments
constituted close to one-quarter of pump manufacturing



105

output in 1977. Given the often difficult climatic and
environmental conditions in which pumps must operate,
and the abrasiveness of fluids often transferred by them,
speedy replacement of worn and damaged parts consti­
tutes a vital function of the manufacturer, and is the
reason for the high proportion of shipments of parts
and attachments.
Air compressors accounted for well over one-quarter
of the shipments of compressor manufacturers, accord­
ing to the 1977 census, gas compressors for just under
one-tenth. They consisted preponderantly of the station­
ary type. Portable compressors, which are relatively
small machines, made up one-fifth of total air and gas
compressor shipments. Industrial spraying equipment
also added one-fifth to compressor manufacturers’ ship­
ments. Compressors, like pumps, are frequently exposed
to rough operating and environmental conditions, hence
a comparatively high proportion of shipments (20 per­
cent) represented parts and attachments in 1977.

Factors underlying output growth
In general, growth in the output of pumps and com­
pressors was related to expansion in industrial and
public utility demand, particularly during the boom
years of the early and mid-1960’s; gains in residential
and associated public works construction, such as sew­
age and waterworks, during the 1960’s and 1970’s; and
intensified needs of energy-related extractive and pipe­
line industries, especially during the 1970’s. Foreign
trade, too, played an important role in sustaining out­
put: about one-fifth of pump and compressor produc­
tion was exported between 1972 and 1978.
Expansion in the productive activities of a wide array
of users lay at the base of output growth of pumps and
compressors. No precise statistical link can be
established between the former and the latter. However,
movements in the plant and equipment expenditures,
adjusted for price changes, by major pump and com­
pressor users are indicative, as are put-in-place data for
construction.
Among large-scale users of pumps and compressors
was the chemical industry, which accounts for about
one-tenth of total pump and compressor output.4 Chem­
icals nearly doubled plant and equipment outlays (ad­
justed for price changes) in the early 1960’s, then
reduced them. After 1973, however, outlays were once
again raised, so that in 1979 they stood nearly twice
above the 1973 level. The industry has increasingly used
pumps made of fiberglas, plastics, and stainless steel to
transfer salt solutions, acid, and chlorine.5
Steel mills and blast furnaces, whose capital spending
patterns compared roughly with that of the chemical in­
dustry over the review period, purchase about 7 percent
of pump and compressor output. They use a variety of
industrial and hydraulic pumps as well as compressors

to move sources of energy such as liquid fuels, as well
as water to absorb waste energy. Installation of multi­
stage pumps to achieve higher pressure has, in part,
been prompted by the shift from open-hearth to basicoxygen and electric-arc steelmaking processes. The par­
tial replacement of slabbing mills by continuous casting
has required more water, hence a larger number of and
more powerful centrifugal pumps.6
More than 18 percent of pumps and compressors are
bought by energy-related extracting, processing, and
distributing industries. Thus, growth in extractive activ­
ities spurred the demand for industrial as well as oil
well and oilfield pumps. Between 1960 and 1970, the
number of crude oil and gas wells drilled dropped
sharply (by nearly two-fifths), as did footage drilled (by
27 percent). After 1970, the decline was reversed; in
1978, the two indicators ran 72 percent and 68 percent
above 1971 levels. Concomitantly, output of oil well
and oilfield pumps, which had risen at an average annu­
al rate of less than 4 percent between 1958 and 1973,
soared to a rate of more than 10 percent between 1973
and 1980. Oil extraction also requires reciprocal pumps
for mud circulation; submersible centrifugal units to lift
the crude oil; and centrifugal pumps for waterflooding
(to prevent subsidence and maintain pressure).7
Compressors are required in oil drilling and oilfield
maintenance operations, and particularly in secondary
recovery efforts. The continued expansion of natural gas
pipelines (whose mileage increased 9 percent between
1973 and 1978) spelled the installation of additional
large compressors for gas transmission; and increases in
new wells—more than twofold between 1973 and 1978
— required numerous smaller compressors for gas gath­
ering, as did the prohibition of flaring of waste gas
(which now must be stored in tanks). Also, steep in­
creases in capital expenditures of the coal mining indus­
try— 162 percent between 1958 and 1972 (after
adjustment for price changes), and 169 percent between
1973 and 1977—indicate expansion in this industry’s
demand for compressors.
Expansion of petroleum pipeline capacity also raised
the demand for pumps, particularly of the high-horse­
power centrifugal kind, and for stationary compressors.
While the network of petroleum pipelines operated by
petroleum pipeline companies increased 16 percent be­
tween 1960 and 1970, and contracted somewhat thereaf­
ter, total oil transported rose 81 percent during the
1960’s, and 48 percent in the 1970’s.8At the same time,
the average diameter of pipes was enlarged by onethird, roughly doubling capacity.9 This required signifi­
cant increases in the size and capacity of pumping
equipment and compressors.
Expanding electrical generating capacity spurred the
output growth especially of centrifugal pumps. These
are used as boiler-feed pumps, as well as in many other



Table 1. Productivity and reiated indexes for pump and
compressor manufacturing, 1958-80
[1977 = 100]
Output per
employee hour

Output

Employee hours

Employees

1958 ...............
1959 ...............
1960 ...............

64.5
68.8
66.6

41.1
43.4
44.6

63.7
63.1
67.0

63.1
62.6
66.1

1961...............
1962 ...............
1963 ...............
1964 ...............
1965 ...............

69.2
73.6
78.1
79.4
80.9

43.9
48.6
51.7
59.0
65.2

63.4
66.0
66.2
74.3
80.6

62.6
64.9
64.6
71.5
77.9

1966
1967
1968
1969
1970

...............
...............
...............
...............
...............

81.1
82.5
82.3
86.3
85.8

70.5
70.1
68.3
74.0
74.8

86.9
85.0
83.0
85.7
87.2

82.9
82.4
80.3
83.3
85.8

1971
1972
1973
1974
1975

...............
...............
...............
...............
...............

87.1
91.1
97.8
96.7
91.3

69.4
76.2
87.7
94.0
87.0

79.7
83.6
89.7
97.2
95.3

79.4
82.5
88.6
97.3
96.0

1976
1977
1978
1979
1980

...............
...............
...............
...............
...............

96.8
100.0
102.6
102.5
99.8

91.4
100.0
107.1
114.5
112.9

94.4
100.0
104.4
111.7
113.1

94.3
100.0
105.1
112.7
113.9

Average annual
rates of change
(in percent):
1958-80 .........
1975-80 .........

2.1
1.9

4.7
6.0

2.6
4.1

2.7
4.2

Year

operations requiring the circulation and condensation of
steam and water. While the total number of electrical
generating stations did not advance very much over the
review period, the proportion of stations generating
500,000 kilowatts or more rose from under 3 percent in
1960 to 12 percent in 1979. Nuclear and gas-turbine
driven, power-generating plants likewise increased. The
rise in the number of larger electric generating plants
spelled a shift to larger, more powerful pumps.1
0
Construction accounts for another 18 percent of
pump and compressor output. Centrifugal and trash
pumps (which accommodate up to 25 percent of small
solids in the water being pumped) are used in the clear­
ing and preparing of construction sites.1 Portable com­
1
pressors are indispensable in the many pneumatical
operations at construction sites. Between 1960 and
1973, the volume of total construction put in place rose
at an average annual rate of 2.9 percent; thereafter it
declined at a rate of 1.5 percent. However, some con­
struction sectors with high demand for pumps continued
to expand— for example, sewage system construction
(spurred by more stringent environmental regulations).

Employmemt

sum !

hours

Employment in the pump and compressor manufac­
turing industry currently numbers approximately 91,000
persons. It rose 81 percent between 1958 and 1980, or
at an average annual rate of 2.7 percent (compared with
1.1 percent for all manufacturing).
106

2.9 percent a year (versus 2.7 percent). Nonproduction
workers account for a comparatively high proportion of
the industry’s employment—41 percent in 1980, as
against 30 percent for all manufacturing. The propor­
tion did not change significantly over the review period.
One of the reasons for the high proportion of
nonproduction workers resides in the larger share
accounted for by mechanical engineers in the industry
groups’ occupational makeup (the data are, again, for
the general industrial machinery group). Such engineers
represented 6 percent of all white-collar workers in the
group in 1980— three times the comparable manufac­
turing ratio. Engineering and science technicians,
among them drafters, made up 11 percent of white-col­
lar workers in the group, as against 8 percent for manu­
facturing. The group also employed a somewhat higher
proportion of clerical and secretarial workers (42 versus
40 percent). The share of blue-collar nonproduction
workers, such as truckdrivers and service employees,
was generally lower than for manufacturing.

The long-term trend in employee hours in the indus­
try did not differ significantly from the long-term trend
in employment. They rose at a rate of 2.6 percent a
year over the period, compared with 1.1 percent for
manufacturing as a whole.
Production worker employment rose somewhat faster
over the 1958-80 period than production workers’
hours (2.7 percent a year versus 2.4 percent). Year-toyear changes ranged from an increase of 12 percent in
1974 to a decline of 10 percent for production worker
employment; the range was wider still for hours. Over­
time exceeded the manufacturing durables average in 17
of the 22 years examined here.1 Comparatively high
2
overtime hours were probably related to hiring and sep­
aration policies which, judging by the pertinent labor
turnover data, have been such as to ensure retention of
a relatively skilled work force. Labor turnover in the in­
dustry ran less than three-fifths of the manufacturing
average for the period.1 High overtime and low turn­
3
over rates were probably also related to the skill com­
position of the industry’s work force.
Data on the skill composition of employees in pump
and compressor manufacturing are not directly avail­
able. Such data have been compiled by the BLS only for
the general industrial machinery group (sic 356), of
which pumps and compressors represent 29 percent by
employment. Craft and related workers accounted for
30 percent of the production workers employed by es­
tablishments in this group in 1980, compared with 26
percent for total manufacturing. Metalworking craftworkers represented 12 percent of all production work­
ers in the group, compared with 5 percent for manufac­
turing; and machinists 3 percent, compared with 1
percent. Operatives accounted for slightly more than
three-fifths of all production workers in the general in­
dustrial machinery group, the same as in manufacturing
as a whole. But metalworking operatives in industrial
machinery, constituting one-third of production work­
ers, had three times the share of their counterparts in
all manufacturing. Laborers, with 6 percent of produc­
tion workers in the group, had little more than half
their share for all manufacturing.
Wage differentials also suggest a somewhat higher
skill composition for production workers in pump and
compressor establishments than in all manufacturing. In
1980, hourly earnings of the former ran 10 percent
above the manufacturing average, and 3 percent above
the manufacturing durables average. These ratios re­
mained substantially unchanged during 1958-80. (Hour­
ly earnings were about the same for production workers
in the industry and in the general industrial machinery
group of which the industry is part.)
Employment of nonproduction workers by pump and
compressor manufacturing establishments rose at a
slightly faster rate than that of production workers—



Technological changes
Small lot production is the rule in pump and com­
pressor establishments. Pumps and compressors are
often large machines, manufactured to customer specifi­
cation. While many of these machines are composed of
standard parts, the economies associated with mass pro­
duction are generally not available in producing pumps
and compressors. The production process must con­
stantly be adapted so as to cope with the many design,
casting, and machining requirements that arise. Such
adaptation was facilitated by the advent of numerically
controlled machine tools in the 1960’s, and the intro­
duction of computer-aided design into engineering prac­
tice.1 Numerical controls and computer-aided design
4
have been important sources of labor productivity ad­
vances in the industry. The impact of these technologi­
cal changes will be outlined, following a brief survey of
the kinds and age of the metalworking machinery used
in manufacturing pumps and compressors.
According to the 12th American Machinist Inventory
of Metalworking Equipment for 1976-78 (latest avail­
able), about one-third of all metal cutting and metal
forming machine tools in the pump and compressor
manufacturing industry were less than 10 years old; 70
percent were less than 20 years old. Comparable data
for earlier years are available only for the general indus­
trial machinery group (sic 356). For general industrial
machinery, no distinct trend in the age composition of
metalworking machinery is observable. Thus, in 1958,
34 percent of such machinery installed in the plants of
this group was less than 10 years old, 74 percent was
less than 20 years. In 1968, as well as in 1978, the com­
parable figures read 33 and 72 percent.1
5
Despite the absence of a trend toward a more mod107

ern stock of metalworking equipment in terms of age,
output capability per machine tool unit improved con­
siderably. According to the American Machinist's 10th
Inventory of Metalworking Equipment (1968), “For the
last 5 years, the number of machine tools has increased
by 4.5 percent, while the value of production, as meas­
ured in constant dollars by the American Machinist pro­
duction index, has gone up by 39 percent.” In the text
accompanying its 12th Inventory (1976-78), the Ameri­
can Machinist again confirmed this trend. It noted that
while the total machine tool “population” had declined
by about one-tenth between 1968 and 1978, the produc­
tion index had risen 40 percent.1
6

Table 2. Productivity and related indexes for pumps and
pumping equipment manufacturing, 1S72-80
[1977 = 100]
Year
1972
1973
1974
1975
1976
1977
1978
1979
1980

......................
......................
......................
......................
......................
......................
......................
......................
......................

Average annual rates of
change (in percent):
1972-80 .................

Output per
employee hour

Output

Employee hours

Employees

90.8
94.1
93.6
89.9
92.7
100.0
101.1
100.7
97.2

81.0
91.7
91.9
90.4
92.9
100.0
106.1
111.6
112.5

89.2
97.4
98.2
100.6
100.2
100.0
104.9
110.8
115.8

88.1
94.8
98.3
101.4
99.7
100.0
106.2
113.0
117.6

1.2

3.9

2.6

3.1

MacMmmg time eet
The increase in the output capacity of machine tools
has undoubtedly contributed to gains in the labor pro­
ductivity of pump and compressor manufacturing. For
example, machining time for pump casings, which often
are of great weight and size, has in the leading plants
been drastically reduced by specially designed milling
machines. These milling machines also require less setup
time, and a smaller number of setups than formerly. In
one case, machining time for large centrifugal pump
casings, weighing up to 18,000 pounds, was reduced
from 48 to 17 employee hours; in other words, where
three 16-hour shifts, involving two operators, were re­
quired earlier, only one operator working 17 hours is
needed now.1 However, electric energy requirements are
7
considerably greater.1
8
Reductions in machining time are frequently achieved
by combining in one large metalworking operation sev­
eral previously separate ones. An example is the simul­
taneous milling, radial drilling, and facing (smoothing)
of different parts of the same workpiece. Sequential op­
erations on a given workpiece are speeded up by means
of automatic tool changers, commanded by taped in­
structions, causing different kinds of tools (or different
configurations of the same kind of tool) to be advanced,
retracted, and changed, as programmed. (Such appara­
tus may be bypassed by manual controls, when neces­
sary in the operator’s judgment.)1
9
Reductions in setup time have also been made
possible for many single-purpose machines, for example,
grinders. Pump shafts must in some cases be tapered,
and this has usually required several setups depending
upon the length and desired fit of the shaft. In some of
the industry’s plants, separate setups for this purpose
have been eliminated by grinders that adapt automati­
cally and will grind several fits simultaneously.2
0
Advances in the foundry operations of pump and
compressor manufacturers have also contributed to la­
bor productivity gains. In the technically more ad­
vanced plants, molding and coremaking have been
speeded up by rapid-cycle machinery, and by discarding



the time-consuming sand baking process. The no-bake
process uses a resin binder and a catalyst to produce
the sand mold, saving energy as well as unit labor re­
quirements.2 Several of the same core patterns (from
1
which pump casings and other pump and compressor
parts are cast) can be cut simultaneously by means of
synchronous fabricating machinery, operating on the
principle of key-making apparatus.
Engineering plays a key role in pump and compressor
manufacturing. As noted, much of the industry’s output
is manufactured to customer specifications, which of ne­
cessity involves engineering staff. Additionally, the ad­
vent of numerically controlled machine tools, and of
computer numerical controls, has centered more pro­
duction responsibilities in engineering departments,
away from the shop floor. The growth of engineering
staff has intensified concern with promoting its ef­
ficiency. Engineering efficiency has been raised in the
more advanced establishments of the industry by apply­
ing certain computer technologies; designing production
processes which economize on engineering time; and
standardizing common parts. Efforts have also been
made to bypass engineering where feasible.2
2
Computer graphics have simplified drafting by
allowing corrections to be made to the draft without
manually redrawing it. Detailed drawings can be made
within minutes, where before it took hours. Comput­
erized data banks permit access to all drawings on file.
Computer graphics has permitted the elimination of 7
to 8 drafter jobs in one of the establishments visited by
BLS staff. The computer-aided design can be program­
med directly upon tape, and fed to the machine tool.
This represents a considerable advance for numerical
controls, inasmuch as programs previously had to be
punched, or prepunched programs had to be purchased.
With design and production closely linked, owing to
the computer and numerical controls, engineers con­
ceive of computer-aided design and computer-assisted
manufacturing as integral operations. Calculation of for­
mulae, design of the product, and production are
108

viewed and operated as a single process. Uniformity of
product dimension and quality are ensured. Changes in
the detail of design are quickly and inexpensively incor­
porated. Engineering time saved by computer-aided de­
sign and computer-assisted manufacturing has been
estimated at two-thirds of conventional engineering pro­
cedures.2
3
As noted, replacement of parts and attachments ac­
counts for a sizable proportion of the output of pump
and compressor manufacturing. Computer-aided design
and computer-assisted manufacturing helps ensure that
replacement parts are dimensionally accurate, while
economizing on engineering time. Dimensional confor­
mance is further ensured by certain process innovations.
Thus, cores or molds for impellers and other pump and
compressor components are now frequently ceramic in­
stead of wood.
Capital expenditures
Plant and equipment outlays by pump and compres­
sor manufacturers rose at an average annual rate of 8.1
percent between 1958 and 1980—compared with 4.9
percent per year for all manufacturing. (The expenditure
data underlying these rates have been adjusted for price
changes.2) The industry’s capital spending rose at a
4
particularly high rate during the 1960’s, nearly tripling
between 1958 and 1969. For a few years thereafter, such
spending receded from the 1969 level, but it resumed its
rise in 1972, and doubled between 1972 and 1980.
Comparable figures for all manufacturing are considera­
bly more modest, as the tabulation shows (average an­
nual rates in percent):

1958-80 ...........
1958-69 . . . .
1969-80 ___
1969-72 . . '
1972-80 . .

Pumps and
compressors
8.1
12.2
6.9
-10.2
7.6

Manufacturing
4.9
8.2
4.6
-3.0
5.4

[1977 = 100]
Output per
employe® hour

Output

Employee hours

Employees

...............
...............
...............
...............
...............
...............
...............
...............
...............

92.1
108.8
103.0
96.7
108.4
100.0
105.5
106.0
105.7

66.7
79.8
98.2
80.5
88.4
100.0
109.1
120.3
113.7

72.4
74.7
95.3
85.0
83.1
100.0
103.4
113.5
107.6

71.6
76.6
95.3
85.3
83.8
100.0
102.8
112.2
106.6

Average annual
rates of change
(in percent):
1972-80 .........

1
.1

6.5

5.4

5.2

Year
1972
1973
1974
1975
1976
1977
1978
1979
1980

with 100 workers or more represented less than onequarter of the total number of establishments in the in­
dustry but well over four-fifths of total employment and
value of shipments.
Concentration was high. The industry’s four largest
companies employed more than half of its workers in
1977, and accounted for half of its value of shipments.
For manufacturing as a whole, the comparable ratios
were 6 and 7 percent.
Even so, the establishments are mostly small,
employing fewer than 100 persons. The smaller plants
accounted for 79 percent (pumps) and 70 percent (com­
pressors) of all industry establishments in 1977. At the
same time, however, they recorded only 14 and 9 per­
cent of total industry employment. These relationships
had not changed much from earlier phases of the review
period.
Outlook
Continued advances in the labor productivity of
pump and compressor manufacturing are likely over the
longer term. The diffusion of numerically controlled ma­
chine tools and computer-aided design within the indus­
try’s establishments, as well as among them, has still
some way to go. The age distribution of metalworking
machinery should continue to favor higher-capacity,
modernized equipment. Organizational changes result­
ing from a widening scope of computer applications—
for example, more centralized decisionmaking in refer­
ence to machining processes—will probably also im­
prove productivity.2
6
So far, robots appear not to have been introduced
widely. Even in the more advanced shops, they are used
chiefly for paint spraying and other marginal opera­
tions. Industry observers, however, expect that robots,
as their costs decline, will handle workpieces more and
more during the noncutting portion of the work cycle.2
7
Such a development is also bound to raise labor pro­
ductivity.

Strasftmr® of the industry
In 1977, pumps and pumping equipment were
manufactured in 613 establishments, air and gas com­
pressors in 175. The former had increased 10 percent
since 1972, the latter had more than doubled. In the
preceding 9 years, no change in the number of estab­
lishments making pumps and compressors had oc­
curred. The number of companies in the industry
owning these establishments barely changed during the
1970’s.2
5
Pumps and compressors are manufactured mostly in
larger plants. Five percent of all establishments in the
industry employed 45 percent of its workers in 1977,
and accounted for about the same proportion of the to­
tal value of shipments. More generally, establishments



Table 3. Productivity and reiated indexes for air and gas
compressor manufacturing, 1S72-80

m

The nearer-term outlook is somewhat clouded, how­
ever. The industry’s output is likely to suffer from
weakened demand from major users of pumps and com­
pressors. When output slackens, a slowed rate of pro­
ductivity advance, even declines in the rate, are more
probable. A source of weakened demand is the stagna­
tion in housing starts, which tends to diminish the need
for pumps and compressors used in construction, as
well as for such public works as water and sewage,
which often require pumps and related equipment on a
large scale. Another source of declining needs for (hence
output of) pumps and compressors are reductions in
projected increases in oilfield exploration and develop­
ment. (These reductions have been linked to smallerthan-expected energy demand increases, and lessened
price pressures.)2
8
At the same time, the widespread concern with cut­
ting energy costs may bolster the demand (and output)
of more energy-efficient pumps and compressors. For
example, variable displacement pumps may to some ex­
tent replace fixed displacement pumps. The latter rejects
excess flows by means of a relief valve, dumping them
back into a reservoir. This wastes pump energy, which a
variable displacement pump can avert.2 Piston pumps,
9
furthermore, are thought by industry observers to be
also favored over fixed displacement pumps, as high-

pressure hydraulics is more widely adopted in industry
and transportation (especially in aircraft and mobile
equipment). High-pressure hydraulics permits the use of
lighter pipes, pumps, and actuators.3
0
Industry observers believe that pumps and equipment
of larger size will continue to be installed in such uses
as steampower generation, pipelines, and petroleum re­
fining. The shift from gasoline to heavy fuel refining3
1
requires heavier rotary rather than lighter centrifugal
pumps. Slurry pipelines— which move water-suspended
solids such as coal and wood chips—are believed to
gain wider acceptance, because they offer important
economies in transportation.
The BLS has projected a somewhat faster rise in the
number of nonproduction workers than production
workers for the general industrial machinery group.3 In
2
1990, professional and technical workers will make up
12.4 percent of all of the group’s employees, according
to the projections, compared with 11.4 percent in 1980;
and the share of clerical and related workers will rise
slightly. The proportion of craftworkers will remain
unchanged, and that of operatives will edge downward.
It seems reasonable to assume that changes in occupa­
tional pattern projected for the general industrial ma­
chinery group will, by and large, be repeated by pump
and compressor manufacturing.

1The pump and compressor manufacturing industry consists of two
segments, pumps and pumping equipment, designated as SIC 3561 of
the Standard Industrial Classification Manual 1972 of the Office of
Management and Budget; and air and gas compressors, SIC 3563.
SIC 3561 consists of establishments primarily engaged in manufactur­
ing pumps and pumping equipment for general industrial use. Meas­
uring and dispensing pumps for gasoline stations are not included,
nor are pumps installed in automobiles. SIC 3563 consists of estab­
lishments primarily engaged in manufacturing air and gas compres­
sors for general industrial use. Refrigeration compressor units are not
included. Prior to 1972, pumps and compressors were classified to­
gether in SIC 3561.
Average annual rates of change are based on the linear least
squares of the logarithm of the index numbers. Extensions of the in­
dexes will appear in the annual BLS Bulletin, Productivity Measures
for Selected Industries.
2William C. Krutzsch, “Introduction and Classification of Pumps,”
in Igor I. Karassik and others, Pump Handbook (New York,
McGraw-Hill, 1973), p. 1 ff.
3John P. Rollins, ed., Compressed Air and Gas Handbook (New
York, Compressed Air and Gas Institute, 1973), p. 1. The range of
compressed air uses are discussed on pp. 1-44.
4 U. S. Department of Commerce, Bureau of Economic Analysis,
The Detailed Input-Output Structure o f the U S. Economy: 1972
(Washington, D.C., Government Printing Office, 1979).
5John R. Birk and James H. Peacock, “Chemical Industry,” Pump
Handbook, p. 10-74 ff. Also, conversation with industry observer.
6 E. R. Pritchett, “Steel Mills,” Pump Handbook, p. 10-159; and
telephone conversation with author.

ty Measures for Selected Industries, 1954-80 (Washington, D.C., Gov­
ernment Printing Office, 1982), table 179.
9 Mary Vickery, “Petroleum Pipeline Transportation,” U.S. Depart­
ment of Labor, Bureau of Labor Statistics, Technological Change and
its Labor Impact in Five Energy Industries, BLS Bulletin 2005 (Wash­
ington, D.C., Government Printing Office, 1979), pp. 39 and 42.
1 Telephone conversation with Krutzsch, an author of Pump Hand­
0
book.
" Benjes, H.H., “Sewage,” Pump Handbook,
telephone conversation with author.

10-2.

Also,

1 Overtime in pump and compressor manufacturing compared with
2
overtime for all of manufacturing durables (all manufacturing = 100)
as follows:
Pumps and compressors
1958 ...........
1959 ...........
1960 ...........
1 9 6 1 ...........
1962 ...........
1963 ...........
1964 ...........
1965 ...........
1966 ...........
1967 ...........
1968 ...........
1969 ...........
1970 ...........
1 9 7 1 ...........
1972 ...........
1973 ...........

7Elvitsky, Pump Handbook.
8U. S. Department of Labor, Bureau of Labor Statistics, Productivi-




p.

110

____
____
____
____
____
____
____
........
.. ..
____
.. . .
.. . .
____
____
. . . .
____

63
111
104
83
96
90
103
113
123
114
103
103
110
93
103
105

Pumps
1974
1975
1976
1977
1978
1979
1980

..
..
. .
..
..
. .
..

.
.
.
.
.
.
.

Compressors

126
119
131
114
100
103
100

112
115
103
116
113
114
161

1
5
Labor turnover in pump and compressor manufacturing com­ versity of Michigan, 1977). See also John Duke and Horst Brand,
“Cyclical Behavior of Productivity in the Machine Tool Industry,”
pared with manufacturing (all manufacturing = 100) as follows (data
Monthly Labor Review, November 1981, pp. 27-34.
from 1972 forward for pumps and pumping equipment only; data for
air and gas compressors are not available):
1 William H. Parker, “Cutting time out of pump machining,”
7
American Machinist, January 1979, pp. 112-13.
Accessions
Separations
1 “The Machine Tools that are Building America,” Iron Age, Aug.
8
1958 ........................................
52
66
30, 1976, p. 163. According to the report, electric horsepower require­
1959 ........................................
69
56
ments for lathes rose from 150 in the 1950’s to 400 to 600 in the
1960 ........................................
55
65
1970’s. Many other examples are also cited in the article.
1961 ........................................
54
58
1 Observation of industry operations. See also Iron Age, cited
9
1962 ........................................
56
51
above.
1963 ........................................
56
56
2 Observation of industry operations.
0
1964 ........................................
63
46
2 Observation of industry operations. See also Richard W. Lyon,
1
1965 ........................................
60
58
“Foundries,” in U.S. Department of Labor, Bureau of Labor Statis­
1966 ........................................
68
63
tics, Technology and Labor in Four Industries, Bulletin 2104 (Washing­
1967 . . . . .............................
57
59
ton, D.C., Government Printing Office, January 1982), p. 12.
1968 ........................................
57
57
1969 ........................................
68
65
2 Industry sources, and observation of industry operations. See also
2
1970 ........................................
60
76
A. Harvey Belitsky, “Major technology changes in metalworking ma­
1971 ........................................
54
56
chinery,” Technology and Labor in Four Industries, pp. 20-33.
2 Industry source.
3
60
49
1972 ........................................
2 Adjustment for price changes was made by using the implicit de­
4
1973 .............................
67
55
flator for nonresidential investment in structures and producers’
71
67
1974 ........................................
durable equipment. See Economic Report o f the President, February
1975 ........................................
51
68
1982, p. 236.
1976 ........................................
59
53
2 U.S. Department of Commerce, Bureau of the Census, General
5
1977 ........................................
63
55
Report on Industrial Organization, 1977 Enterprise Statistics (Washing­
1978 ........................................
51
49
ton, D.C., Government Printing Office, 1981).
1979 ........................................
53
54
1980 ........................................
54
58
2 A. Harvey Belitsky, “Major technology changes,” especially pp.
6
24—
25.
1 See Comptroller General of the United States, Manufacturing
4
2 See American Machinist, June 1980, p. 147 fF
7
.
Technology— A Changing Challenge to Improved Productivity, Report
2 “Biggest U.S. Oil Concerns Likely to React to Glut by Cutting
8
to the Congress, Washington, June 3, 1976, especially p. 37 ff.
1982 Capital Budgets,” The Wall Street Journal, Apr. 7, 1982, p. 7.
1 The Eighth American Machinist Inventory of Metalworking Equip­
5
2 “Curbing the Energy Appetite of Hydraulic Systems,” Machine
9
ment— 1958, New York, McGraw-Hill. Reprinted from the American
Design, June 26, 1980, p. 95.
Machinist, Nov. 17, 1958; The Tenth American Machinist Inventory of
3 “Modem Hydraulic Systems: the Pressure Mounts,” Machine De­
0
Metalworking Equipment— 1968, New York, McGraw-Hill, 1968. The
sign, Jan. 24, 1980, p. 81 ff.
data cited for pump and compressor manufacturing from the 12th In­
3 Rose Zeisel and Michael D. Dymmel, “Petroleum refining,” Tech­
1
ventory are based on unpublished printouts.
nological Change and Its Impact in Five Energy Industries, p. 26.
1 American Machinist, December 1978, p. 135. The reduction in
6
3 See the articles on the Bureau’s projections in Monthly Labor
2
machining time is confirmed in Donald N. Smith and Larry Evans,
Review, August 1981, pp. 9-42.
Management Standards for Computers and Numerical Controls (Uni­




11J

Output per unit of labor input
in the retail food store industry
Productivity, as measured by output
per hour o f all persons, increased
2A percent annually during 1958-75,
because o f industry structural changes
and some technological improvements
Jo h n l . C a r e y

and

Ph y l l is F l o h r O t t o

Output per hour of all persons in the retail food in­
dustry increased at an average annual rate of 2.4
percent from 1958 to 1975, compared with a gain of
2.3 percent for the nonfarm business sector.1 This
growth reflect increases of 2.4 percent in output
and 0.1 percent in hours.
Growth in output per hour of all persons has
been influenced by a trend to fewer and larger
stores serving a growing population, a correspond­
ing decline in the number of small stores, as well as
some changes in technology and store operations
designed to increase efficiency. The decline in the
number of small stores has led to decreases in the
number of partners, proprietors, and unpaid family
workers in the industry, which has counteracted the
increase in the number of paid employees. Hours of
paid employees grew at an average rate of 1.9 per­
cent from 1958 to 1975, offsetting the decline of 4.5
percent per year in the hours of the self-employed
and unpaid family workers.
There are widely disparate growth rates in the
two components of the retail food store industry.
From 1958 to 1975, the output of grocery stores*
John L. Carey and Phyllis Flohr Otto are economists in the Division of
Industry Productivity Studies, Bureau of Labor Statistics. Robert S.
Robinowitz, an economist formerly with the Division, and Mark Segal,
a student, assisted with the research for the article. This study was fi­
nanced in part by the National Center for Productivity and Quality of
Working Life.
Reprinted from the
M onthly L abor Review, January 1977.




112

(which comprises the largest portion of the industry
and includes supermarkets as well as small grocery
stores) increased at an average rate of 2.8 percent
each year. Specialty food stores (such as bakeries
and meat markets) averaged only 0.7 percent each
year. The number of specialty stores continued to
decline as they lost business to supermarkets, which
offer similar products with the convenience of onestop shopping. On the other hand, supermarkets
have lost some business to the convenience food
stores. Although many items are higher priced,
these stores are conveniently located and provide
fast service.
Efficiency in the retail food industry has been
aided by improvements in displaying merchandise
and electronic innovations, such as computerized
scales and cash registers. Also contributing to effi­
ciency is the trend toward prepackaging merchan­
dise at the distributor, cutting and packaging poul­
try at a centralized location, and price marking at
the warehouse.
Trends in output per hour
During 1958-72, output per hour showed in­
creases, ranging from 1.2 to 5.5 percent, every year
except 1969, when a decline occurred. (See table 1.)
During 1958-72, the average annual rate of increase
was 3.0 percent. (In the nonfarm business sector the
rate for the same period was 2.6 percent.) However,
as shown in the following tabulation, the 1973-74

price levels, increased about 10 percent from
1958-71, though it has dropped each year since
then.3
Because most products can be packaged in differ­
ent sizes, packaging practices by the m anufacturer
have an effect on the utilization of labor in retail
food stores. For example, economy size packages
result in a larger volume being sold in a single
transaction. In a similar fashion, disposable cans
and bottles have largely superseded deposit bottles
and eliminated much handling time. However, the
growing concern with environmental and conserva­
tion factors may reverse this trend.
In 1973 and 1974, as the economy moved into a
recession, coupled with large increases in food
prices, output per hour fell in retail food stores—
5.1 percent in 1973, and an additional 2.1 percent
in 1974. O utput fell 3.3 and 0.5 percent, respec­
tively; despite this drop, hours rose in both years,
1.8 percent in 1973, and 1.6 percent in 1974. Food
prices, as measured by the Consumer Price Index
for food at home, soared 16 percent and 15 percent
in 1973 and 1974, higher than increases in per cap­
ita disposable income (12 and 8 percent). Am ong
the reasons for the increase in hours (despite the
drop in output) was the continued industry trend
tow ard Sunday openings, the longer hours of opera­
tion during the week, and the growth of serviceoriented operations in supermarkets.
As the economy began to improve in 1975, out­
put per hour in the retail food store industry re­
corded a gain of 2.7 percent— output grew 0.8 per­
cent and hours declined 1.9 percent, as hours of
employees fell slightly and the hours of partners,
proprietors, and unpaid family workers continued
their historic movement downward. Per capita dis­
posable income rose 9 percent in 1975 as com pared
with the slightly lower gain of 8 percent in food
prices.
Employment and hours
The total num ber of persons working in retail
food stores increased 21 percent from 1958-75,
while total hours increased less than 1 percent. This
disparity can be attributed directly to the growth of
part-tim e help, along with the decline in the num ­
ber of partners, proprietors, and unpaid family
workers in the retail food industry. The average
weekly hours for nonsupervisory workers fell from
36.3 in 1958 to 32.3 in 1975, even though there was
a general trend to longer store hours and Sunday
sales. In June 1966, part-tim e help accounted for
over 40 percent of the work force in retail food.4
Today, one survey shows that p art-tim e help ac­
counts for over 50 percent of the work force in
superm arkets.5

recession slowed the growth to an average annual
rate of 2.4 percent for 1958-75:
O utput p e r hour
o f a ll persons
1 9 5 8 -7 5 ................
1 9 5 8 -7 2 ................

2.4
3.0

O utput

H ours o f
a ll persons

2.4
2.8

0.1
- .2

In the overall economy, the period 1958-72 was
m arked by increasing per capita income, relatively
stable prices (the Consumer Price Index rose at an
average annual rate of 2.7 percent), and increasing
per capita food consumption. Retail food store out­
put increased every year. From 1958-72, people
continued to switch from small food stores to su­
perm arkets and the average superm arket became
larger, as firms closed their small, inefficient opera­
tions. The num ber of retail food stores fell 26 per­
cent from 1958 to 1972. Grocery stores with sales
greater than $500,000 a year increased their share
of total retail food store sales from 51 percent in
1958 to 70 percent in 1972.2
Supermarkets accounted for alm ost 76 percent of
all grocery store sales and about 70 percent of all
retail food store sales in 1972. Today, nearly 9,000
items are sold in superm arkets— a 50-percent in­
crease since 1960. This vast num ber of items, along
with the procedures that are necessary to order,
stock, and m arket them have acted to limit the
gains in output per hour. However, this has been
offset to some extent by an upward trend in the size
of the average custom er transaction. A lthough pre­
cise figures are not available, estimates indicate that
the average superm arket sale, adjusted for changing

T a b le

1.

O u t p u t p e r h o u r o f a ll p e r s o n s a n d

r e la te d

d a ta ,

re ta il f o o d s to r e s , 1 9 5 8 -7 5
[1 9 6 7 = 1 0 0 ]

Y es

O perhour
utput
of all persons

O
utput

H of
ours
all persons

1 9 5 8 ....................................................................
1 9 5 9 ....................................................................
1 9 6 0 ....................................................................

75.4
78.4
80.9

78.4
81.9
84.1

104.0
104.4
103.9

196 1 ....................................................................
1 9 6 2 ....................................................................
1 9 6 3 ....................................................................
1 9 6 4 ....................................................................
1 9 6 5 ....................................................................

84.0
85.3
89.4
91.4
93.8

86.1
88.0
88.7
93.0
96.4

102.5
103.2
99.2
101.8
102.8

196 6 ....................................................................
196 7 ....................................................................
196 8 ....................................................................
196 9 ....................................................................
1 9 7 0 ....................................................................

96.3
100.0
105.1
104.8
110.5

98.0
100.0
104.6
105.6
111.7

101.8
100.0
99.5
100.8
101.1

197 1 ....................................................................
197 2 ....................................................................
19 7 3 ....................................................................
1974P..................................................................
1975P..................................................................

111.9
113.3
107.5
105.2
108.1

114.1
116.8
112.9
112.3
113.2

102.0
103.1
105.0
106.7
104.7

p= p in
relim ary.




113

The 21-percent increase in the num ber of persons
engaged in retail food sales consists of a 54—
percent
increase in the num ber of paid employees (nearly
2.0 million in 1975) offset by a 52-percent decrease
in the number of partners, proprietors, and unpaid
family workers (272,000 in 1975). Partners, propri­
etors, and unpaid family workers now account for
only about 12 percent of the total number, as com­
pared with 31 percent in 1958. Some of the decrease
in the num ber of partners and proprietors may be
due to changes in business organization. (For exam­
ple, a num ber of firms may have incorporated and
the partners or proprietors may have become paid
employees, but there are no specific data on the
change.) It is believed, however, that most of the
decline was due to the decrease in the num ber of
single-store firms— a 34-percent decline from 1958
to 1972.
Retaining experienced personnel is a m ajor prob­
lem confronted by retail food stores. Labor turn­
over is extremely high. In 1968, the overall separa­
tion rate for superm arkets was over 60 per 100
employees.6 (In comparison, the separation rate for
the m anufacturing sector is around 4 to 5 per 100
employees.) In 1971, the highest turnover among
full-tim e superm arket employees was in cashiers
(14 separations per 100) and m eatcutters (10 sepa­
rations per 100). Turnover in store m anagers was
negligible.7 The high turnover rate among store em­
ployees is one of the factors hindering the gain in
output per hour in this industry, as new employees
m ust be trained and, therefore, are not as produc­
tive during the training period as more experienced
persons.
One factor that might be contributing to the high
turnover is the low hourly earnings in retail food.
In 1975, wages in retail food stores were 13 percent
below the hourly average for the total private nonagricultural sector, and 18 percent lower than the
m anufacturing average.

all retail food stores with employment was 10 in
1972. More than 30 percent of the retail food stores
are operated entirely by proprietors or partners
and, therefore, have no paid employees.
Most superm arkets are in suburban locations—
frequently in shopping centers. Eighty percent of
the superm arkets opened in 1971 were located in
shopping centers, com pared with 47 percent in
1958. Developers of shopping centers are usually
willing to lease sites only to firms with high finan­
cial ratings, making it difficult for small grocery
stores to lease space.8
To generate traffic and increase sales within the
store, many superm arkets have installed large spe­
cialty food and nonfood departm ents. Bakeries, del­
icatessens, pharmacies, and liquor departm ents are
common. Although longer hours and Sunday sales
are the most visible changes, services, such as check
cashing, money orders, film processing, and cater­
ing services are now provided by many superm ar­
kets.
Consumers choose to shop in superm arkets
rather than small, proprietor-run stores for many
reasons. Small stores frequently cannot compete
with large stores on price and variety of selection
because their low sales volume and small space
limit volume discounts. In addition, many chains9
own some of their own production facilities, or
carry private label goods which usually are offered
at a lower price than widely advertised, national
brand name goods. Some independent stores are
able to overcome price problems by joining or
forming a wholesale affiliation group (voluntary or
cooperative chain), which can obtain volume dis­
counts for all the member stores. Frequently, they
also carry some types of private label goods.
Voluntary chains lack in-store data about trading
areas and customer shopping habits, but have more
flexibility with respect to store locations than cor­
porate chains. Corporate chains operate in a consis­
tent m anner and the use of experimental tests can
be conducted in a controlled environm ent.1
0
The small retailer who does not belong to a chain
generally pays higher prices for merchandise; there­
fore, in order to compete, the retailer m ust be very
efficient or provide a service not readily available in
a supermarket. A significant num ber of speciality
stores and small grocery stores survive by locating
in areas which are not profitable for large retailers
(such as inner city, high-density areas, and large
apartm ent buildings) because some custom ers are
willing to pay extra for service or credit.
A new trend in food retailing is the grow th of
small (2,000-4,000 sq. ft.) convenience food stores.
These stores, located prim arily near residential a r­

C hanges in industry structure

Retail food distribution has been dom inated for
quite some time by superm arkets. Supermarkets
came into existence around 1930 and were well es­
tablished by the beginning of W orld W ar II.
G row th was slowed during the war but increased
rapidly during the postw ar years. Supermarkets
combine the functions of separate establishments
(or specialty stores) which, for the m ost part, are
small and independently owned.
Small specialty stores and grocery stores account
for a large portion of employment, about 30 percent
in 1972. While a superm arket usually employs 25 to
75 people, the average num ber of paid employees in



114

Also affecting the utilization of labor, are trends
tow ard prepackaged produce, precut and packaged
poultry, and some increase in centralized cutting
and packaging of fresh meat. These trends vary
greatly, however, because consumer acceptance has
not been uniform. For example, although much
m ore produce is w rapped by distributors and grow­
ers than in the past, m ost retailers continue to dis­
play, package, and weigh more than half of their
produce within individual stores. On the other
hand, precut and prepackaged poultry has gained
wide consumer acceptance; in fact, much of the
poultry now sold in retail food stores is cut, pack­
aged, and weighed at poultry processing plants.1
3
A t the present time, most meat is cut and pack­
aged at the retail store, after it has been broken into
primal and subprim al cuts at a centralized location.
Only about 3 percent of the firms operating retail
m arkets have centralized the cutting and packaging
of final cuts. Frozen meat, which can be cut, pack­
aged and shipped by the m eatpacking plant, has
not gained sizable consumer acceptance.1
4
O ther changes in technology include the use of
conveyor belts in the warehouse and overhead “rail­
way” systems for unloading and transporting sides
of beef. Central warehouses are used m ore effi­
ciently now because com puters are used for inven­
tory control and space m anagement as well as for
forecasting sales.
Detrim ental to the growth ?n output per hour of
all persons over the period studied was the time
spent by employees issuing stam ps and games to
customers. Although this trend has been replaced
by the use of coupons and foods stam ps by shop­
pers, the effect remains the same— increased check­
out time. New shopper aids, such as unit pricing,
also can slow growth in output per hour during pe­
riods of rapid price changes because of the addi­
tional num ber of labels that have to be changed.

eas, provide the consumer with ease of access, quick
selection, and virtually immediate checkout. Items
requiring extensive labor time, such as fresh meat
and most fresh produce, are not carried. The selec­
tion is much smaller than in a supermarket; in fact,
four products (tobacco, beer and wine, soft drinks,
and dairy products) make up more than one-half of
the sales of one convenience store chain.1
1
Convenience stores grew rapidly during the late
sixties. Their net profits before taxes were 4.8 per­
cent of retail sales in 1974— double that of indepen­
dent superm arkets.1 These stores are supplied
2
largely by outside wholesalers and distributors
rather than by their own warehouses.

Changes in technology and store operations
Technological developments, including changes
in distribution practices, have been a source of im­
provem ent in output per hour in this industry. The
technological changes that occurred during
1958-75 were small improvements, rather than in­
novations that greatly altered store operations.
Am ong such improvements were refrigeration sys­
tems with fast defrost, self-defrosting freezers, spe­
cial surface floors in the meat departm ent that help
to reduce cleanup time, and faster m eat sheers and
grinders. O ther improvements included meat wrap­
ping machines, shrink and heat pressure wrapping
film, and scales that autom atically print a weight
and price tag.
Improvem ents have also been m ade in display
techniques. Produce, for example, is now kept in
refrigerated display cases, eliminating the need to
remove the perishable produce to the back room
each night. Other changes include standup refriger­
ato r cases with doors that face the custom er and
backs that open into a refrigerated storeroom. (A
variation is the roller cart which can be placed in a
refrigerator case direct from delivery.) O ther tech­
niques include the use of cardboard display cases
provided by m anufacturers and “dum p” displays
where goods are simply put in wire baskets in no
particular order.
M ost superm arkets are arranged so that the typi­
cal custom er will have to walk though m ost of the
store to buy groceries. (Exposing the custom er to
the full array of products encourages larger transac­
tions.)
Certain m anufacturing and distribution tech­
niques have also aided growth in output per hour of
all persons within the retail food store. The use of
shrink film (a form of clear plastic wrap) rather
than complete cardboard cartons has decreased the
time spent opening boxes and disposing of waste
cardboard. Shrink film also allows instant identifi­
cation of merchandise.



O utlook
A growing trend in the retail food industry is the
emergence of the com bination or “superstore” with
a large nonfood departm ent. This departm ent may
cover 40 to 50 percent of the total selling area. To­
tal space in these “superstores” is around
40,000-50,000 square feet, or twice the size of the
average superm arket today. One trade association
reported that 9 percent of all superm arkets opened
in 1973 were com bination stores. Conventional
stores, however, still account for over 95 percent of
all superm arkets.1
5
The trend tow ard central, as opposed to retail,
m eat cutting may grow. The prim ary advantage is
the reduction in labor time at the retail level. With
115

central meat cutting, the waste problem is reduced
and retailers can be more specific in their orders.1
6
Electronic cash registers and the uniform product
code are beginning to be used in retail food stores.
W ith this system, a unique, m achine-readable sym­
bol is placed on each product, and is translated into
a price by an optical scanner. The price can be read
on a screen and the data are then transm itted to a
com puter which prints out a detailed receipt for the
customer. In the most prevalent system, the checker
runs the symbol over the optical scanner and bags
at the same time, decreasing checkout time for
shoppers. In addition to eliminating cashier errors,
the com puter can also keep inventory autom atically
and assist in reordering for the store when levels

decline to a predetermined point. Labor savings can
be obtained by eliminating the need to m ark prices
on items, and to take inventory manually. Training
time for cashiers can also be reduced significantly.
The development of scales fo r weighing and simul­
taneously marking meat and produce with the uni­
form product code symbol will assist the diffusion
of this technology. The main obstacle to the imme­
diate use of this system is the high cost of the regis­
ters and computers. Also, there have been some ob­
jections to the lack of price m arkings on individual
items. Retailers may have to continue to m ark
prices, thereby losing some of the labor saving ad­
vantages of this system.

1The study covered paid and unpaid persons (including paid em­
ployees, partners, proprietors, and unpaid family members) working in
retail food establishments. (Data for 1974 and 1975 are preliminary.)
Retail food stores are defined as those establishments primarily en­
gaged in selling food for home preparation and consumption (major
group 54 in the 1972 Standard Industrial Classification Manual). All
average annual rates of change are based on the linear least squares
trend of the logarithms of the index numbers. Current indexes for this
industry are published in the annual BLS bulletin, Productivity Indexes
for Selected Industries.

6 Super Market Industry Speaks: 1969, pp. 25-26.
7 Super Market Industry Speaks: 1972, pp. 15-16.
s See Food: From Farmer to Consumer (National Commission on
Food Marketing), pp. 69-83.
4 The definition of a “chain” differs among industry sources. A chain
is usually considered to be either 4 or more, or 11 or more stores under
some form of common operation or control.
1 Saul B. Cohen, “Location Research Programming for Voluntary
0
Food Chains, ” Economic Geography, January 1961, pp. 1-11.

2 The definition of supermarkets in terms of annual sales volume
differs among industry sources. The amount of $500,000 is used by the
Department of Agriculture and some trade publications. The Super
Market Institute, however, uses a $1 million annual sales volume. Dis­
cussions and data pertaining to that source in this article, therefore,
should be viewed with the difference in mind.

1 See “The Threat to Southland’s Growth,” Business Week, Oct. 28,
1
1972, pp. 60-62.
1 See “Forty-second Annual Report of Grocery Industry,” Progress­
2
ive Grocer, April 1975.
1 See The Chicken Broiler Industry: Structure, Practices, and Costs,
3
Marketing Research Report 930 (U.S. Department of Agriculture,
Economic Research Service), May 1971, p. 34.

3 See annual issues of The Super Market Industry Speaks (Super Mar­
ket Institute, Inc.). This should be interpreted with care since many
factors can influence the size of a transaction—number of trips to the
store per week and shifts to higher (or lower) valued merchandise, for
example.

1 See “Chain Store Age 1972 Meat Study,” Chain Store Age, No­
4
vember 1972, beginning on page 55.
1 Super Market Industry Speaks: 1972, p. 10.
5

4 Employee Earnings and Hours in Retail Food Stores, Bulletin
1584-3, (Bureau of Labor Statistics, 1968).

1 See “Central Cutting: The Only Way to G o,” Chain Store Age,
6
November 1972, beginning on page 72, and “Central Prepackage Meat
Operations Holds Line On Expenses for a Three-Store Organization,”
Progressive Grocer, February 1961, beginning on page 50.

5 See The Super Market Industry Speaks: 1969 {Super Market Insti­
tute, Inc.) p. 25.

APPENDIX:

Measurement techniques and limitations
Thus, those goods which require more retail labor are
given more importance in the output index.
Data on the quantities of goods sold usually are not
available for trade industries, including retail food.
Therefore, real output was estimated by removing the
effects of changing price levels from the current dollar
value of sales. This technique was used at various levels
of aggregation for both grocery stores and for specialty
food stores. Because an adjustment for changing price
levels usually lowers the dollar value, such a series is
usually referred to as a deflated value measure. Output
measures based on deflated value have two major char­

Indexes of output per hour of all persons measure the
change in the relationship between the output of an in­
dustry and the hours expended on that output. An index
of output per hour is derived by dividing an index of out­
put by an index of industry hours.
The preferred output index for retail trade industries
would be obtained from data on quantities of the various
goods sold by the industry, each weighted (that is, multi­
plied) by the employee-hours required to sell one unit of
each good in some specified base period. This concept
also embodies the services associated with moving the
goods from the retail establishment to the consumer.



116

tions. This has shifted some of the hours in retailing
from the employee to the consumer. However, data are
not available to measure the impact of this change. Like­
wise, adjustments could not be made for changes in the
quality of products sold. Such adjustments are implicit,
however, to the extent that changes in quality have been
accounted for in the price indexes used to deflate the cur­
rent dollar value of sales.
In the noncommodity producing sector, there are
many more conceptual problems with the definition of
an industry’s “output” than in most other sectors of the
economy. There are differences of opinion about many of
these concepts. One problem is the definition of quality
change. In the retail food industry, for example, soft
drinks are now sold largely in cans and “no return”
bottles, while previously they were sold in deposit bot­
tles. This can be treated as a factor that aids output per
hour of all persons because cans and “no return” bottles
eliminate much handling time in the store. Such packag­
ing, however, can also be considered as a change in qual­
ity to the consumer, and, therefore, a different product
requiring an adjustment to the measure. In any case,
available data did not permit adjustments to be made for
this change.
Other changes in store operations can also affect the
utilization of labor. The shift to produce, poultry, and
other products that are prepackaged by food processing
industries eliminates much retail labor time. Although
the available data do not allow adjustments to be made
each year, the measure is adjusted periodically through
the use of relative labor importance weights. This re­
duces the possibility for bias in the measure, but does
not, of course, eliminate it entirely.
The basic sources for the output series for this mea­
sure consist of the total sales data and sales by merchan­
dise line data reported by the U.S. Department of Com­
merce. The deflators were developed using various
Consumer Price Indexes published by the Bureau of La­
bor Statistics. The labor importance weights were devel­
oped from data reported by the U.S. Department of
Commerce and the U.S. Department of Agriculture.
The basic sources for the all-persons-hour series con­
sist of data on employment and hours published by the
Bureau of Labor Statistics, supplemented by data re­
ported by the Internal Revenue Service and special tabu­
lations compiled for the Bureau of Labor Statistics by
the Bureau of the Census.

acteristics. First, shifts in sales can occur among prod­
ucts of different value which have the same unit labor
requirements. (For example, if customers begin to pur­
chase more “nationally advertised” brands instead of
store brands, dollar sales will increase if the “nationally
advertised” brand is priced higher.) Thus, a change can
occur in the output per hour index even if the labor re­
quired to sell the merchandise does not change.
Second, the sales level, both in current and constant
dollars, reflects differences in unit values for identical
products sold in different types of establishments. For
example, the unit values associated with a product sold
in a self-service “discount” store may be lower than the
unit value associated with the same product sold in a
store that provides many sales clerks and delivery serv­
ice. The output measure, therefore, reflects changes in
the level of service provided to customers insofar as dif­
ferences in unit values reflect the differences in service
among the various types of establishments.
In addition to the deflated value technique, weights
relating to labor importance (that is, labor cost and em­
ployment) were used to combine segments of the output
index into a total output measure. These procedures re­
sult in a final output index that is closer, conceptually, to
the preferred output measure.
The index of hours for the retail food industry is for
all persons—that is, hours for paid employees, partners
and proprietors, and unpaid family workers. As in all of
the output per hour measures published by the Bureau of
Labor Statistics, hours and employment in retail food
stores are each considered homogeneous and additive.
Adequate information does not exist to separately weight
the various types of labor.
The indexes of output per hour relate total output to
one input—labor time. The indexes do not measure the
specific contribution of labor, capital, or any other single
factor. Rather, they reflect the joint effect of many inter­
related influences such as changes in technology, capital
investment, capacity utilization, store design and layout,
skill and effort of the work force, managerial ability, and
labor management relations.
No explicit adjustments were made to the measure for
retail food to take into account increases or decreases in
service provided to the consumer. With the growth of
supermarkets and convenience food stores, there has
been a continuation of the trend to self-service opera­




117

During 1958-77, annual productivity
increased an average o f 2 .9 percent,
as the industry responded to a strong
dem and fo r soap and detergent products
and was aided by improved technology
P a t r i c i a S. W i l d e r

Productivity in the soaps and detergents industry has
increased in line with the rise in output per employee
hour for the manufacturing sector since 1958.'
While annual output doubled, employee hours in­
creased by more than one-fourth between 1958 and
1977. The average annual increase in productivity was
2.9 percent.
The rise in productivity was associated with an annu­
al increase in output of 4.1 percent coupled with a
1.2-percent average annual increase in employee-hours.
Productivity gains have resulted prim arily from
sustained high levels of capital investment for new ma­
chinery and equipment, and improvements in produc­
tion and packaging operations.
Output per employee hour has fluctuated during the
period of this study. Since 1958, annual increases in
productivity have ranged from 1.0 to 10.6 percent. De­
clines in productivity have occurred in 4 years, includ­
ing 1977. For the most recent 5-year period, 1973-77,
productivity has declined at an annual rate of 0.6 per­
cent. (See table 1.)
From 1958 to 1965, average growth in productivity
was 1.9 percent; output rose 4.6 percent, and hours ad­
vanced 2.7 percent annually. During this period, the in­
dustry experienced a general expansion. The number of
establishments manufacturing soaps and detergents in­
creased from 608 in 1958 to 704 in 1963.
Patricia S. Wilder is an economist in the Division of Industry Produc­
tivity Studies, Bureau of Labor Statistics.

Reprinted from the
M onthly L abor Review, February 1980.




From 1965 to 1974, productivity grew much faster,
averaging 4.3 percent each year. The acceleration was in
sharp contrast to the productivity movements of other
industries in the economy. More than two-thirds of the
industries for which productivity measures are available
showed slackening productivity growth since 1966. Pro­
ductivity growth in the soaps and detergents industry
during 1965-74 reflected average annual increases of
4.9 percent in output and 0.6 percent in employeehours. The slower growth in employee-hours was asso­
ciated with an overall decline in the number of estab­
lishments—from 704 in 1963 to 642 by 1972.
In 1975, a recession year, productivity fell 7.1
percent. Output recorded its largest decline of 9.4 per­
cent, and employee-hours declined 2.4 percent. In 1976,
productivity growth resumed with a 3.0 percent gain
with both output (5.8 percent) and hours (2.8 percent)
increasing over the depressed levels of the preceding
year. In 1977, however, output growth slowed to 2.2
percent, while employee hours increased 2.8 percent.
This resulted in a 0.6-percent decline in productivity.

Output doubles
Productivity gains in the soaps a '^detergents indus­
try have been closely linked to output expansion, which
doubled between 1958 and 1977. Some significant fac­
tors affecting this growth are expanded use of home
laundry equipment and dishwashing appliances, popula­
tion growth, and successful advertising and sales pro­
motions.2

Table 1. Productivity and related indexes for the soaps
and detergents industry, 1958-77

temperatures. The household laundry equipment indus­
try followed the development of permanent press gar­
ments within a few months by the introduction of
properly matched cycles in washers and dryers to han­
dle this new concept in clothing.5 At present, many au­
tomatic washers include permanent press cycles and
various combinations of wash and rinse temperatures.
The increase in the sales of home laundering equip­
ment, as well as the increase in wash and wear fabrics,
favorably affected the demand for soap and detergent
products. The output of the household laundry equip­
ment industry is estimated to have increased nearly 70
percent between 1958 and 1976. In 1975, more than 4
million home washing machines were sold, increasing
market penetration to 70 percent, from 53 percent in
I960.6

i

[1967 = 100]
Employee-hours

O
utput per employee-hour
Y
ear

1958 ...
1959 . . .
1960 . . .

A
ll
Produc­ Nonpro­
employ­
duction
tion
ees
workers workers

O
utput

77.7
84.4
81.7

78.3
85.4
81.5

76.3
82.2
82.0

64.7
71.8
71.9

83.3
85.1
88.0

82.6
84.1
88.2

84.8
87.3
87.7

A
ll
Produc­ Nonpro­
employ­
tion
duction
workers workers
ees

1961
1962
1963
1964
1965
1966
1967
1968
1969
1970

...
...
...
...
...
...
...
...
...
...

82.6
83.9
90.7
90.7
88.1
94.2
100.0
101.1
101.1
105.7

81.7
81.7
87.5
88.3
87.0
94.0
100.0
102.4
104.1
110.4

84.2
89.7
98.8
96.5
90.7
94.6
100.0
98.2
95.0
96.3

75.7
78.8
85.2
88.7
88.5
93.8
100.0
106.0
109.9
115.3

91.7
93.9
93.9
97.8
100.4
99.6
100.0
104.8
108.7
109.1

92.6
96.5
97.4
100.4
101.7
99.8
100.0
103.5
105.6
104.4

89.9
87.8
86.2
91.9
97.6
99.2
100.0
107.9
115.7
119.7

1971
1972
1973
1974
1975
1976
1977

...
...
...
...
...
...
...

108.6
120.0
127.5
132.7
123.3
127.0
126.2

114.8
125.1
134.4
139.6
129.0
135.0
135.6

96.5
110.1
114.2
119.3
112.1
112.0
109.1

111.7
125.9
135.1
137.9
125.0
132.3
135.2

102.9
104.9
106.0
103.9
101.4
104.2
107.1

97.3
100.6
100.5
98.8
96.9
98.0
99.7

115.7
114.3
118.3
115.6
111.5
118.1
123.9

;

Employment shows moderate rise
Employment in the soaps and detergents industry,
currently at 40,000, has increased moderately since
1958, when employment was at 32,000. This change is
equivalent to an average increase of 1.1 percent each
year. The growth in employee hours—an average annu­
al rate of 1.2 percent— reflected a very small increase in
average hours per employee.
Labor turnover in the industry has been comparative­
ly low, providing a stable and experienced work force.
Since 1958, accessions have averaged 2.5 per 100 em­
ployees annually, compared with 3.6 for all manufactur­
ing. Separation rates have been 2.4 per 100 employees,
compared with 4.1 for all manufacturing. Lower layoff
and quit rates occurred in the industry than for all
manufacturing almost every year. Average hourly earn­
ings for production workers in the soaps and detergents
industry have risen steadily. Hourly earnings averaged
$7.81 in 1977, compared with the manufacturing aver­
age of $5.68.
The proportion of nonproduction workers in the in­
dustry is somewhat higher than is the case in other
manufacturing industries— 37 percent of total employ­
ment in 1977, compared with 28 percent for all manu­
facturing. The higher proportion reflects the larger
number of professional and technical, clerical, and sales
personnel employed.
Although data on the occupational composition of
employees in the industry are not available, some in­
sights can be obtained from the broader aggregation,
soaps and cosmetics.7 In 1976, an estimated 6 percent of
all workers employed in soaps and cosmetics were chem­
ical and industrial engineers, chemists, and chemical
technicians. Sales and clerical personnel accounted for
26 percent of total employment. The industry also em­
ploys a large number of semi-skilled workers, such as
packers, wrappers, examiners, assemblers, and mixers
who accounted for 32 percent of the work force in 1976.

Average annual rates of change (in percent)
1958-77
1973-77

2.9
-0.6

3.4
-0.2

1.9
-1.5

4.1
-0.4

1.2
0.2

0.7
-0.2

2.1
1.1

!

The growth in output has also been influenced by the
availability of a wide variety of soap and detergent
products which can handle different types of cleaning
problems. Among synthetic detergent products are
light-duty, mild, sudsing detergents mainly used for
dishwashing by hand; all-purpose and heavy-duty laun­
dry detergents, which' can be used for a number of
tasks; presoak products; and automatic dishwashing de­
tergents. Laundry soaps are also available as flakes and
blown granules. The predominantly used soap product
is the refined bar of toilet soap. As shoppers are aware,
these bars are available in a variety of sizes, colors, and
scents, some containing additives such as cold creams
and deodorants. The industry is very competitive and
has been able to gain public acceptance of new products
through advertising, and by dispensing free samples in
large numbers when new products are introduced.3
The growth in output has also been influenced by the
interactions among the household laundry equipment,
textile, and detergent industries. The development of
permanent press garments in the mid-1960’s by the
textile industry was followed by reformulations in deter­
gent products. Because oily soils are more difficult to
remove from synthetic fibers, their increased use in
clothing required improved detergent products. Also,
because higher wash temperatures may cause oily soils
in some synthetics to become “set,” lower wash temper­
atures are often recommended for wash and wear gar­
ments.4 The detergent industry developed improved
products that would perform adequately at lower wash



119

Larger plants dominate output

German scientists are credited with developing the
first synthetic detergents during World War I.1 Syn­
0
thetic detergents were introduced into the United States
during the early 1930’s. The first synthetic detergents
performed well in hard water; however, their cleaning
ability was limited in laundry usage. In the 1940’s, the
discovery and development of phosphates, primarily so­
dium tripolyphosphate, led to the first “built” synthetic
detergents which not only performed well in hard wa­
ter, but provided the cleaning power necessary for laun­
dry use.1 By 1958, soap for many centuries the chief
1
cleansing agent for household laundry and dishwashing
use, had been largely replaced by synthetic detergents.

Most of the soaps and detergents industry’s output is
produced by large establishments. By 1972, more than
80 percent of the value of shipments was accounted for
by units having 100 employees or more. These units
represented only 9 percent of the industry’s establish­
ments because most of the industry’s establishments are
small. In 1972, 69 percent of the 642 manufacturing es­
tablishments had fewer than 20 employees.
Prior to the introduction of synthetic detergents, the
soaps and detergents industry tended to concentrate
near the sources of its principal raw materials. In more
recent years, with increased detergent usage, more em­
phasis is given to locations near distribution centers
when new sites are considered. Although production es­
tablishments are located throughout the Nation, about
half of the industry’s production originates in the North
Central region of the United States.
Increases in labor productivity are frequently related
to increases in the stock of capital. Over the period of
this study, new capital expenditures per employee in the
soaps and detergents industry increased at an average
annual rate of 10 percent, compared with 7.8 percent
for all manufacturing. Moreover, the levels were sub­
stantially above the average for all manufacturing in al­
most every year. By 1976, capital expenditures per
employee were 82 percent higher than the manufactur­
ing average ($4,191, compared with $2,300). About
three-fourths of the expenditures have been for new ma­
chinery and equipment, the same as for all manufactur­
ing.

Detergents reformulated. Developments over the past 15
to 20 years have resulted in many changes in product
composition. Because of environmental concerns, deter­
gent products have been and are still being reformu­
lated. One of the first changes in detergent composition
occurred in 1964-65 involving the replacement of the
organic surfactant with a type which degrades rapidly
in the environment.1 Specifically, “hard” branched2
chain alkylbenzenesulfonate (ABS) was replaced by
“soft” biodegradable linear alkylbenzenesulfonate
(LAS). LAS is still a major detergent ingredient.
By the late 1960’s, the focus of environmental con­
cern shifted to phosphate levels in detergent products
because of the controversy over the effect of phosphates
upon rivers, streams, fish, and other wildlife. Legislation
restricting phosphate levels in detergents was intro­
duced, including a total ban on phosphate in detergents
in several States. To maintain detergent performance
with reduced phosphate levels, surfactant levels are gen­
erally increased. Also, the use of surfactants, which are
even less sensitive to water hardness than LAS, helps to
maintain cleaning performance. For this reason, surfac­
tants based on long-chain alcohols have become more
popular.
These detergent formulation changes which occurred
in the mid- and late 1960’s coincided with years in
which productivity grew substantially less than the in­
dustry long-term average. Also, exceptionally large an­
nual increases in nonproduction workers occurred
which suggests that the industry, in response to the en­
vironmental concerns, expanded its research efforts into
the development of environmentally more acceptable
products.
Although sodium tripolyphosphate is still the leading
detergent builder, new builders are beginning to appear
and are currently used as phosphate substitutes. These
include sodium carbonate, sodium silicate, and various
surfactant blends. Other possible phosphate replace­
ments are being developed and tested, but none of these
materials has proved to be a total replacement on a
one-to-one basis.

Technology changes
Soap has always been made by combining the basic
ingredients, fat and alkali. The early American commer­
cial soapmakers made soap outdoors in large iron ket­
tles over an open fire, according to a uniform formula.
The kettle method of soapmaking was used until 1940
when a major improvement was achieved’in soap pro­
duction technology. A continuous process was perfected
which reduced soapmaking time from about a week to
less than a day.8 Today, the continuous process is dom­
inant, although the “kettle” process is still used in some
establishments.
Soap reacts with the minerals in hard water to form
lime soap, which sometimes appears as a white scum in
the wash water. Synthetic detergents, however, do not
react this way. The term detergent usually refers to a
product, which for heavy-duty laundry use, generally
contains an organic surface active agent (surfactant), an
inorganic builder, and various other ingredients. Also,
“detergent” is sometimes used to denote the organic
surfactant.9



120

Production processes improved. By 1958, virtually all of
the basic equipment currently used in soap and deter­
gent making had been developed. Most of the improve­
ments which became available later were technological
refinements of the basic equipment and production pro­
cesses. However, some notable improvements have been
introduced.
One of the major processes in the manufacture of
synthetic surfactants is sulfonation. In this process, a
nonsurface-active hydrocarbon (alkylbenzene, for in­
stance) is converted into surface-active alkylbenzenesulfonic acid, and subsequently neutralized to a salt.
Oleum is the sulfonating agent.1 In the mid-1950’s, an
4
innovation was developed which permitted the industry
to convert batch sulfonation into a continuous process.
With the continuous oleum process, a high-quality, uni­
form product could be obtained which met the impor­
tant production criteria, principally light color and low
free oil (unconverted hydrocarbon) in the final sulfona­
tion product. Time saving is another advantage; the
continuous sulfonation process is completed in a matter
of minutes, whereas the batch process requires 6 to 10
hours.1
5
In the mid-1960’s, a further improvement was intro­
duced in the continuous sulfonation process involving a
change from oleum to sulfur trioxide (S03) gas, mixed
with air, as the sulfonating agent. The industry-wide
trend towards the use of continuous S 03 has occurred
mainly because former sulfonating agents, such as ole­
um, have higher chemical costs, and present disposal
problems of spent sulfuric acid. Also, some of the newer
types of raw materials mentioned earlier cannot be pro­
cessed efficiently except with S 0 3. This process provides
a high-quality product by minimizing product degrada­
tion due to the short reaction time, reducing costs, and
realizing labor savings in the handling of the acid dis­
posal product.1
6
Continuous sulfonation processes with automatic con­
trols minimize labor requirements. An entire continuous
sulfonation plant can be operated with one operator,
rather than the two or three operators needed in the
batch and semi-continuous plants.1
7
Packaging operations in the industry have long used
automatic equipment. However, some technological
modifications have been introduced. For example, ma­
chines have been developed to handle larger powder
packs and at the same time are capable of erecting car­
tons, and filling and closing them at higher speeds. For
liquids, machines have also been introduced that can
achieve higher filling speeds.
High-speed soap bar production. Changes have also been
made in soap bar finishing operations.1 Although con­
8
tinuous soap production lines have been in operation
for many years, the need for faster production rates and



121

the development of more complex shapes of bar soaps
spurred improvements over the past 15 years. Extensive
changes have occurred in the design of the equipment
and the line configurations.
New high-speed lines for production of simple or uni­
form type bar soap formulas have broken the tradition­
al line speed barrier of 150-200 bars per minute. With
the high-speed lines, 200-300 soap bars can be pro­
duced each minute. Modern specialty lines are available
which provide flexible processing capability. A variety
of toilet bar formulations such as synthetic detergent
bars, soap-synthetic bars, and translucent soaps can
now be produced at reasonable speeds. New high-speed
stamping machines have also been developed which can
produce up to 400 bars per minute either banded or
bandless. In addition, refrigerated stamping dies have
become standard in the industry. They serve to improve
product appearance, to lessen die-fouling, and to im­
prove production rates.
New developments have also been made in the ma­
chinery that is widely used to package bar soaps. These
developments complement the development of the high­
speed finishing lines and have been directed primarily
toward wrappers, cartoners, and bar soap transfer units.
This new equipment has the capability of attaining
higher speeds and has the flexibility of handling various
shapes of bar soap.1 In the past, bar soap transfer units
9
were limited to maximum speeds of 200 bars a minute.
In the last 4 to 5 years, the speed has been increased.
Wrappers, cartoners, and bar soap transfer units are be­
ing introduced that are capable of average production
speeds of up to 300 bars a minute for the mass-pro­
duced soaps.
Computer technology has made possible the central­
ized instrumentation of the production processes,
although the industry has always been highly mecha­
nized. Computers are increasingly being used for jobs
such as inventory control, flow and measurement of raw
materials, formula calculations, and in mixing opera­
tions to assure uniformity of soap and detergent mixes.
Marketing analysis can more easily be accomplished
with computer-based information systems. The use of
computer processing provides information that can be
used to better allocate the time required for many activ­
ities, resulting in improved utilization of labor.
Shortrun changes in productivity in the soaps and de­
tergents industry will continue to be affected by changes
in demand. Over the longrun, the high levels of capital
investment per employee should help to keep industry
productivity gains in line with the average for all manu­
facturing.
Substantial demand for virtually all of the products
produced by the industry should continue into the im­
mediate future. The output of dishwasher detergents
should especially show growth as the utilization of

existing dishwashing machines is increased, and the
ownership of home dishwashers is expanded. The num­
ber of washing machines in U.S. households is also
expected to increase, thus generating additional growth
for the soaps and detergents industry.

8“About Soap,” Procter and Gamble Service Bulletin, Procter and
Gamble, Cincinnati, Ohio.
9 Based on information provided by Dr. Arno Cahn, Development
Director, Household Products, Lever Brothers Co.
1 “Some Facts About Procter and Gamble Detergents,” Procter and
0
Gamble Information Bulletin.
1 Anne L. Lyng, “Detergents in Review,” Detergents— in Depth, a
1
symposium sponsored by The Soap and Detergent Association, Wash­
ington, D.C., Mar. 28-29, 1974, pp. 2 -7 .
1 T. E. Brenner, “Soaps and Detergents: North American Trends,”
2
The Soap and Detergent Association, in Proceedings— World Confer­
ence on Soaps and Detergents, Oct. 9 -1 5 , 1977, Montreux, Switzer­
land, pp. 5 -8 . Reprinted in Journal o f the American Oil Chemists'
Society, January 1978. Also see, O. Carl Kerfoot and H. R. Flammer,
“Synthetic Detergents: Basics,” Hydrocarbon Processing, March 1975,
pp. 74-78.
1 Brenner, “Soaps and Detergents.”
3
1 Based on information provided by Dr. Arno Cahn, Development
4
Director, Household Products, Lever Brothers Co.
1 Oleum Sulfonation Process Equipment, The Chemithon Corp., Se­
5
attle, Washington, 1968. Also conversation with respresentative of
The Chemithon Corporation.
1 Sulphur Trioxide Detergent Process Equipment, The Chemithon
6
Corp. Ibid.
1 Oleum Sulfonation Process Equipment and Sulphiir Trioxide Deter­
7
gent Process Equipment. Ibid.
1 A. B. Herrick, “Bar Soap Finishing— New Trends in Soap Pro­
8
cessing Line Design and Layouts,” Armour-Dial Company, in
Proceedings— World Conference on Soaps and Detergents, Oct. 9 15, 1977, Montreux, Switzerland. Reprinted in Journal o f the Ameri­
can Oil Chemists' Society, January 1978, pp. 147-50.
1 L. Spitz, “Bar Soap Packaging,” ACMA S.p.A., in Proceedings —
9
World Conference on Soaps and Detergents, Oct. 9 -1 5 , 1977,
Montreux, Switzerland. Reprinted in Journal o f the American Oil
Chemists’ Society, January 1978, pp. 151-55.

1The soap and other detergents industry comprises establishments
primarily engaged in manufacturing soap, synthetic organic deter­
gents, inorganic alkaline detergents, or any combination thereof, and
refined glycerine from vegetable and animal fats and oils. The indus­
try is designated as number 2841 in the Office of Management and
Budget’s Standard Industrial Classification M anual (SIC), 1972 edition.
Data prior to 1958 are not comparable. All average annual rates of
change are based on the linear least squares trends of the logarithms
of the index numbers. Extensions of the indexes will appear in the an­
nual BLS Bulletin, Productivity Indexes fo r Selected Industries. A tech­
nical note describing the m ethods used to develop the indexes is
available from the Division of Industry Productivity Studies.
2 U.S. Industrial Outlook, various issues.
3 Industrial Outlook, 1970, p. 181.
4 Dieter H. Von Hennig, “The Role of Detergent Alcohols in the
Soap and Detergents Industry, A Bicentennial Update,” Shell Chemi­
cal Company, at Chemical Industry Association, Inc. Workshop
Meeting, Absecon, New Jersey, June 14, 1976.
5 Richard C. Davis, “Washer-detergent-textile Interactions,” H ydro­
carbon Processing, March 1975, pp. 9 0 -9 2 .
6 Richard B. Carnes, “Laundry and cleaning services pressed to
post productivity gains,” Monthly Labor Review, February 1978; and
“ 1976 Statistical and Marketing Report,” Merchandising, March 1976,
pp. 3 8 -4 2 .
7 Bureau of Labor Statistics, unpublished data for 1 9 7 0 -8 5 , N ation­
al Industry Occupational Matrix.




122

Measuring P’ oiictivity in
r
Government
Federal, State, and Local
JEROME A. MARK

S a result o f the growth o f government, which now em ­
ploys one out o f every six members o f the workforce, and
public concern over the rising costs o f government, the
need to develop measures o f productivity for public agencies has
becom e increasingly important. However, the concepts underlying
productivity measurement are so com plex, that experts do not
share a com m on perception o f the field; rather, they define the
nature and purpose o f productivity measures in a number o f ways,
particularly with regard to the public sector. The public sector
literature variously defines productivity measurement as measure­
ment o f efficiency, effectiveness, cost reduction, input-output,
management im provem ent, performance, m ethods improvement,
work standards, and program evaluation.
Efforts to evaluate productivity can be grouped into three broad
areas; programs em ploy either efficiency type measures, opera­
tional type measures, or effectiveness type measures. Efficiency
type measures compare the inputs or resources an organization
uses with the final goods or services it produces. This measurement
approach does not, however, determ ine whether these products
should be produced or relate them to some desired goal.
Where efficiency type measures assess the productivity o f a given
activity by relating that activity to its result or som e kind o f fin­
ished product, work m easurem ent or operational type measures
are mainly concerned with the activity itself. As Joh n P. Ross and
Jesse Burkhead point out, “ work measures exam ine the work ac­
tivity itself rather than its results and are usually measured in terms
o f activity per unit o f tim e.” 1 Work measures evaluate interm ediate
activity by assessing resource requirements under a given tech no­
logy or set o f conditions; in contrast, productivity measures relate
inputs to final products.

A

Unlike efficiency or operational type measures, effectiveness
type measures quantify a program’s im pact on society and deter­
mine whether that program makes optim um use o f inputs or re­
sources to achieve its goals. The output indicator for this approach
to m easurement is the effect that an activity has on a portion o f
the public. The thrust o f measurem ent changes from a comparison
betw een the goods and services that a program or activity produces
and the inputs used, to a study o f the public that consum es those
goods and services.
The principal differences betw een the three productivity evalu­
ation system s lie in the definition and m easurem ent o f outputs.
Bradford, Malt, and Oates have classified efficiency type measures
as measures o f “ direct ou tp u ts” and effectiveness type measures
as measures o f “ consequ en ces.” 2 In m any cases, however, this dis­
tinction is more apparent than real.
Although there is a substantial am ount o f confusion about the
specific meaning o f “ produ ctivity,” it seems clear that all o f these
performance measures help public sector managers make decisions.
The degree o f difference between efficien cy and effectiveness
measures is greater in the public sector than in the private sector.
In the private sector, com petition and market forces encourage
businesses to operate efficien tly and effectively so that they can
provide the goods the public demands at prices consumers are
willing to pay. Because market forces and com petition do not af­
fect the public sector, the relationship between efficiency and
effectiveness becom es more tenuous. For this reason, it is far more
im portant to separate these tw o measures in analyses o f the public
sector than in analyses o f the private sector.
Each o f the three kinds o f m easurement system s—efficiency,
work, and effectiveness—is im portant in its own right, and each
plays a role in evaluating productivity. However, much o f the re-

* J e r o m e A . M a r k is A s s i s t a n t C o m m i s s i o n e r f o r P r o d u c t i v i t y a n d T e c h n o l o g y a t t h e
B u r e a u o f L a b o r S t a t i s t i c s , U .S . G o v e r n m e n t D e p a r t m e n t o f L a b o r . T h e a u t h o r g r a t e ­
1 John

f u l l y a c k n o w l e d g e s t h e a s s i s t a n c e o f C h a r l e s A r d o l i n i , D o n F i s k , a n d J a m e s U r i s k o in

Dimensions
o f Productivity Research: Proceedings o f the Conference on Productivity Research,
April 21-24, 1980, E d i t e d b y J o h n D . H o g a n , ( H o u s t o n , T X : A m e r i c a n P r o d u c t i v i t y

v ic e s:

C e n t e r , 1 9 8 0 ) , V o l u m e II.

1 8 5 -2 0 2 .

th e

p re p a ra tio n

o f th is

a r tic le . A s lig h tly

Productivity in the Local G overnm ent Sector

2 D .F . B r a d f o r d , R . A . M a l t , a n d W .E . O a t e s , “ T h e R is i n g C o s t o f L o c a l P u b l i c S e r ­

Reprinted from the P u b l i c P r o d u c t i v i t y R e v ie w , March 1981.




P. R oss a n d J e s s e B u rk h e a d ,

( L e x in g to n , M A : L e x in g to n B o o k s , 1 9 7 4 ) , 1 4 .

d i f f e r e n t v e r s io n a p p e a r e d in

123

Som e

E v id e n c e

and

R e fle c tio n s ,"

National Tax Journal,

X X II

(Ju n e

1969)

search on productivity focuses on efficiency type measures and
these are the measures which this paper principally addresses.

F is c a l Y e a r

1971

Productivity M easurement at the Federal Level
For several years, the principal effort to measure the produc­
tivity o f agencies in the federal government has been the Bureau
o f Labor Statistics’ program to develop labor productivity indexes
for all governm ent agencies with 200 or more em ployees. This
program was undertaken first in conjunction with the General
A ccounting O ffice, the O ffice o f Management and Budget, and the
Civil Service Com m ission, later with the N ational Center for Pro­
ductivity and Quality of Working Life, and is currently administered
with the O ffice o f Personnel M anagem ent.3 The program currently
obtains data anddevelops measures for approxim ately 350 agencies
representing about sixty-five percent o f federal em ploym ent.
Table 1 shows the growth in the coverage o f the project over the
years.
The measures developed are indexes o f output per unit o f labor
input (generally, em ployee-year) and the agency measures are
grouped into functional categories that have com m on characteris­
tics. Table 2 shows the functional categories for which measures
are derived and the associated average annual rates o f change for
productivity and related series.
The productivity indexes developed compare the current ou t­
put-input relationship with that o f a previous reference period, in
this case fiscal year 1967, and the measures reflect the changes
which have taken place in labor input per unit o f output regardless
of the m ission o f the organization.
Where it is possible, the relevant concept o f outp ut for a govern­
m ent agency is its final product—that is, what the given organiza­
tion produces for use by the public or other governm ent agencies.
The output data included in the overall sample used for this study
is final from the perspective o f both the agency itself and the func­
tional groupings in which these agencies are classified.
However, since one federal agency may consum e all or some
o f the outputs o f another federal agency to produce its own final
outputs, all output indicators are not final from the perspective o f
the entire federal government. Therefore, the overall statistics
presented in the study do not represent “federal governm ent pro­
du ctivity,” but rather the average o f the productivity changes o f
the measured federal organizations included in the sam ple.4
It is particularly im portant to understand the meaning o f these
statistics in order to use this data for improving the measurement
o f general government output in the National Incom e and Product
accounts. Real gross product originating in general governm ent at
present is measured by moving the base year com pensation o f
government em ployees by changes in em ploym ent adjusted for
shifts in grade structure. As a result, the implied productivity
change for this measure is biased toward zero reflecting only the
impact o f the change in grade structure o f federal em ployees. Since
real gross product is a net measure, it is necessary to net out inter­
mediate output in order to use it in the developm ent o f improved
GNP m easures.5

3 F o r d e s c r ip tio n s o f th e w o r k o n th is p r o j e c t a n d th e p r e s e n ta tio n o f re s u lts , se e
U .S . C o n g r e s s , J o i n t E c o n o m i c C o m m i t t e e ,

the Federal Sector
A.

M ark

and

Measuring and Enhancing Productivity in

( W a s h i n g t o n , D .C .: U .S . G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 7 2 ) ; J e r o m e

C h a rle s W . A r d o lin i, “ D e v e lo p m e n ts in M e a s u rin g P r o d u c tiv ity

in th e

F e d e r a l S e c t o r , ” Proceedings o f the Business and Economics Section o f the American
Statistical Association ( W a s h i n g t o n , D . G , 1 9 7 4 ) 2 3 6 - 4 5 ; J o i n t F i n a n c i a l M a n a g e m e n t
I m p r o v e m e n t P r o g r a m , Productivity Programs in the Federal G overnm ent ( W a s h i n g t o n ,
D . C . : U .S . G o v e r n m e n t P r i n t i n g O f f i c e , 1 9 7 5 ) .
4 M a rk a n d A r d o lin i, “ D e v e lo p m e n ts in M e a s u rin g P ro d u c tiv ity .”
5 F o r a d is c u s s io n o f t h e p r o b le m s o f u s in g th is d a ta in th e n a tio n a l in c o m e a n d
p r o d u c t a c c o u n t s , s e e C h a r l e s A . W a ite a n d A l l a n D . S e a r l e , “ C u r r e n t E f f o r t s t o M e a s u r e
P r o d u c tiv ity

in

th e

P u b lic S e c to r :

H ow

A d e q u a te

fo r th e

N a tio n a l A c c o u n ts ,” a n d

New D evelopments in Productiv­
ity Measurement and Analysis; Studies in Incom e and Wealth, E d i t e d b y J o h n W . K e n ­

J e r o m e A . M a r k , “ C o m m e n t s o n S e a r le a n d W a i t e , ” i n

d r i c k a n d B e a t r i c e N . V a c c a r a ( C h i c a g o : U n i v e r s i t y o f C h ic a g o P r e s s , 1 9 8 0 ) .




124

1972

1973

1974

1975

1976

1977

197 8

17

45

46

49

51

51

50

50

O r g a n iz a tio n a l U n its

114

187

200

245

279

307

319

347

O u t p u t I n d ic a t o r s

948

1120

1201

1581

1747

2035

2334

2660

C iv ilia n E m p lo y e e -

1 .6

1 .7

1.7

1.8

1 .9

1 .9

1 .8

1 .8

2 .9

2 .8

2 .8

2 .8

2 .8

2 .8

2 .8

2 .8

A g e n c ie s

Y e a rs C o v e re d
( m illio n s )
T o t a l C iv ilia n E m ­
p lo y e e - Y e a r s
( m illio n s )
E m p lo y e e - Y e a r s

54%

60%

61%

65%

66%

66%

64%

65%

C o v e re d as P e rc e n t
o f T o t a l ( m illio n s )

T a b le 1: C o v e ra g e o f t h e F e d e r a l P r o d u c tiv ity M e a s u r e m e n t P ro g ra m

Although it would be desirable to have a net measure o f federal
government output and productivity, this is not possible at present.
While it is possible to identify some o f the measured outputs that
are consum ed inside the federal governm ent, sufficient data is not
available to determine the degree o f internal consum ption.
To determine final output indicators, the agencies and the
Bureau o f Labor Statistics (BLS) must identify specific units o f
services which are countable, fairly hom ogeneous over time,
flexible enough to be adjusted for quality changes, and represen­
tative o f the agencies’ workload. In addition, since historical
trends are o f interest, it is im portant that the measures are easy to
contruct from readily available records.
The indicators vary substantially. They include such diverse
items as trademarks disposed, tanks repaired, weather observations
made, square feet o f buildings cleaned, electrical pow er generated,
and deportable aliens located. The output volum es range from sev­
eral hundred units com pleted per year (e.g. river basin studies) to
billions (e.g. mail delivery items).
Em ployee-indexes are developed from the data each agen­
cy submits. As in all labor input measures used to develop produc­
tivity indexes, em ployee-years are treated as hom ogeneous and
additive and no distinction is made betw een different groups of
em ployees.
Since the productivity measure is a labor productivity series,
the line item output indicators arc com bined with the correspon­
ding labor input weights. These weights arc constructed from the
detailed output and input data provided by each participating
organization.
BLS productivity indexes measure both the overall sample and
also tw enty-eight major functions representing relatively h om ogen­
eous groups o f activities such as library services, procurem ent,
finance and accounting, electric pow er production, and postal ser­
vice. Although indexes for each o f the 350 organizational units
have also been constructed, they arc not published. Rather, they
arc returned to each organization for its own use (e.g.to stim ulate
further exam ination o f productivity changes).
M E A S U R E M E N T P R O B L E M S . There are several im portant prob­

lems in measuring productivity in the federal sector. Because
most federal activities are service oriented, it is often difficult to
define and quantify outputs; governm ent agencies usually do not
produce clearly specified physical products comparable to the
goods produced by the private sector. Moreover, after the output
indicators are specified, they m ust be in su fficien t detail to repre­
sent a hom ogeneous group o f services. If the outp ut units repre­
sented by the output indicator are not hom ogeneous, and if over
time the proportion changes betw een units that are more or less

F u n c tio n a l G roupings
T o ta l
A u d it o f O perations
B uild in g s & G rounds M aintenance
C o m m u n ica tio n s'*
E d u ca tio n & T ra in in g !
E le c tric Power P ro d u c tio n & D is trib u tio n
E q u ip m e n t M a in te n a n ce !
Finance & A c c o u n tin g
Genera! S u p p o rt Services
In fo rm a tio n Services
Legal & Ju d icia l A c tiv itie s
L ib ra ry Services
Loans & Grants
M edical Services
M ilita ry Base Services
N atura l Resources <k E n v iro n m e n t Managem ent
Personnel Investigations
Personnel Managem ent
Postal Service
P rin tin g & D u p lic a tio n
P rocurem ent
Records M anagem ent
R eg u latio n -C o m p lia n ce & E n fo rce m e n t
R eg u latio n -R u le m a kin g & Licensing
Social Services & B enefits
Specialized M a n u fa ctu rin g
S u p p ly & In v e n to ry C o n tro l
T r a ffic M a n a g e m e n t**
T ra n s p o rta tio n

O u tp u t p e r
E m p lo y e e -Y e a r

O u tp u t

E m p lo y e e
Years

C om pensation
Per E m p lo y e e
Year

U n it L a b o r
C ost

1.4

1.3

-0.1

8.7

7.2

2.5
2.5
10.1
0.9
0.3
0.4
2.4
3.9
0.3
-0.3
5.3
4.2
-0.1
-0.4
1.3
3.6
1.8
1.3
-1.5
2.0
3.1
2.0
2.6
2.8
1.9
2.0
1.3
2.6

-1.2
0.6
9.8
-1.2
7.9
-3.2
0.2
6.0
1.7
3.9
8 .6
5.1
1.9
-4.6
1.0
12.6
6.6
0.8
-3.5
-0.9
-1.4
5.0
4.6
7.5
-0.7
-4.8
-3.4
3.4

-3 .6
-1.9
-0.2
-2.0
7 .6
-3.6
-2.1
2.1
1.5
4.2
3.2
0.9
2 .0
-4.2
-0.3
8.7
4.7
-0.4
-2 .0
-2.9
-4.4
3.0
2 .0
4.6
-2.6
-6.7
-4.6
0.8

7.2
9.5
6.8
8.1
8.1
7.8
7.7
7 .3
5.7
5 .6
8 .4
7.9
7.9
7 .4
7.0
6 .6
5 .4
10.2
9 .4
5 .4
8 .3
6.9
6.7
7 .0
8.5
7.4
5.7
8.5

4.6
6.8
-3.0
7.1
7.8
7.4
5.2
3.3
5.5
5.9
3.0
3.6
8 .0
7.9
5.6
2.9
3.5
8.8
11.0
3.3
5.0
4.9
4 .0
4.1
6.5
5.3
4 .4
5.7

Source: Bureau o f Lab o r S tatistics, U.S. D e p a rtm e n t o f Labor. Average annual p ercent change based on lin e a r least squares tre n d o f the lo g a rith m s o f the
index num bers. * is fiscal year 1973 -1 97 8 . t is fiscal year 1968 -1 97 8 . * * is fiscal year 1972-1 978.

Table 2: Functional and summary average annual rates o f change in output per em ployee-year and related data for the measured
portion o f the federal civilian government, fiscal years 196 7-1978

labor intensive, the resultant output per em ployee-year measure
will reflect not only the change in productivity—the change in the
am ount o f labor required to produce the base year com posite—
but also shifts in the types o f output.
A nother related problem is that some reported outputs do not
reflect changes in output quality. Adjustm ents for changes in ou t­
put quality are necessary in order to appropriately measure the
changes in resources used per unit o f similar type good or service.
If a reduction in labor requirements per unit o f outp ut results from
a change in the dim ensions o f the outp ut or in the quality o f the
service, then the resultant measure does not reflect productivity
improvement.
For purposes o f labor productivity m easurement, changes in
outp ut quality, which reflect an altered production process with
different base period labor requirements, can be considered as ba­
sic changes in the output. Similarly, changes in outp ut character­
istics, which affect the value o f the output to the user but not an
altered production process or different base period labor require­
ments, do not require special treatment.
The main remedy for these difficulties is to collect more de­
tailed data. For exam ple, initially the only output indicator used
to assess the productivity o f the Postal Service was the number o f
pieces o f mail handled. By exam ining other Postal Service data,
BLS learned that output indicators covering various types o f mail
(e.g. registered and first class) and other services (e.g. m oney orders)
were available. The measure for the Postal Service is now based on
this detailed data and takes into account shifts in the importance
o f the different types o f mail services, all o f which require differ­
ent levels o f unit labor.
A lthough the use o f this additional product detail has improved
the accuracy o f the measure for this function, opportunities for
further im provem ent remain. For exam ple, at the present time t h e




125

measure reflects shifts betw een the less labor intensive deliveries
to large office buildings which have centralized mail distribution
facilities, and the more labor intensive deliveries to scattered pri­
vate, suburban residences, as well as productivity changes in these
com ponent activities. It would be desirable to adjust the measures
for these shifts.
BLS has sought, and in some cases obtained, detailed output
indicators, but in many cases the reported measures only summar­
ize an agency’s activities. For exam ple, the indicators record the
total number o f weapons overhauled rather than breaking this
figure down by type o f w eapon and specifying type o f repair. In
order to ensure that units o f reported outp ut were hom ogeneous,
BLS asked each participating agency to specify typical maximum
and minimum am ounts o f time required to produce one output
unit. A narrow range indicated that the labor requirements for each
unit o f output were relatively hom ogeneous within a given year
and that the effects o f a shift in output mix would be minimal.
Conversely, a wide range indicated that the units measuring ou t­
put were n ot standard and thus could not reliably reflect changes
in a shift in output mix. M ost outputs had a relatively narrow range;
however, a lim ited number o f output indicators showing wide
ranges had to be dropped from the study.
O utputs having long production cycles (i.e. requiring many
m onths or in som e cases, years to com plete) also present difficult
measurement problems. Q uantifying such outputs only in the year
they are com pleted produces output measures that are inconsistent
with the associated factor input. When production time extended
beyond one year, estim ates o f the proportion o f long term outputs
produced in each year were made.
Closely related to the problem o f quantifying output are the
difficulties that emerge when the federal governm ent contracts
out work. In these cases, the final output measure for a govern­

m ent agency may reflect the activities o f not only the agency’s
em ployees, but also its private contractors. It is important to de­
termine which output is exclusively associated with government
em ployees, because the input measure should be lim ited to them.
For exam ple, the Em ploym ent and Training Adm inistration o f the
Departm ent o f Labor administers contractor-operated job training
programs. Analyses o f the productivity o f these programs did not
directly use the outputs o f the contractors since it could not be
established that the outp ut o f the federal em ployees was propor­
tional to the results or the activities o f the contractors. The project
attem pted to com pensate for this problem.
Departm ental reorganizations also present difficulties in ob­
taining consistent output and em ployee-year data. Whenever a
reorganization occurred, respondents were asked to subm it consis­
tent data on the basis o f either the new or the old organizational
structure. Within the Departm ent o f Labor’s Em ploym ent and
Training Adm inistration, for exam ple, the offices overseeing train­
ing, em ploym ent, and un em p loym ent com pensation services were
reorganized several tim es during the fiscal years 1967-1978. BLS
has developed a separate measure for the un em p loym ent insurance
activities, but since the training and em ploym ent counseling ser­
vices are so closely related and their records are so intertwined,
it was considered more appropriate to construct a com bined m ea­
sure for these activities. Using budget and other docum ents, BLS
developed indexes based on the current organizational structure
for the fiscal years 1967-1978.
C U R R E N T E F F O R T S . BLS is attem pting to expand the program’s
coverage o f the federal sector. It is believed that the sam ple’s cov­
erage can be expanded from the sixty-five percent o f federal em ­
ployees it n o w includes to a m axim um o f perhaps eighty-five per­
cen t o f federal em ployees. Much work needs to be done before
measures can ever be developed for areas such as the research activ­
ities in the N ational Aeronautical and Space A dm inistration and
the National Bureau o f Standards.
Therefore, BLS concentrates on improving the outp ut indica­
tors currently reported. This process involves developing indicators
that more accurately represent the results of an agency’s activities;
these improved measures will use more clearly defined and de­
tailed product inform ation than some o f the indicators presently
in use.
On the input side, BLS devotes some attention to exploring
ways o f accounting for changes in the characteristics o f the feder­
al work force. In the present set o f measures, all em ployee-years
are treated as hom ogeneous and additive. It w ould be desirable
to learn how shifts in the com position o f the work force effect
changes in output per unit o f labor unit input.
In addition to the work o f BLS and the O ffice o f Personnel
M anagement (OPM), various federal agencies have undertaken other
activities generally reflecting efficiency-type measures.
O T H E R E F F O R T S A T F E D E R A L L E V E L . While BLS conducts
the annual productivity data call to construct measures for the
overall federal governm ent and the functional categories, partici­
pating federal agencies conduct their ow n research efforts to develop
m easurem ent system s. These measurem ent system s are developed
internally but often are technically supported by BLS or aided by
outside contractors. As the system s becom e operational, BLS will
review the m ethods they use and may incorporate their results in
the overall federal statistics. The exam ples that follow dem onstrate
the approaches that are being developed to overcom e conceptual
problems in productivity measurement.
The Naval System s Command has enlisted BLS support to de­
velop meaningful measures o f productivity for its ship overhaul
activities. Ship overhaul presents tw o significant m easurem ent pro­
blems. This activity involves outputs with a long production cycle
time; it som etim es takes more than eighteen m onths to retrofit a




ship. Recording these outputs only in the year they are com pleted
produces an output measure that is inconsistent with the associated
input measure. In addition, the custom ized nature o f the outp ut
precludes the developm ent o f a standard unit o f output that re­
mains hom ogeneous over time.
These problems can be addressed by estim ating the proportion
o f long-term output produced in each year or by breaking the final
outputs down into com ponent parts, each o f which is com pleted
within a m easurement period. The thrust o f this productivity meas­
urement program is to analyze ship overhaul in terms o f repairs to
the ship’s major system s such as its hull, armament, and propulsion
and electrical system s. Within each o f the system s, the navy has
selected sample indicators that reflect the output o f the system
and are hom ogeneous over time. Data is now being collected and
the final results should yield an accurate and useful measure o f pro­
ductivity in ship overhauling.
A research project in the area o f hospital care is underway at
the Veteran’s Adm inistration. Traditionally, productivity meas­
urement has been weak in this area because the output indicators
generally used in these activities (beds occupied, patient-days, or
patients treated) have severe lim itations. The measures which result
from the use o f these indicators do not reflect the varying resource
requirements needed by type o f illness (e.g. fractures versus cardio­
vascular diseases). The measures also contain a bias o f unknown
direction and magnitude because the length o f stay per type o f ill­
ness varies over time. A more accurate measure o f productivity
would distinguish betw een type o f illness, diagnostic category, etc.,
and would make allowance for the fact that varying units o f labor
are required to treat each type o f illness.
BLS staff have provided technical assistance to the Veterans
Adm inistration in its effort to portray trends in productivity more
accurately. This project is part o f a new integrated m anagement
inform ation system. The purpose o f the system is to provide more
tim ely data on costs and to develop a firmer basis for estim ating
em ploym ent needs.6
Som e agencies are conducting measurem ent activities that are
designed along the lines o f the M undel hierarchy o f work un its.7
This approach classifies service outputs into well-defined quantifi­
able units o f work. The various classifications serve as a guide for
tracing the production cycle from resource input to the organiza­
tion’s outputs or achievements. The m eth odology begins with time
and m otion studies, moves to the measurement o f interm ediate and
final output, and ends with an attem pt to measure effectiveness.
While the basic model postulates eight levels o f work units, it can
collapse or expand to m eet the needs o f the organization. The V et­
erinary Service o f the Departm ent o f Agriculture, which has the
mission o f ensuring that livestock herds are free o f infectious di­
seases, uses this system to evaluate animal inspection activities.
The Department o f the Interior, HEW, and the Secret Service have
also em ployed this measurement technique.
At the Department of Housing and Urban D evelopm ent (H UD),
a new management control system allocates regional staff. The
system classifies major programs into categories depending on the
activities required to com plete defined outputs. K nown as the
operating plan system, this program planning tool permits HUD to
allocate staff on the basis o f workload projections. For instance,
the system develops standard times required to process a housing
application. Given a reasonably accurate estim ate o f the number
o f applications it will have to process in the upcom ing year, HUD

S26

6 A r e c e n t d is c u s s io n o f th e v a r io u s a p p r o a c h e s to

p r o d u c tiv ity

m e a s u r e m e n t in

h o s p i t a l s is p r o v i d e d b y W . R i c h a r d S c o t t , “ M e a s u r i n g O u t p u t s i n H o s p i t a l s , ” i n

m ent and Interpretation o f Productivity

Measure­

(W a s h in g to n , D .C .: N a ti o n a l R e s e a r c h C o u n c il,

N a tio n a l A c a d e m y o f S c ie n c e s , 1 9 7 9 ) , 2 5 5 - 7 5 .
7

M a r v in

E.

M u n d e l,

G overnm ent Organizations

Measuring and Enhancing the Productivity o f Service and
(H o n g K o n g : A s ia n P r o d u c tiv ity O r g a n iz a tio n , 1 9 7 5 ).

can determine labor requirements for that output. The operating
plan is set at the beginning o f the year and each m onth accom plish­
ments and staff-years utilized are compared to targeted forecasts.
This permits HUD to exam ine staff utilization rates and re-allocate
resources when unplanned disruptions occur. While the HUD pro­
gram springs from detailed work measurement concepts, the system
is flexible to permit data aggregation at any level o f analytical
need. Thus, when the system becom es fully operational, BUS will
be able to derive productivity data from an autom ated system.
Lastly, mention must be made o f OPM’s experim ental project
in the area o f administrative services. The project is designed to
measure productivity o f a com m on support service and has stan­
dardized definitions to facilitate inter and intra-agency compari­
sons so that managers have an adequate am ount o f inform ation
on which to base decisions. The project began with the function
o f personnel services'and is now in the second year o f data collec­
tion .8 Data is being collected on outputs and inputs within the
categories o f staffing actions, position classification, labor rela­
tions, em ployee developm ent, and general administration. In con ­
trast to the existing Federal Productivity Measurement System ,
which directly collects outputs and associated inputs, OPM’s meas­
urement system em ploys random sampling techniques to generate
weighting factors. As the pilot project proves feasible, OPM o f­
ficials hope to extend it to other administrative services such as
finance, procurement, and data processing.
Som e federal agencies have aggressively pursued efficiency re­
search activities and are exam ining indicators that may dem on­
strate how effectively they are m eeting their mission statem ents.
These measures go beyond the efficiency measures contained in
the federal productivity data base. For the purposes o f efficiency,
the Drug Enforcem ent Agency is measuring the number o f nar­
cotics arrests forwarded for prosecution. For the purposes o f effec­
tiveness, that is, how well the agency is discharging its mission to
protect society from illegal drugs, the measure may be an index o f
narcotics prices (i.e. rising prices indicate a declining supply) or a
measure o f drug-related deaths and hospitalizations.
Som e agencies involved in activities such as awarding grants
have been attem pting to develop effectiveness measures. Where
the grant is for research, as in the physical or behavioral sciences,
one approach has been to have a peer group evaluate the quality
o f the research or to record the number o f times the findings o f
the research are cited in professional journals. The National Science
Foundation uses this approach to measure the quality o f the re­
search activities it funds.
In agencies that disburse m onies, com m on measures o f e ffe c ­
tiveness are the percent o f paym ents made on tim e, the timeliness
with which lost checks are reissued, and the percent o f financial
reports for public use com pleted on time.
Within the E m ploym ent and Training Adm inistration (ETA)
o f the Department o f Labor, a primary function is to issue con ­
tracts for the operation o f J ob Corps centers. ETA measures this
effort by expressing the number o f trainees placed in J o b Corps
centers as a percent o f expected slots to be filled. ETA believes
this approach permits it to analyze its success in m eeting expansion
targets o f the Job Corps program.
Som e organizations attem pt to com bine efficiency and effec­
tiveness into a single overall performance indicator. The Army
Material Command, for exam ple, has adopted a procedure which
develops one measure for effectiveness and a separate measure for
efficiency for each o f the com m and’s depots. The results o f all the
depots are then numerically ranked on a scale from one to fifteen.
The procedure assigns one score for effectiveness and one for e f­

8

A l l a n S . U d l c r , “ A n E x p e r i m e n t in t h e P e r s o n n e l O f f i c e , ”

Civil Service Journal

X IX (J a n u a ry /M a rc h 1 9 7 9 ), 3 2 -3 5 .




127

ficien cy; the com bined score represents each d ep ot’s performance.
Thus, with one indicator, the ArmyM aterial Com m and shows both
how well a mission is accom plished (effectiveness) and how com ­
pletely resources are utilized (efficiency).
It seems better to keep the tw o types o f productivity measure­
ment systems separate. D ifferent types o f productivity measures
serve different purposes. For exam ple, assessing the am ount o f
resources used per unit o f output in order to improve resource
allocation may call for an efficiency type measure. Evaluating the
impact o f resource use on program goals may call for an effective­
ness type measure. It is doubtful that com bining different types
o f measures into a single measure would produce satisfactory
measures in m ost cases.
F U T U R E D I R E C T I O N S . While BLS has extended productivity
measurement to cover agencies representing two-thirds o f the fed­
eral governm ent’s civilian work force, further expansion is required.
In the future, however, the measurement program’s rate o f growth
will slow because it is difficult to define and quantify the output
o f many o f the government agencies not currently covered by the
program. Research and developm ent and policy-oriented activities,
such as those in the National Institutes o f Health, the State Depart­
ment, and defense activities fall into this category. Innovative ap­
proaches to measurement must be developed for these activities.
BLS must continue to refine and improve existing output
measures in the present m easurement system . It will be necessary
to m odify reporting system s to provide more suitable indicators.
Moreover, as the federal establishm ent places increased em pha­
sis on gauging the performance o f organizations, federal agencies
will have to expand and develop existing measures so that they
become far more detailed. As a result, the governm ent will begin
measuring outputs that arc not currently measured because, under
the present reporting system , they are considered as intermediate
rather than final outputs. For exam ple, an overall productivity

As the federal establishment places increased
emphasis on gauging the performance of
organizations, federal agencies will have to
expand and develop existing measures so that
they become far more detailed.
measure o f an agency may be useful when the entire agency is re­
viewed. However, such a measure, which masks the performance
o f the agency’s many bureaus, administrations, or programs fails
to supply enough inform ation to the executives directing the sub­
activities o f the agency. They need more detailed data.
This need is amply demonstrated by a request to BLS from the
Department o f Agriculture. In the department, one o f the many
organizations furnishing productivity inform ation is the Soil Con­
servation Service. The outputs and associated inputs for the entire
Department o f Agriculture are collected, processed, and returned
to the service for its analysis. This measure is useful for analyzing
the Department o f Agriculture, but the executives o f the Soil Con­
servation Service felt additional inform ation would be more use­
ful. They wished to have productivity measures for the service’s
many programs. BLS m et this need by developing productivity
measures for individual programs such as flood prevention services,
snow forcasts, and soil mapping. Because the Soil Conservation
Service now has its own detailed set o f productivity measures, it
is possible to analyze the specific programs o f the service as well
as the overall performance o f the Departm ent o f Agriculture.
It is clear that we must adopt this type o f approach in order
to provide managers with m eaningful data. BLS has always offered
to provide technical assistance in developing measures for detailed

agency use, but unfortunately, the response from the agencies has
been poor. Finer levels o f measurem ent must be stressed and en ­
couraged if the federal agencies are to have reasonably accurate
productivity measures for use in resource allocation and budgeting.
Currently, the m easurement system relates output to a single
factor, labor input. Since there are as many productivity measures
as there are factors o f production, Som e people are interested in
professional em ployees or shifts in levels o f educational attain­
ment, the results are reflected as productivity m ovem ents and not
as factor input changes. Productivity measures could be refined so
that they are able to adjust for such things as the age, sex, and e x ­
perience o f the work force. This area o f research should be co n ­
sidered one o f the more feasible expansions o f the federal produc­
tivity measurement system .
It is generally agreed that the m ost suitable unit o f labor input
measure is the number o f hours worked. The current input meas­
ure, em ployee-years, is based on the number o f hours paid. Thus,
the denom inator o f the productivity expression includes actual
hours expended to produce the output plus hours paid for sick
and annual leave, overtim e, and accrued leave paid at separation.
This probably does not present a measurement bias if the ratio
o f hours worked to total hours paid is stable over time. However,
in the public sector, this issue has never been analyzed. While the
exten t o f any potential bias is thought to be minimal, this subject
should be researched.
The O ffice o f Personnel M anagement, which is responsible for
improving the productivity o f the federal government, would like
in see effectiveness measures supplem ent tire existing data base.9
As m entioned earlier, this is a most difficult area o f productivity
measurement. The com plexities o f defining and quantifying ou t­
put measures involve normative judgem ents and a variety o f as­
sum ptions. The research literature has extensively questioned the
value and availability o f standards, norms, or targets.10 Even if it
is possible to develop multiple measures of effectiveness, there is
no clear conceptual basis for aggregating them with an efficiency
measure into an overall performance m easure.I
11
An exam ple will illustrate these problems. Suppose the focus
o f the productivity analysis is the Social Security Adm inistration.
For an efficiency type measure, the final output indicator is the
exploring alternative measures o f productivity for the federal gov­
ernment. The m ost com m on alternative to single factor measures
is output per com bined unit o f labor and capital. There has been
a great deal o f research on m ultifactor productivity estim ates, but
these efforts have been lim ited to analyses o f the private sector.
W ithout listing every problem inherent in the developm ent o f cap­
ital estim ates, it is necessary to discuss a few difficulties that make
the federal government unique when considered from the perspec­
tive o f m ultifactor analysis.
The com m on form of the capital input estimate in m ultifactor
productivity measurement is a capital stock measure. However,
the governm ent’s current accounting system precludes the devel­
opm ent o f such a measure because the system charges purchases
o f equipm ent and structures in the year o f acquisition. Thus, there
is no convenient m ethod for estim ating the value o f the capital
stock available for the production o f goods and services^ For e x ­
ample, a vehicle acquired for a m otor pool fleet is expensed in the
year it is purchased. That vehicle has a useful life o f several years

I S ec

O P M m e m o ra n d u m fo r h e a d s o f e x e c u tiv e d e p a r tm e n ts a n d a g e n c ie s , “ F e d e ra l

P r o d u c t i v i t y M e a s u r e m e n t P r o j e c t : F is c a l Y e a r 1 9 7 9 I n s t r u c t i o n P a c k a g e , ” d a t e d J a n u ­

and will be contributing a flow o f capital services over the period
o f time it is in use. However, there is now no m ethod o f measuring
the depreciated value o f the vehicle in any year after the year o f
acquisition. Thus, it is necessary to introduce a system for fixing
the value o f capital equipm ent and structures before beginning to
research m ultifactor productivity in the federal governm ent.
An official o f the Departm ent o f the Treasury has suggested
the G overnment Financial Operations unit within the Treasury
Department, a heavily com puterized organization, as a possible
candidate for a m ultifactor study. Perhaps som e initial pilot stud­
ies will provide insights on developing a governm ent-wide frame­
work for extending productivity measures beyond com parisons
o f output per unit o f labor input.
New ways o f measuring labor input in the present system can
also be explored. Current m ethodology considers labor input as
hom ogeneous and additive; that is, an hour’s worth o f work per­
formed by any general schedule em ployee regardless o f his position,
expertise, or experience, is weighted equally in the measure. This
procedure, therefore, does not take into account the com position
o f the work force. Furthermore, as changes occur in the co m p o ­
sition o f labor, such as changes in the ratio o f professional to n on ­
number o f checks disbursed. For purposes o f effectiveness type
measures, however, the choice o f the proper indicator is not as ob­
vious. The manager o f the program might evaluate effectiveness
by measuring the percent o f the number o f recipients who receive
their checks on the first o f the m onth. The political scientist or
welfare econom ist may define effectiveness as the degree to which
the incom e o f the population served is above or below the poverty
line. Thus, various disciplines would select different measures o f
effectiveness. In addition, the feasibility o f com bining the measures
o f effectiveness and efficiency into a meaningful overall measure
o f performance remains questionable.
The work which some federal agencies have carried out on e f­
fectiveness measurement is still in the experim entation stage. Much
research remains to be done on the standards against which e ffe c ­
tiveness can be measured and on possible techniques for data col­
lection. The disciplines concerned with public sector productivity
face the challenge o f developing m odels for measuring effective­
ness which accurately convey to both the public and policy planners
the degree to which a public sector organization is accom plishing
its goals.
Productivity Measurement at the State and Local Level
Compared with our knowledge o f measurement in the federal
government, we know very little about state and local government
productivity measurement. The diversity and m ultiplicity o f state
and local governments and the absence o f a continuing productivi­
ty information system makes it difficult to assess their productivity.
The knowledge we have com es from three sources: local govern­
ment surveys, academic studies, and work by the federal govern­
ment on state and local government enterprise funds.
Most surveys o f state and local government have concluded
that the majority o f governments collect and use efficiency-type
productivity measures, although in many cases the measures arc
limited to a small group o f activities. These measures are important
for budgeting, auditing, review, and daily operations. A 1976 sur­
vey o f local government found that about sixty-five percent o f the
cities and fifty percent o f the counties regularly used efficiency
measures at some stage o f their operations.12 More recent surveys
in North Carolina and the Denver M etropolitan area alsouncovered

ary 2 1 , 1 9 8 0 .
10

H a r r y P. H a tr y , “ T h e S ta tu s o f P r o d u c tiv ity in t h e P u b lic S e c t o r ,"

istration Review,
II J e s s e
D e fin itio n

X X X V III (J a n u a ry /F e b ru a ry

B u rk h e a d
and

and

O r d e r ,"

P a tric k J .

Public A d m in ­

1 9 7 8 ), 2 8 .

H e n n ig a n , “ P r o d u c tiv ity A n a ly s is : A S e a r c h f o r

Public Adm inistration Review,

12

X X X V III (J a n u a ry /F e b ru a ry




R ackham

F u k h u h a ra ,

The Status o f Local G overnm ent Productivity

D .C : T h e I n t e r n a tio n a l C ity M a n a g e m e n t A s s o c ia tio n , 1 9 7 7 ) .

1 9 7 8 ), 3 4 .

128

(W a s h in g to n ,

frequent but uneven use o f efficiency m easures.13 On the other
hand, a 1976 exam ination o f budget docum ents o f 247 cities and
counties found that only ten percent contained any efficiency
m easures.14 Many local governm ents use efficiency measures but
do n ot include them in their budget docum ents.
Inform ation on state use o f efficiency measures is even more
lim ited than that on local governm ent. A survey o f state budget
docum ents and statistical reports in 1975 found m ixed use o f ef­
ficiency-type m easures.15 In fields such as corrections and edu­
cation, about one-third o f the states included atleast one efficiencytype measure in their budgets. These included such measures as
“ dollar cost o f caring for inmate per inmate day” and “number o f
offenders per parole officer.” For functions such as tourism and
licensing,less than ten percent o f the states included any efficiencytype measures in their budgets.
Several conclusions emerge from these surveys o f state and local
governm ent use o f efficiency-type measures. First o f all, larger jur­
isdictions are more likely to develop and use efficiency-type meas­
ures than smaller jurisdictions. Secondly, the use o f efficiency-type
measures is very uneven from service area to service area. Thirdly,
service areas subsidized or regulated by the federal governm ent are
more likely to have efficiency-type measures than other areas. In
addition, services with tangible outputs, such as solid waste collec­
tion, are more likely to be measured than those with intangible
outputs, such as general management. Tw o other conclusions emerge from these surveys which bear m entioning. State and local
governm ent docum ents include more effectiveness-type measures
tiian efficiency-type measures. A lso, state and local governments
are increasingly interested in the developm ent o f engineered work
standards. A lthough they can be a useful tool in increasing pro­
ductivity, engineered work standards do n o t measure productivity
themselves; rather, they measure normal, optim al, expected, or
other am ounts o f time required to perform individual tasks.
Cost measures calculated by state and local government do
not assess productivity either. Changes in unit costs, such as “cost
per stu dent” reflect factor price changes, as well as changes in
resources used.
Over the past decade, several members o f the academic com ­
m unity have prepared state and local government productivity in­
dexes. For the m ost part, these indexes have been parts o f larger
research issues, such as Robert D. Reischauer’s projections o f state
and local governm ent cost and revenue.16
These studies have taken one o f tw o approaches. One approach
has been to com pute a cost per unit o f service such as “ cost per
pupil day” in the case o f education and infer changes in productivi­
ty from changes in unit costs over tim e.17 The other approach has
been to specify unit cost as an independent variable and identify
changes in productivity (and quality) through the residual term .18
There are a number o f problems with the academic research.
F orm ost service areas, researchers have had to use “persons served”
or “ persons elegible for service” as the indicator o f output. For
exam ple, in public safety, the “population at risk” is often used
as the output measure. Such measures, unfortunately, do not
define or distinguish the number o f people served or the type o f

R e s e a r c h T r i a n g l e I n s t i t u t e , Comparative Performance Measures fo r Municipal
Services ( R a l e i g h , N . C : D e c e m b e r 1 9 7 8 ) ; D e n v e r R e g i o n a l C o u n c i l o f G o v e r n m e n t s ,
I D em onstration o f Comparative Productivity Measurement, ( D e n v e r : R e g i o n a l C o u n 1 :l o f G o v e r n m e n t s , 1 9 7 8 ) .
14

i l a t r y , “ T h e S ta t u s o f P r o d u c tiv ity in th e P u b lic S e c t o r ,” 2 9 .
U rb a n I n s titu te ,

In Initial Examination,
16 R o b e r t

1972 Budget

D.

The Status o f Productivity Measurement in State Government:
( W a s h in g to n , D .C .: T h e U r b a n I n s t i t u t e , 1 9 7 5 ) .

R e is h a u e r in

C h a rle s

I ,. S c h u l t z e ,

Setting National Priorities: The

( W a s h in g to n , D .C .: T h e B r o o k in g s I n s t i t u t i o n , 1 9 7 1 ) .

17 B r a d f o r d , M a l t , a n d O a t e s , “ T h e R is i n g C o s t o f L o c a l P u b l i c S e r v i c e s . ”
18 R o s s a n d B u r k h c a d ,

Productivity in the Local Government Sector.




129

service provided. In these measures, a shift in population could
appear incorrectly as a change in productivity.
The availability o f data has dictated the selection o f outputs,
and aggregate data on state and local governm ent operations is
not generally available. Data that is collected nationally, such as
crime rates, does not lend itself to productivity calculations, and
data which would be useful for output measurem ent, such as tons
o f solid waste collected, is not tabulated nationally.
O utput data is available for state and local government Enter­
prise Funds, which account for about seven percent o f all state and
local government expenditures. Governm ent services included in
Enterprise Funds are: toll highways, water transport, airports, li­
quor stores, parking, off-track betting, lotteries, transit and water
supply system s, gas and electical utilities, housing and urban re­
newal, and sewerage. Enterprise Funds form the only area where the
federal governm ent calculates state and local government produc­
tivity on a continuing basis. These services are managed like private
businesses; they sell goods and services, make profits or losses, and
pay taxes. Changes in Enterprise Fund activity are reflected in the
BLS quarterly index o f private sector productivity.
The process used in calculating Enterprise Fund outputs is to
take total sales, add any governm ent subsidies, and deflate to ob­
tain the outputs. These statistics are drawn from the annual census
o f governments and are supplem ented with trade association in­
formation and other data. The labor data is also taken from the
same sources. E m ploym ent data is taken from the current em ploy­
ment survey by BLS.
State and local governm ent Enterprise Funds are currently in­
cluded in private sector non-farm business productivity calcula­
tions. It would be possible to identify Enterprise Funds separately,
by service, and calculate a specific state and local government
Enterprise Fund productivity index if that were desired.
C U R R E N T E F F O R T S . Research on state and local government
productivity m easurement has decreased in recent years, partly due
to the term ination o f the N ational Center on Productivity and
Quality o f Working Life. Financial support from the N ational Sci­
ence Foundation and the D epartm ent o f H ousing and Urban
D evelopm ent has also decreased. However, BLS has launched a
new effort. Its investigation assesses the possibility o f calculating
state and local governm ent productivity indexes and gross output
measures. The effort is similar to those underway in individual
industries and the federal government. The services focused on
include state liquor stores, electric pow er distribution, mass
transit, elem entary and secondary education, and hospitals and
roads.
In each case, researchers exam ine prior studies, specify a series
o f outputs, identify quality or service factors that m ust be adjusted
for, survey the available data, id en tify labor input data, and docu­
m ent the results. If researchers can not calculate an index from
existing data, they suggest procedures to accum ulate data and esti­
mate resources.
The project has been underway for only a short time, and has
focused almost entirely on the Enterprise Funds, but it has yielded
several tentative findings. Som e service areas, such as mass transit
and public power have been extensively researched and there is
general agreement about m ethods o f m easurement for these areas.
In other service areas, such as liquor stores, there has been little
research on productivity techniques, but evidently measurement
o f these activities does not present serious problems. In other
areas, such as education, there is considerable research but little
agreement about productivity measurement. Also, some o f the data
necessary for com puting a state and local productivity index is
readily available from federal sources (e.g. public power statistics);
other data appears tob e available from the states (e.g.liquor stores);
and some data is not currently available but the Departm ent o f

Transportation is now devising a system that should eventually pro­
vide the needed inform ation. Finally, the study indicates that each
service area must be exam ined in depth. Very little knowledge or
data can be transferred from one service area to another.
As far as the federal government is concerned, the clim ate is
not bright for new initiatives in this area. The National Productivity
Council recently concluded that the federal government should not
make a major investm ent in the measurement o f state and local
governm ent productivity:
Th e costs and relative benefits o f a federal effort in this area com pa re d to
o th e r forms of federal s uppor t do n o t w arra nt a m ajor federal invest ment at
this time. More analysis o f the c once ptual proble ms and approaches should
be made. In the meantime , federal agencies should be encour aged to c ontinue
their individual efforts to s upp ort state and local gove rnment produc tiv ity
m e asu re ment and, where possible within their own budget priorities, ex pan d
such e ffo r ts .19
F U T U R E D I R E C T I O N S . One thrust o f the BLS study o f state and
local government productivity measurement is to identify those
areas needing additional research. The m ost difficult issue, o f
course, is output measurement, particularly in areas, such as edu­
cation, where there is a large body o f research hut little agreement
on measurement approaches, or parks and recreation, where there
is little research and no agreement on productivity approaches.
Most state and local govenm ent services fall into these two
categories.
Related to the quantity o f outputs is the quality and level o f
service. The need to account for quality differences when calcu­
lating state and local productivity has been discussed at length.
There has been little research on the impact quality variations have
on the production costs o f specific state and local government ser­
vices. Properly adjusting for changes requires a thorough under­
standing o f these relationships.
Data availability is an im portant research issue in its own right.
Given tod ay’s fiscal restraints, it is almost impossible to collect vast
am ounts o f data. In some cases, it should be possible to use existing
data to com pute state and local government productivity indexes.
In other instances, it may be necessary to establish new data col­
lection procedures. For each service area, it is necessary to deter­
mine the m ost cost-effective way to fill the gaps in existing data.

19

N a tio n a l P r o d u c tiv ity C o u r.c il,

m ent 1‘roductivity Im provem ent

Federal Actions to Support State and Local Govern­

(W a s h in g to n ,

D .C .:

N a tio n a l

P ro d u c tiv ity

C o u n c il,

1 9 7 9 ) , v ii.




130

Inform ation on the quantity o f labor hours is necessary for
calculations o f state and local government labor productivity.
Reports on the state and local governm ent work force in current
surveys, such as the Census o f G overnments, Current Population
Survey, and U nem ploym ent Insurance Report, have major lim ita­
tions. A system atic exam ination o f the current situation and the
opportunities for m odification is needed. This exam ination should
assess m ethods o f indicating shifts in the com position o f the work
force.
For m ost state and local governm ent operations, labor is the
m ost im portant resource, and in som e cases, it accounts for eightyfive to ninety percent o f operating costs. Thus, it is only appropri­
ate that state and local governm ent productivity programs start
by focusing on labor productivity. However, for som e services such
as streets, pow er utilities, and water and sewer system s, where
capital input is a substantial factor, an additional measure o f m ulti­
factor productivity including labor and capital is desirable. By
addressing a lim ited number o f state and local governm ent service
areas, it should be possible to develop a better appreciation for the
problems and opportunities in developing efficiency-based produc­
tivity indexes. BLS has used this approach for its industry work
and its state and local governm ent work.

Conclusion
A lthough public sector productivity measurem ent has been e x ­
tensively studied and a system of productivity reporting has been
established at the federal level, much research and additional
work remain. A m bitious output definitions and measures for e f­
ficiency-type productivity measurement must be elim inated or
refined, and innovative approaches to output measurement must
be developed so that the system o f productivity reporting can e x ­
tend to areas such as research and developm ent and defense acti­
vities. Future research efforts on input measures should be directed
to developing measures that take into account changes in com p o­
sition o f labor input and include other factor inputs such as capital.
These activities are applicable to productivity m easurement for
federal and state and local governm ent activities although the ap­
proaches will have to vary depending on the nature o f the activity.
At the same tim e, research directed toward the developm ent
o f effectiveness-type measures should be continued. However,
resolution o f the conceptual and measurement problems for these
types o f measures may be less promising.

Part IV„ International Comparisons

Comparable measures of productivity and labor costs
for industrial countries are described here. The first
report represents the latest in a regular, annual series,
and shows how the measures are applied and analyzed.
The broad international setting within which significant
differences in productivity trends occur is discuss­
ed, and some of the reasons for the differences are
suggested.
Background
From its inception, the Bureau has collected and
published statistical information on labor conditions
and developments abroad. Foreign labor research and
statistical analyses have been undertaken because (1) in­
formation on labor conditions published by a majority
of foreign countries is not readily available to U.S.
labor representatives, employers, Government officials,
and others, and is often not available in English; (2)
often, only an expert can judge the quality of foreign
statistical sources; (3) comparisons between U.S. and
foreign labor conditions shed light on U.S. economic
performance relative to other industrial nations; and (4)
comparisons provide information on the competitive
position of the United States in foreign trade, which has
an important influence on the U.S. economy and
employment.
Description of measures
The b l s foreign labor statistical reports cover a
variety of international comparative measures, mainly
for the Western industrial countries. The principal
measures cover the labor force, employment, and
unemployment; productivity and labor costs; hourly
compensation of manufacturing production workers;
and trends in consumer prices.

Hourly compensation. Measures of total compensation
per hour worked for production workers in all manufac­
turing and in over 30 selected manufacturing industries
are prepared annually for about 30 countries. The
measures are developed from data on average earnings,
as published by each country, plus information on other
direct payments to the workers and employer expen­
ditures for legally required insurance programs and con­
tractual and private benefit plans. They are expressed in
national currency and in U.S. dollars at prevailing com­
mercial exchange rates. Hourly compensation, when
converted to U.S. dollars at commercial exchange rates,
indicates comparative levels of employer labor costs. It
does not indicate relative living standards of workers or
the purchasing power of their income. Prices of goods
and services vary greatly among countries and commer­
cial exchange rates are not reliable indicators of relative
differences in prices.

Productivity and labor costs. Comparative trends in
manufacturing productivity (output per hour), hourly
compensation, unit labor costs (labor compensation per
unit of output), and related measures are compiled on
an annual-average basis for the United States, Canada,
Japan, Belgium, Denmark, France, Germany, Italy, the
Netherlands, Sweden, and the United Kingdom. Trends
are expressd in index form (1977= 100) and as percen­
tage changes at annual rates. For most countries, the
series begin with 1950. Indexes of unit labor costs for
foreign countries are calculated in national currency and



in U.S. dollars converted at prevailing commercial ex­
changed rates.
Comparative levels and trends in productivity and
labor costs in the iron and steel industry in the United
States, Japan, France, Germany, and the United
Kingdom have been compiled annually beginning with
1964. The measures express levels of foreign output per
hour, hourly compensation, and unit labor costs relative
to the U.S. level (United States = 100). They also show
trends in index form (1977 = 100) and at annual rates of
change.
Comparative levels (United States =100) and trends
(1977= 100) in gross domestic product (GDP), g d p per
capita, and g d p per employed person are calculated on
an average-annual basis for th United States, Canada,
Ja p an , Belgium, France, G erm any, Italy, the
Netherlands, and the United Kingdom beginning with
1950. The g d p level comparisons, which are based on
estimated purchasing-power-parity exchange rates, are
benchmarked to data from the United Nations Interna­
tional Comparison Project. Purchasing-power-parity
exchange rates represent the number of foreign currency
units required to buy goods and services equivalent to
what can be purchased with one unit of U.S. currency.
A common practice has been to base such comparisons
on official market exchange rates. However, market ex­
change rates seldom reflect the relative purchasing
power of different currencies.

131

Analysis and presentation

The presentation of foreign labor statistics varies with
the degree of analysis and major use of the data. Com­
prehensive bulletins have been published, covering
manufacturing productivity and labor cost trends, steel
productivity and costs, unemployment and labor force
comparisons, and youth unemployment comparisons.
For more current developments, articles are published
periodically in the Monthly Labor Review. Also, an an­
nual news release is issued on comparative trends in
manufacturing productivity and labor costs. The b l s
Handbook o f Labor Statistics and the Bureau of the
Census’ Statistical Abstract o f the United States publish
many of the principal foreign data series, and some
series are published in the annual Economic Report o f
the President. Many unpublished tabulations of current
comparative data are available on request.

Analyses of international labor statistics focus upon
comparisons with U.S. data. Wherever possible, foreign
data are adjusted to U.S. definitions and concepts to
facilitate comparisons; for example, the adjustment of
foreign unemployment rates to approximate U.S. con­
cepts and the adjustment of production worker earnings
to total hourly compensation.
Productivity and unit labor cost data are analyzed to
explain the relative contributions of changes in output,
employment, average hours, compensation, and ex­
change rates upon changes in the measures. Changes in
employee compensation are analyzed to determine the
relative contributions of direct pay and other elements
of compensation.




132

International trends in
productivity and labor costs
Output per employee hour in manufacturing
generally improved and unit labor cost trends
moderated in the US. and 10 other nations
in 1981; relative productivity and
labor cost indexes are introduced
P a t r ic ia C a p d e v ie l l e , D o n a t o A l v a r e z ,
a n d B r ia n C o o p e r

Manufacturing productivity increased in 1981 in the
United States, Japan, and most European countries
studied, with gains ranging from about 2 to 4 percent in
the United States, Japan, France, Germany,1 Italy, and
the Netherlands, to almost 6 percent in the United
Kingdom and Denmark, and more than 7 percent in
Belgium. In Canada and Sweden, productivity remained
essentially unchanged.
These productivity changes occurred in what was for
most countries the second year of recession. In most
European countries, productivity rose because employ­
ment and hours declined more than output. In the
United States, Canada, and Japan, productivity gains
were accompanied by modest output growth— tempo­
rary recoveries from 1980 declines in the United States
and Canada.
Unit labor cost increases, which reflect changes in
both productivity and hourly compensation costs,
ranged from 2 to 5 percent in Japan, Germany, Bel­
gium, Denmark, and the Netherlands, up to 15 percent
in France and 18 percent in Italy. When measured in
U.S. dollars, however, unit labor costs declined substan­
tially in all the European countries— 5 to 20 percent—
because of the sharp appreciation of the dollar, while
rising 7 to 8 percent in Canada and Japan as well as in
the United States.

Patricia Capdevielle, Donato Alvarez, and Brian Cooper are econo­
mists in the Division of Foreign Labor Statistics and Trade, Bureau of
Labor Statistics.

Reprinted from the
M o n th ly L a b o r R eview , December 1982.




133

While the 1981 appreciation of the dollar partially
offset the lower long-term U.S. cost trend, unit labor
costs in the United States nevertheless declined 29 per­
cent between 1970 and 1981, relative to the average
costs of our trade competitors. Unit labor costs in Can­
ada, Belgium, Denmark, the Netherlands, and Italy also
declined relative to those of their trade competitors
while those of Japan, France, Germany, the United
Kingdom, and Sweden increased.
This article describes developments in manufacturing
productivity (as measured by output per hour), hourly
compensation, and unit labor costs in 1981, and com­
pares the 1980-81 trends with those of the 1974— re­
75
cession, for the United States, Canada, Japan, France,
Germany, Italy, the United Kingdom, and four smaller
European countries— Belgium, Denmark, the Nether­
lands, and Sweden.2 Percent changes in productivity, la­
bor costs, and related measures for selected periods and
for each year from 1973 are shown in tables 1 through
3;3 percent changes are also presented for the eight Eu­
ropean countries and for the 10 foreign countries com­
bined.4 (Annual indexes for the years 1950 to 1981 are
available from the authors.) The data for 1981 are
based on preliminary underlying statistics, while those
for other recent years reflect revised underlying statistics
for several countries.
Although the productivity measure relates output to
the hours of persons employed in manufacturing, it
does not measure the specific contributions of labor as a
single factor of production. Rather, it reflects the joint
effects of many influences, including new technology,

Output. With the exception of a small gain in Denmark,
manufacturing output fell in each of the European
countries in 1981—by more than 6 percent in the Unit­
ed Kingdom and about 1 to 4 percent in the other
countries. In the non-Scandinavian countries, productiv­
ity increased because employment and hours declined
even more than output. Most of Denmark’s productivi­
ty gain also resulted from decreases in employment and
hours. In Sweden, hours and output fell equally.
The 1981 drop in British output followed an even
larger 1980 decline of 9 percent. For France and Bel­
gium, 1981 marked the second consecutive year of de­
clining output, but the 1980 declines were under 1
percent. Germany, Denmark, Sweden, and the Nether­
lands had zero or only slight 1980 output increases —
under 1 percent— while Italy had a more substantial
gain. In most countries, output turned down during the
first half of 1980, and showed little if any recovery by
late 1981 or early 1982. Only in Italy did output recov­
er in late 1980 and turn down again in 1981.
In the United States and Canada, 1980 manufactur­
ing output levels declined about 3 to 4 percent from
previous year levels, but 1981 annual output levels were
up 2 percent. In both countries, manufacturing produc­
tion dropped in the second quarter of 1980, recovered
in the fourth quarter, then turned down again during
the second half of 1981. In Japan, manufacturing out­
put increased more than 9 percent in 1980, and rose an­
other 3 percent in 1981, but then turned down during
the first half of 1982.

capital investment, the level of output, capacity utiliza­
tion, energy use, and managerial effectiveness, as well as
the skills and efforts of the work force.
This article also introduces new measures of relative
trends in productivity and labor costs. Table 5 presents
indexes of relative output per hour, hourly compensa­
tion, and unit labor costs in national currency and in
U.S. dollars for the 11 countries. Each relative index
represents the ratio of a country’s own index to a
weighted geometric average of the corresponding index­
es for the other 10 countries; the weights used to com­
bine the other country indexes reflect the relative
importance of each country as a manufacturing trade
competitor (table 4).

Productivity trends
In 1981, manufacturing productivity increased by
more than 7 percent in Belgium, almost 6 percent in the
United Kingdom and Denmark, and about 2 to 4 per­
cent in the United States, Japan, France, Germany, Ita­
ly, and the Netherlands. In Canada and Sweden, it rose
less than 0.5 percent. (See table 1.)
For the United States, the 1981 productivity gain was
the largest annual increase since 1976. And for Belgium
and the United Kingdom, the 1981 gains were the larg­
est in many years. For Japan and Italy, the 1981 in­
creases represent substantial slowdowns from large 1980
gains, but for most other countries, they were improve­
ments over small gains or productivity declines in the
previous year.

Table 1.

Annual percent changes in manufacturing productivity and output, 11 countries, 1960-81
Eight
European
countries

Ten
foreign
countries

United
States

Canada

Japan

France

Germany

Italy

United
Kingdom

Belgium

2.7
3.0
1.7

3.6
4.5
1.4

9.2
10.7
6.8

5.5
6.0
4.6

5.2
5.5
4.5

5.8
6.9
3.7

3.6
4.3
2.2

7.2
7.0
6.2

6.1
6.4
4.1

7.1
7.6
5.1

5.0
6.7
2.2

5.3
5.8
4.1

5.9
6.4
4.7

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

-2.4
2.9
4.4
2.5
.9
.7
.2
2.8

2.2
-2.6
5.3
4.0
1.6
1.7
-3.3
.3

2.4
3.9
9.4
7.2
7.9
8.9
6.8
3.2

3.5
3.1
8.2
5.1
5.7
4.9
1.6
1.6

5.4
5.3
7.1
4.9
3.3
4.9
1.4
2.7

4.9
-4.4
8.6
1.1
3.0
7.3
5.8
3.4

.8
-2.0
4.0
1.6
3.3
3.3
5.9

5.8
4.4
10.4
6.5
5.0
6.5
3.1
7.3

3.3
10.4
3.8
2.1
2.4
5.8
1.4
5.6

8.3
-1.8
12.8
4.1
6.6
4.9
1.3
3.1

3.6
-.4
1.0
-1.5
4.3
8.4
1.2
.1

4.1
1.6
7.2
3.3
4.0
5.3
2.8
3.8

3.8
2.0
7.5
4.3
4.9
6.1
3.6
3.3

O
utput:
1960-81 .....................
1960-73 .....................
1973-81 ......................

3.6
4.7
2.3

4.8
6.3
2.0

10.0
13.0
6.5

5.2
6.6
2.3

3.8
5.2
1.9

5.4
6.8
3.3

1.6
3.0
-1.7

5.0
6.5
1.1

4.0
5.2
1.8

4.7
6.4
1.7

3.2
5.1
-.3

4.0
5.4
1.5

5.3
6.8
2.9

-4.2
-7.1
9.6
6.9
5.3
2.7
-4.3
2.3

3.6
-5.9
5.9
2.0
5.0
4.7
-3.1
1.6

-2.0
-4.0
13.3
7.3
7.3
9.9
9.4
3.2

3.2
-2.1
7.0
3.7
3.2
2.6
-.1
-2.7

-.3
-4.8
8.0
2.4
1.3
4.8
.5
-1.4

6.4
-9.7
12.6
2.1
1.8
6.7
6.3
-1.0

-1.2
-7.0
2.0
1.9

4.6
-7.4
8.6
.7
.9
3.7
-1.4
-2.5

1.5
-2.1
4.8
.6
.7
6.5
.0
.5

4.4
-6.7
8.0
.9
2.8
2.7
.9
-.9

4.8
-1.5
-.4
-5.6
-1.3
6.9
.0
-3.6

1.8
-5.2
7.0
2.1
1.6
3.8
-.4
-2.4

-.9
-5.0
8.5
3.5
3.4
5.6
2.4
-.4

Year

O
utput per hour:
1960-81 .....................
1960-73 ......................
1973-81 ......................
1974
1975
1976
1977
1978
1979
1980
1981

1974
1975
1976
1977
1978
1979
1980
1981
N o te:

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

Rates of change com
puted fromthe least squares trend of the logarithm of the index num
s
bers.




134

.6

.6

.2
-9.1
-6.3

Denmark Netherlands

Sweden

Employment and hours. Manufacturing employment and
aggregate hours both increased only in .Canada in 1981;
in Japan, employment rose slightly but total hours were
essentially unchanged. In 1980, hours had increased
slightly in Canada and by more than 2 percent in Ja­
pan. In the United States, employment and hours de­
clined only slightly in 1981, after falling more than 3
percent in 1980. (See table 2.)
In Europe, employment declined 10 percent in the
United Kingdom and 2 to 6 percent in the other
countries in 1981. Those declines followed 1980 drops
of 6 percent in the United Kingdom and 1 to 2 percent
in most of the other countries. Employment had in­
creased slightly in Germany in 1980 and was essentially
unchanged in Italy and Sweden. Aggregate hours fell
even more than employment in 1981— except in Den­
mark—as average hours were also reduced.
Comparisons with 1974-75. Comparisons of develop­
ments during the years 1980 and 1981 with the
recession of 1974-75 cannot be precise, particularly
when dealing with annual average data, because of dif­
ferences among countries in the extent and timing of
the 1974-75 recession and the 1980-81 downturns.
Nevertheless, certain broad comparisons can be made.
Over the 1974— period, manufacturing output fell
75
in one or both years in all 11 countries studied. During
1980-81, neither Japan nor Denmark experienced annu­
al average declines in output, although Denmark had
virtually no output growth over the period and Japa­
nese output slowed sharply in 1981; most of the other
countries had smaller output declines than in 1974—
75.
However, there were exceptions. The recent output de­
clines in the United Kingdom were substantially greater
than during 1974-75, and those in France and Sweden
also appear to have been larger. Only in the United
States did output regain its pre-1974 average rate of
growth during the 1976-79 recovery period.
As in the case of output, manufacturing employment
and hours declined less sharply during 1980-81 than
during 1974-75 in most of the countries studied. For
example, German employment declined about 2 percent
in 1980-81, compared with 9 percent in 1974— and
75,
total hours declined 5 percent versus 15 percent. Again,
major exceptions were France, where employment and
hours declined somewhat more in 1980-81, and the
United Kingdom, where the recent declines— 16 per­
cent for employment and 21 percent for total hours—
were substantially greater than those in 1974—
75. In
Sweden, the employment effects of the 1974— reces­
75
sion were delayed; therefore, direct comparison between
the two periods is not appropriate.
Although employment losses over the 2-year period
of 1980-81 were less severe in most countries than in
1974— employment in most of Europe also declined
75,



during the intervening 1976-79 period. The rate of de­
cline ranged from about 1 percent per year in France
and Germany to almost 4 percent annually in Belgium.
Only in Denmark and Italy was employment essentially
stable during the recovery period. By 1981, employment
in manufacturing was down 6, 11, and 14 percent from
1973 levels in Sweden, France, and Germany; 17 per­
cent in Denmark and the Netherlands; and almost 25
percent in Belgium and the United Kingdom. In con­
trast, employment in the United States and Canada was
higher in 1981 than in 1973.
All European countries have taken actions, through
collective bargaining or government programs, to short­
en average hours worked to preserve manufacturing
jobs. Most countries have partial unemployment benefit
programs to provide wage replacement to employees on
short work schedules for economic reasons. In addition,
minimum annual holiday (vacation) entitlements have
been increased in Denmark, Germany, the Netherlands,
Sweden, and the United Kingdom (and are scheduled to
increase in France) as a job creation measure as well as
a fringe benefit improvement. (In Italy, on the other
hand, several national holidays were abolished in 1977
as a labor cost cutting measure, although many employ­
ees receive extra annual holidays in lieu of the national
holidays.) In Belgium, the standard workweek was
shortened through collective bargaining from 40 hours
in 1977 to 38 hours for most employees in 1981; the
shorter hours are provided as either a shorter workweek
or a longer annual holiday.
Given the relative output and employee-hours changes,
manufacturing productivity increased in most countries
during both the 1974— recession and in 1980-81. The
75
following tabulation shows average annual productivity
changes over the two periods:
1974-75
1980-81
C anada .................................
Japan ....................................
France .................................
G e rm a n y ...............................
I t a l y .......................................
U nited K ingdom ..............
Belgium ..............................
D e n m a rk ...............................
N e th e rla n d s .........................
S w e d e n .................................

0.2
-.2
3.2
3.3
5.4
.2
-.6
5.1
6.8
3.2
1.6

1.5
-1.5
5.0
1.6
2.1
4.6
3.2
5.2
3.5
2.2
.7

In the United States, Japan, Italy, and the United King­
dom, the productivity trend was higher during 1980-81,
while productivity gains were higher during 1974— in
75
France, Germany, Denmark, the Netherlands, and Swe­
den. In Belgium, productivity rose equally in both peri­
ods. In Canada, productivity declined in both periods.
H o u rly com pensation

Hourly compensation increases in 1981 varied
considerably among the 11 countries studied. The
135

Table 2.

Annua! percent changes in manufacturing employment and hours, 11 countries, 1960-81
Sweden

Eight
European
countries

-2.3
-1.1
-3.2

-1.7
-1.5
-2.5

-1.2
-.4
-2.4

-.6
.4
-1.7

-1.7
-11.3
' 1.0
-1.4
-1.7
.7
-1.4
-4.8

-3.6
-5.0
-4.3
-3.0
-3.6
-2.1
-.4
-3.8

1.2
-1.1
-1.5
-4.1
-5.4
-1.3
-1.2
-3.7

-2.2
-6.7
-.2
-1.2
-2.3
-1.5
-3.1
-6.0

-2.7
-6.8
1.0
-.8
-1.5
-.5
-1.1
-3.7

-.7
.6
-3.6

-.7
.2
-1.8

-1.0
.0
-2.4

-.3
-.2
-1.0

-.2
.5
-1.7

.3
1.1
-1.2

1.9
-3.8
-2.2
-.4
-2.4
-2.5
-6.0
-10.1

1.1
-6.1
-4.1
-3.9
-4.1
-2.7
-1.9
-5.5

-3.6
-8.4
.6
-.5
-.5
.8
-2.0
-4.8

-.4
-3.2
-4.0
-2.7
-2.5
-1.0
-1.2
-3.3

2.4
.9
-.2
-3.5
-2.8
.3
-.1
-3.2

.3
-3.9
-1.7
-.7
-1.5
-.9
-1.6
-4.5

.4
-4.2
-1.0
-.6
-1.2
-.5
-.3
-2.8

-1.4
-1.5
-.3

-.9
-.7
-1.0

-1.4
-1.0
-1.2

-1.3
-1.4
-.5

-1.2
-1.1
-.8

-1.4
-1.3
-1.5

-1.0
-.9
-.9

-.8
-.8
-.5

-1.1
-5.1
3.5
.9
-.2
-1.0
.3
-2.4

-3.8
-1.3
.3
.6
-.2
-.5
-3.9
-1.6

-2.2
-5.6
2.5
-1.6
.3
.1
-2.5
-3.8

2.0
-3.2
.4
-.9
-1.2
-.1
.7
.0

-3.2
-1.8
-.3
-.3
-1.1
-1.1
.8
-.6

-1.1
-2.0
-1.3
-.7
-2.6
-1.6
-1.1
-.6

-2.5
-2.9
1.5
-.4
-.8
-.6
-1.5
-1.6

-3.1
-2.8
2.0
-.2
-.3
.0
-.8
-.9

United
States

Canada

Japan

France

Germany

Italy

United
Kingdom

Aggregate hours:
1960-81 .....................
1960-73 .....................
1973-81 .. ..................

0.9
1.6
.6

1
.1
1.7
.5

0.7
2.1
-.3

-0.2
.6
-2.2

-1.3
-.2
-2.5

-0.4
-.1
-.4

-1.9
-1.2
-3.8

-2.0
-1.1
-2.2

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

-1.9
-9.7
4.9
4.2
4.4
2.0
-4.5
-.5

1.4
-3.4
.5
-2.0
3.4
2.9
.2
1.3

-4.3
-7.6
3.6
.1
-.5
1.0
2.5
-.1

-2.1
-.5
-4.8
-1.2

-.3
-5.0
-1.1
-1.3
-2.3
-2.2
-1.7
-4.3

-5.4
-9.6
.8
-2.4
-1.9
-.1
-.9
-4.0

1.4
-5.5
3.8
1.0
-1.2
-.6
.5
-4.3

-2.0
-5.1
-1.9
.2
-2.6
-3.0
-9.6
-11.5

-11.3
-1.7
-5.4
-3.9
-2.6
-4.3
-9.2

Em
ploym
ent:
1960-81 .....................
1960-73 .....................
1973-81 .....................

.9
1.5
.7

1.4
1.9
.8

1.5
3.0
-.4

.5
1.2
-1.4

-.4
.5
-1.6

1.1
1.4
.0

-1.1
-.5
-2.9

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

-.4
-8.6
3.7
3.6
4.2
2.6
-3.4
-.5

2.0
-2.2
.4
-2.0
3.2
3.7
.3
1.8

.2
-5.1
.4
-.2
-1.1
-.1
2.5
.5

1.3
-2.7
-1.0
-.5
-1.6
-1.8
-1.3
-3.6

-2.6
-6.7
-2.4
-.8
-.6
.3
.6
-2.4

2.5
-.4
.2
.1
-1.0
.5
.2
-1.9

Average hours:
1960-81 .....................
1960-73 .....................
1973-81 ......................

-.1
.1
-.2

-.3
-.2
-.3

-.8
-.9
.1

-.7
-.5
-.8

-.9
-.8
-.9

-1.5
-1.2
1.2
.6
.2
-.6
-1.0
.0

-.6
-1.1
.1
.0
.3
-.7
-.1
-.5

-4.5
-2.6
3.2
.3
.6
1
.1
-.1
-.6

-1.5
-2.3
-.1
-.9
-.7
-.4
-.3
-.7

-2.9
-3.1
3.2
-1.6
-1.4
-.4
-1.5
-1.6

Year

1974
1975
1976
1977
1978
1979
1980
1981

1974
1975
1976
1977
1978
1979
1980
1981

1974
1975
1976
1977
1978
1979
1980
1981
N o te:

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

Belgium

Denmark Netherlands

Ten
foreign
countries

Rates of change com
puted fromthe least squares trend of the logarithm of the index num
s
bers.

smallest gain was 5 percent in the Netherlands and the
largest, 22 percent in Italy. In the United Kingdom and
France, the increases were also large— over 16 percent.
In Japan and Germany, the gains were relatively small—
under 8 percent— while in the United States, Canada,
Belgium, Denmark, and Sweden, they were 9 to 12 per­
cent. (See table 3.)
Four countries—the United States, Germany, Den­
mark, and the United Kingdom— showed some degree
of moderation in hourly compensation gains for 1981.
In the United Kingdom, there was a substantial slow­
down from the 24 percent recorded in 1980. (In the
Netherlands, a significant slowdown occurred in 1980.)
In Canada, Japan, Italy, and Sweden, however, the
1981 increases were higher than those of the previous
year, and in France, Belgium, and the Netherlands, the
increases in both years were virtually the same.
Compared with the hourly compensation trend dur­
ing the 1974-75 recession, annual rates of increase dur­
ing the 1980-81 period were considerably lower in every
country except the United States and France. In the
United States, however, the 1974— increases were rel­
75
atively small. The moderation in wage gains and other



labor costs occurred even though consumer price trends
w ere g e n e ra lly a b o u t as h ig h in 1 9 8 0 -8 1 as in 1974—75

— with Japan and Belgium as principal exceptions.
However, growing concern with moderating labor costs
and containing inflation, as well as preserving manufac­
turing jobs, had a significant impact on recent compen­
sation trends.
Concerted action was taken in several countries to
moderate wage settlements during 1980-81. Temporary
pay freezes were imposed in Belgium and the Nether­
lands and a temporary price freeze was undertaken in
Sweden. The Dutch government subsequently imposed
statutory pay controls. In several countries with wage
indexation systems, the price indexes used were adjust­
ed to exclude fuel and energy prices, or the cost-of-liv­
ing allowances ( c o l a ’s) normally payable were reduced
or rescinded.
In Japan and Germany, annual wage agreements in
1980 and 1981 continued the moderate pattern of recent
years. In Japan, the average manufacturing settlement
was 6.7 percent in 1980 and 7.6 percent in 1981, and in
Germany, the average settlements were 6.7 percent in
1980 and 4.6 percent in 1981. In the United States and

136

additional reductions. Because wage rates are adjusted
to compensate for the shorter workweek, the hours re­
ductions are measured as hourly compensation gains.
Wage rates are also indexed for consumer price in­
creases in Italy, and cost-of-living allowances are paid
under collective agreements in Denmark, Sweden, and
the United Kingdom. In Italy as in Belgium, the
indexation system continued unchanged during 198081. In Denmark and Sweden, COLA payments were re­
stricted. In Denmark, the index used to compute the
c o l a ’ s was changed in December 1979 to exclude fuel
and energy prices, and was also rebased. As a result,
one of the c o l a ’ s was eliminated in 1980. In Sweden,
the 1981 pay agreements specified exclusion of energy
prices from the consumer price index used in COLA
computation. The government imposed a price freeze in
September 1981 and cut value-added taxes in Novem­
ber, and thereby kept the price rise below the COLA
threshold (trigger) specified in the pay agreement.
In Denmark, early 1981 wage settlements at the in­
dustry level provided moderate wage increases and re­
stricted additional company-level wage negotiations. In

the United Kingdom, wage-and-salary concessions were
made in some impacted companies or industries.
In the Netherlands, a pay freeze was imposed from
January through April 1980, followed by statutory con­
trols which were later extended through 1981. No basic
wage increases were allowed. Furthermore, the June
1980 cost-of-living adjustment was restricted to a flatrate amount, and the January 1981 adjustment was re­
duced by 2 percent. In 1981, holiday bonuses were low­
ered slightly, and extra annual holidays delayed.
The Belgian Government imposed a pay freeze in
January 1981. The national wage agreement signed in
February, under threat of statutory pay controls, pro­
vided either a 1-percent wage rate increase or an extra
hour off the standard workweek by 1983. Wages are
indexed for consumer price increases in Belgium, how­
ever, and the indexation system was not changed. The
emphases of recent wage settlements in Belgium have
not been basic wage increases but reductions of stan­
dard hours. Standard weekly hours were reduced from
40 hours per week in 1977 to 38 hours for most work­
ers by 1980, and the 1981 national agreement allowed

Table 3.

Annual percent changes in hourly compensation and unit labor costs in manufacturing, 11 countries, 1960-81
Sweden

Eight
European
countries

Ten
foreign
countries

12.9
12.8
9.7

12.0
10.4
13.0

12.0
9.8
13.7

11.9
10.1
12.4

21.0
19.3
11.7
10.6
10.2
11.8
10.9
9.3

19.2
14.3
12.5
8.6
8.7
7.8
5.0
5.3

17.6
21.2
18.5
9.2
11.3
7.8
10.9
12.4

18.3
18.4
13.0
12.0
11.6
12.4
14.9
13.8

21.4
18.0
11.2
11.3
9.8
10.7
12.0
11.5

5.1
3.5
5.6

6.8
5.1
8.0

5.5
4.8
4.4
10.0

6.7
3.5
10.6

6.3
3.8
9.2

5.8
3.5
7.4

24.1
32.5
12.7
10.8
12.8
15.0
22.9
9.7

15.7
16.3
2.5
5.2
2.9
1.1
6.4
2.1

17.1
8.0
7.6
8.4
7.6
5.7
9.4
3.5

16.4
-.3
4.3
1.9
2.8
3.7
2.1

13.5
21.7
17.3
11.0
6.7
-.5
9.6
12.3

13.7
16.6
5.5
8.4
7.2
6.7
11.8
9.7

17.0
15.6
3.5
6.7
4.7
4.3
8.1
7.9

7.1
2.6
15.0

7.8
4.6
8.6

7.9
5.0
7.7

8.7
6.1
8.0

7.7
4.2
9.6

7.6
4.2
9.9

7.2
3.9
8.8

18.5
25.8
-8.5
7.1
24.0
27.3
34.6
-4 .5

15.5
23.2
-2.5
13.3
17.3
8.4
6.8
-19.5

16.0
14.6
2.1
9.1
17.3
10.6
2.2
-18.0

13.9
23.8
-4 .8
12.3
15.8
10.7
4.8
-18.5

11.5
30.2
11.5
8.2
5.6
4.8
11.1
-5.7

11.4
22.6
-5 .0
10.0
18.8
14.5
13.6
-12.0

13.5
19.3
-3.8
9.9
19.3
8.2
8.6
-7.4

United
States

Canada

Japan

France

Germany

Italy

United
Kingdom

Belgium

Hourly compensation:
1960-81 .............................
1960-73 .............................
1973-81 .............................

6.9
5.0
9.6

8.7
6.4
11.1

14.8
14.6
9.7

11.9
9.2
15.1

10.1
9.3
9.4

16.2
12.3
19.8

13.1
8.6
19.1

12.6
10.7
12.1

13.2
11.8
12.5

1974 ....................................
1975 ....................................
1976 ....................................
1977 ....................................
1978 ....................................
1979 ....................................
1980 ....................................
1981....................................

10.6
11.9
8.0
8.3
8.3
9.7
11.8
10.2

15.8
14.2
14.2
11.0
6.7
10.1
9.1
11.1

31.2
17.0
6.7
9.7
5.9
6.5
6.5
7.4

19.6
19.0
14.1
13.7
12.7
13.8
16.6
16.5

15.0
12.4
7.8
10.5
8.5
7.3
8.6
7.5

24.6
28.9
19.8
18.8
14.5
17.6
18.5
22.3

25.0
29.9
17.2
12.6
16.5
18.9
23.6
16.2

22.5
21.4
13.2
12.0
8.0
7.7
9.6
9.6

Unit labor costs:
1960-81 .............................
1960-73 .............................
1973-81 .............................

4.1
1.9
7.7

4.8
1.8
9.5

5.1
3.5
2.7

6.1
3.1
10.0

4.6
3.7
4.7

9.8
5.1
15.5

9.2
4.1
16.6

1974 ....................................
1975 ....................................
1976 ....................................
1977 ....................................
1978 ....................................
1979 ....................................
1980 ....................................
1981....................................

13.3
8.8
3.4
5.7
7.4
9.0
11.6
7.2

13.3
17.2
8.4
6.7
5.0
8.3
12.8
10.7

28.1
12.6
-2.5
2.4
-1.8
-2.2
-.2
4.0

15.6
15.4
5.5
8.2
6.6
8.5
14.8
14.6

9.1
6.8
.6
5.3
5.0
2.4
7.0
4.7

18.7
34.9
10.4
17.5
11.2
9.6
12.1
18.3

Unit labor costs in U.S. dollars:
1960-81 .............................
1960-73 .............................
1973-81 ......................

4.1
1.9
7.7

4.4
1.9
6.5

7.9
4.9
7.2

6.5
2.8
9.4

9.1
6.1
9.1

1974 ....................................
1975 ....................................
1976 ....................................
1977 ....................................
1978 ....................................
1979 ....................................
1980 ....................................
1981....................................

13.3
8.8
3.4
5.7
7.4
9.0
11.6
7.2

15.8
12.7
11.9
-1 .0
-2.1
5.4
13.0
8.0

19.0
10.7
-2.4
13.3
26.2
-6.5
-3.5
6.7

6.7
29.6
-5.4
5.1
16.5
14.7
15.7
-10.5

11.9
12.3
-1.8
14.2
21.6
12.0
8.1
-15.7

Year

Note:

7.6 ■
5.4
8.1
6.2
34.5
-13.3
10.5
15.6
12.0
8.9
-10.6

Rates of change computed from the least squares trend of the logarithms of the index numbers.




137

Denmark Netherlands

France, there were no government restrictions on wage
increases during 1980-81, and wage rate increases
followed the consumer price index although there is no
formal indexation system. Minimum-wage increases
above the price index rate raised average wages further
in some lower wage industries. In Italy, the major wage
agreements were concluded in 1979 and expired in late
1981. Their wage rate provisions and the indexation
system were not limited, although there were discus­
sions of labor cost reductions and indexation changes
for 1982. In Italy and several other European countries,
actions were taken to cut employers’ social security tax
rates, although in other cases tax rates were raised to fi­
nance system deficits.
U n it labor co sts

Unit labor costs, which reflect the interplay between
hourly compensation and output per hour, increased
about 7 percent in the United States and 10 to 12 per­
cent in Canada, Sweden, and the United Kingdom in
1981, compared with more than 14 percent in France
and 18 percent in Italy, but only 2 to 5 percent in Den­
mark, Japan, Germany, Belgium, and the Netherlands.
(See table 3.)
In every country except Japan, France, Italy, and
Sweden, unit labor costs increased less in 1981 than in
the previous year. In the United Kingdom, the slow­
down from the 23 percent recorded in 1980 was sub­
stantial, and reflected both a smaller compensation
increase and a larger productivity gain. In most other
countries also, the moderation in unit labor costs re­
flects a slowdown in hourly compensation and improve­
ments in productivity. In France, the 1981 increase in
unit labor costs, as well as in productivity and hourly
compensation, was essentially the same as the previous
year’s. In Japan and Italy, the acceleration in unit labor
costs primarily reflects their productivity slowdowns.
The 1980-81 increases in unit labor costs were gener­
ally much smaller than those of 1974-75 because hourly
compensation gains were relatively moderate, in con­
trast to the substantial wage gains during the 1974-75
recession. The average annual unit labor cost increases
for the two periods are shown in the following tabula­
tion:
United States..................
Canada ...........................
Japan .............................
France ...........................
Germany ......................
Italy...............................
United Kingdom ...........
Belgium ........................
Denmark ......................
Netherlands ..................
Sweden...........................



1974-75
11.0
15.2
20.3
15.5
7.9
26.8
28.3
16.0
12.5
13.2
17.6

1980-81
9.4
11.8
1.9
14.7
5.9
15.2
16.3
4.2
6.4
2.9
10.9

138

For some countries— Japan, Belgium, Denmark, and
the Netherlands— the differences are substantial. Even
for the countries with the largest unit labor cost in­
creases in 1980-81— Italy and the United Kingdom—
the recent increases are down considerably from 1974—
75 peaks. The differences are less marked for the United
States and Germany, which had the smallest 1974-75
unit labor cost increases.
In U.S. dollars. In comparing trends in unit labor costs
among countries, an important analytical element is the
shift in relative currency values through international
exchange rate adjustments. In recent years, the number
and extent of such adjustments have been so great as to
constitute a major variable in competitive assessment.
The relationship between exchange rate shifts and
unit labor cost trends is partial and indirect but none­
theless important. The two are linked by the price
mechanism, a main determinant of trade directions and
competitive relationships. Because labor cost is the prin­
cipal cost factor in the production of manufactured
goods, it exerts a strong influence on the price at which
goods can be offered in international markets. Relative
changes in exchange rates alter the effect of relative
changes in costs in national currency. Consequently, in
assessing relative changes in unit labor costs in competi­
tive terms, changes in exchange rates need to be taken
into account.
Changes in currency exchange rates in 1981 had a
significant effect on relative changes in unit labor costs
measured in U.S. dollars. The dollar appreciated sub­
stantially— from about 15 percent to more than 30 per­
cent— relative to the European currencies. (By
September 1982, the dollar had further appreciated—
compared with the annual average for 1981 — 10 per­
cent versus the German mark and Dutch guilder, and 8
to 30 percent versus the other European currencies.)
The dollar also appreciated somewhat relative to the
Canadian dollar, but declined slightly versus the Japa­
nese yen. (By September 1982, however, the dollar had
appreciated 19 percent versus the yen, as well as anoth­
er 3 percent versus the Canadian dollar.)
Therefore, when measured in U.S. dollars, unit labor
costs in the European countries fell about 5 percent in
Sweden and the United Kingdom; 11 percent in France
and Italy; 16 percent in Germany; and 18 to 20 percent
in Belgium, the Netherlands, and Denmark. In U.S.
dollars, unit labor costs increased 8 percent in Canada
and 7 percent in Japan— about the same rate as for
U.S. costs. (See table 3.)
The largest contrast was between Japan and Germa­
ny. On a national currency basis, they had increases of
4 and 5 percent, respectively. On a U.S. dollar basis,
Japanese unit labor costs rose 7 percent while German
unit labor costs fell 16 percent.

lowing section introduces indexes of trade-weighted rel­
ative trends in manufacturing productivity, hourly com­
pensation, and unit labor costs in national currency, as
well as unit labor costs in U.S. dollars.
Because trade involves individual products, the use of
aggregate manufacturing measures as indicators of trade
competitiveness has certain limitations. In general, labor
productivity growth rates in export sectors probably ex­
ceed those for manufacturing as a whole. On the other
hand, hourly compensation tends to grow at similar
rates in all manufacturing sectors within a country.
Overall, therefore, trend measures for the total manu­
facturing sector would be expected to overstate, to some
extent, the growth of unit labor costs for the export sec­
tor. However, this would probably be true for every
country, and, in any case, the measures are intended to
represent relative changes only. In addition, exchange
rate changes have a significant effect on relative unit la­
bor cost developments, and these affect unit labor costs
in all manufacturing industries equally.

While the 5-percent decline in the United Kingdom
was not as large as in the other European countries, it
was the sharpest trend reversal among all the countries,
for British unit labor costs had increased 35 percent in
1980. Unit labor costs in Japan had posted a small de­
cline in 1980; among the other countries, they had risen
2 to 16 percent.
The trend in unit labor costs in U.S. dollars for the
1980-81 period differs significantly from that for the
years 1974-75 in most countries covered. First, unit la­
bor costs in national currency increased much less dur­
ing 1980-1981 in most countries. Secondly, the U.S.
dollar appreciated versus all European currencies and
the Canadian dollar in 1981, while in 1974— the dol­
75,
lar appreciated versus the Japanese yen, Italian lira, and
British pound but depreciated versus all the other cur­
rencies. Therefore, unit labor costs in U.S. dollars in­
creased substantially more in most other countries than
in the United States during the 1974— recession, while
75
in the 1980-81 period, unit labor costs in U.S. dollars
declined in all European countries covered.

Index calculation methods. The indexes of relative
trends in manufacturing productivity and labor costs
represent ratios of each country’s own indexes to
weighted geometric averages of the corresponding in­
dexes for the other 10 “competitor” countries.
The weights used to combine the other 10 countries’
indexes into an average “competitors” index reflect the
relative importance of each country as a manufacturing
trade competitor. The weights are those developed by
the IMF for computation of their own relative cost and
price indicators— except that they have been adjusted
from the 14-country coverage of the IMF series to the
11-country coverage of the BLS series.6 The weights are
based on disaggregated trade data for manufacturers in
1975. They take into account the relative importance of
each country’s trading partners in its direct bilateral
trade with them and the relative importance of those
partners in competition in “third country” markets, ad-

Relative productivity and cost trends
Indexes of manufacturing productivity and labor
costs are often used in analyses of changes in the rela­
tive competitive position of countries in the internation­
al trade of manufactures. Unit labor costs are an
important element in determining the underlying price
competitiveness of manufactured products, with relative
productivity and hourly compensation trends determin­
ing unit labor cost performance. The International
Monetary Fund (IM F) and Organization for Economic
Cooperation and Development ( o e c d ) publish indexes
for key cost and price measures—including unit labor
costs in U.S. dollars— which show the trend of each
country’s own indicators relative to those of other in­
dustrial (competitor) countries.5The BLS unit labor cost
measures are used in the computation of the IMF and
OECD indicators for most countries they cover. The fol­

Table 4.

Trade weights used to compute competitor indexes

[In percent]

Competitor country
Reference country

United S ta te s ...............................
C a nada........................................
Japan ..........................................
Belgium ........................................
D enm ark......................................
France ..........................................
G erm any......................................
Ita ly ...............................................
Netherlands.................................
Sw eden........................................
United Kingdom ...........................
Note:

United
States

Canada

Japan

Belgium

Denmark

France

Germany

Italy

Netherlands

Sweden

United
Kingdom

76.9
36.2
5.7
12.7
16.7
17.5
16.3
11.9
18.0
25.0

19.3
—
2.9
.5
.9
1.1
1.5
1.4
.7
3.5
2.0

17.3
5.1
—
6.2
10.3
11.9
12.1
12.2
9.1
11.6
11.6

3.3
.9
3.8
—
2.9
4.0
7.8
4.5
8.7
3.3
5.4

1.1
.2
1.4
.9
—
1.3
1.2
1.4
1.5
4.8
2.1

13.1
2.5
11.3
22.9
9.6
—
21.0
10.8
16.5
10.3
13.7

18.8
5.3
18.2
34.1
23.4
31.1
—
34.3
33.9
23.4
22.5

7.4
1.7
7.4
7.7
6.4
13.3
12.8
—
4.2
6.5
7.8

4.9
.9
4.4
9.5
4.7
5.0
8.1
4.9
—
3.8
5.3

3.2
2.0
3.7
2.4
13.2
3.0
5.3
2.8
2.7
—
4.7

11.6
4.5
10.8
10.1
15.9
12.5
12.8
11.5
10.7
14.8
—

Because of rounding, sums of individual items may not equal 100.0.




139

Relative productivity trends. The countries in which
manufacturing productivity grew more rapidly than that
of trade competitors since 1970 were Japan, Belgium,
the Netherlands, and Denmark. Productivity had risen
11 to 12 percent more in Denmark and the Netherlands
and 16 percent more in Japan and Belgium by 1976. By
1981, their relative trends had diverged: For Japan, pro­
ductivity gains were 41 percent higher and for Belgium,
29 percent, while in Denmark and the Netherlands the
gains were 12 and 13 percent higher.

justed for the importance of foreign trade to the manu­
facturing sector as a whole in each country.7 Table 4
shows the weights used for each of the 11 countries.
The relative indexes of output per hour, hourly com­
pensation, and unit labor costs in national currency and
in U.S. dollars are shown in table 5. The underlying
“own country” and “competitor countries” indexes
used to compute the relative indexes, and indexes of
trade-weighted exchange rates, not shown in table 5, are
available from the authors.

Table 5. Relative indexes of output per hour, hourly compensation, and unit labor costs in manufacturing, 11 countries,

1970-81

[1970 = 100]
United
States

Canada

Japan

France

Germany

Italy

United
Kingdom

Belgium

Denmark

Netherlands

Sweden

100.0
100.7
98.5
96.5
91.0
92.6

100.0
101.2
100.0
100.3
103.6
98.4

100.0
101.0
105.9
109.5
110.2
112.3

100.0
100.4
98.9
96.9
97.2
98.4

100.0
98.9
98.0
96.2
98.7
102.9

100.0
98.1
99.0
103.8
105.8
98.1

100.0
98.7
99.4
98.4
96.5
92.1

100.0
101.4
105.6
109.2
111.1
113.4

100.0
101.2
102.2
105.2
105.5
114.6

100.0
101.7
102.4
105.5
110.6
105.4

100.0
100.0
98.2
98.3
99.2
96.5

........................................
........................................
........................................
........................................
........................................
........................................

90.3
88.8
85.9
82.3
81.0
81.1

98.7
99.8
99.6
99.4
95.4
93.0

115.9
120.4
126.3
132.9
140.3
140.5

99.6
101.0
103.0
103.0
102.3
100.6

102.8
104.2
103.2
103.1
102.1
101.6

99.6
96.6
95.8
98.3
102.1
102.3

89.5
87.5
87.1
86.0
84.8
87.4

116.5
119.2
120.0
121.7
123.1
128.5

112.0
110.7
109.1
109.8
109.2
111.9

111.2
111.0
113.9
113.9
113.1
112.9

91.4
86.6
87.2
90.4
89.8
87.0

Hourly compensation:
1970 ........................................
1971 ........................................
1972 ........................................
1973 ........................................
1974 ........................................
1975 ........................................

100.0
94.2
88.7
82.4
75.4
71.4

100.0
99.8
99.9
100.7
103.0
103.5

100.0
104.8
110.5
120.8
136.3
136.5

100.0
99.3
99.2
98.4
98.7
99.7

100.0
99.8
98.7
96.3
91.8
86.1

100.0
103.0
106.1
118.3
124.4
137.9

100.0
102.9
105.1
102.5
108.7
121.7

100.0
101.8
105.6
107.4
110.5
114.2

100.0
102.2
101.3
107.8
109.3
109.9

100.0
101.8
104.6
109.4
109.7
106.7

100.0
100.2
100.7
98.8
97.4
100.3

........................................
........................................
........................................
........................................
........................................
........................................

68.6
66.6
65.8
65.2
65.3
64.5

108.4
110.4
108.2
108.3
105.5
106.1

130.6
129.4
124.3
119.2
112.5
108.1

102.3
104.4
107.0
110.2
114.8
120.5

82.0
80.9
79.2
75.9
72.6
69.0

149.7
161.0
168.3
180.4
192.1
213.5

128.8
130.8
139.5
151.2
168.4
176.6

115.6
115.9
113.2
109.9
107.2
105.3

109.2
108.7
108.4
109.4
107.8
105.6

108.1
105.4
104.1
101.7
95.2
90.2

106.8
105.0
106.1
103.1
101.6
102.7

Unit labor costs in national
currency:
1970 ........................................
1971 ........................................
1972 ........................................
1973 ........................................
1974 ........................................
1975 ........................................

100.0
93.6
90.0
85.4
82.9
77.1

100.0
98.6
99.9
100.4
99.4
105.2

100.0
103.7
104.4
110.3
123.6
121.5

100.0
98.9
100.3
101.6
101.6
101.3

100.0
101.0
100.7
100.1
93.0
83.6

100.0
105.0
107.2
113.9
117.7
140.6

100.0
104.3
105.8
104.2
112.7
132.1

100.0
100.4
100.0
98.4
99.5
100.7

100.0
101.0
99.1
102.5
103.6
95.9

100.0
100.1
102.1
103.7
99.1
101.3

100.0
100.2
102.4
100.5
98.2
103.9

........................................
........................................
........................................
........................................
........................................
........................................

76.0
75.0
76.6
79.3
80.5
79.6

109.9
110.6
108.6
108.9
110.6
114.0

112.8
107.5
98.4
89.7
80.2
76.9

102.7
103.4
103.9
107.0
112.2
119.8

79.7
77.7
76.7
73.7
71.1
68.0

150.3
166.5
175.6
183.5
188.1
208.6

143.9
149.6
160.1
175.7
198.5
202.2

99.2
97.2
94.3
90.3
87.1
82.0

97.5
98.2
99.3
99.7
98.7
94.3

97.2
94.9
91.4
89.3
84.2
79.9

116.9
121.2,
121.7
114.1
113.2
118.1

Unit labor costs in U.S. dollars:
1970 ........................................
1971 ........................................
1972 ........................................
1973 ........................................
1974 ........................................
1975 ........................................

100.0
91.1
81.5
71.1
70.9
64.9

100.0
101.3
102.6
99.9
102.1
103.3

100.0
105.2
116.1
129.0
136.6
128.8

100.0
96.6
100.3
106.0
99.9
110.2

100.0
104.3
106.4
119.2
118.6
109.1

100.0
103.7
104.6
100.3
93.8
108.5

100.0
104.2
101.2
88.9
93.8
100.8

100.0
100.0
102.0
101.4
104.9
107.5

100.0
100.0
97.7
107.0
109.8
105.0

100.0
101.1
104.4
109.9
111.2
116.4

100.0
99.5
102.6
101.1
99.4
110.3

68.1
66.4
61.7
61.8
62.4
70.8

113.5
105.2
93.5
90.2
91.5
95.4

127.4
133.5
148.2
123.6
105.7
118.7

107.6
102.2
101.0
105.1
110.3
106.6

111.0
116.9
122.7
124.3
120.2
107.8

95.9
96.1
94.2
95.5
94.1
91.1

93.5
91.5
98.4
115.4
143.2
146.0

109.1
112.2
111.1
107.2
102.8
92.5

109.2
108.6
109.0
108.8
99.2
87.6

113.9
116.1
113.7
113.2
106.7
95.6

126.4
125.0
112.4
106.1
105.9
106.8

Year

Output per hour:
1970 ........................................
1971 ........................................
1972 ........................................
1973 ........................................
1974 ........................................
1975 ........................................
1976
1977
1978
1979
1980
1981

1976
1977
1978
1979
1980
1981

1976
1977
1978
1979
1980
1981

1976
1977
1978
1979
1980
1981
Note:

........................................
........................................
........................................
........................................
........................................
........................................

Relative indexes are calculated from the ratio of the reference country index to a trade-weighted average index for the other 10 countries.




140

In France, Germany, and Italy, productivity in­
creased at about the same rate as that of trade competi­
tors from 1970 to 1981. Their relative rates of change
varied during the period, however. In the early 1970’s,
productivity in France and Germany rose somewhat
less rapidly, and in Italy it rose more rapidly, but dur­
ing the late 1970’s, the relative rates were reversed.
Productivity rose less rapidly than in competitor
countries for the United States, Canada, Sweden, and
the United Kingdom. From 1970 to 1981, U.S. relative
productivity had increased 19 percent less, while in
Sweden and the United Kingdom, gains were 13 per­
cent lower, and in Canada, 7 percent lower. The slower
gains were quite consistent throughout the entire peri­
od.
Relative compensation trends. Hourly compensation rose
less than in competitor countries in the United States,
Germany, and the Netherlands. From 1970 to 1981,
compensation increased about 35 percent less in the
United States, 30 percent less in Germany, and 10 per­
cent less in the Netherlands. For the United States and
Germany, the slower relative trend was fairly consistent
over the whole period. For the Netherlands, however,
compensation rose more rapidly than competitors’ dur­
ing the early 1970’s, then less rapidly after 1976, with
the greatest relative declines occurring in 1980-81, fol­
lowing the imposition of wage controls.
Hourly compensation rose more rapidly than in com­
petitor countries in Italy, the United Kingdom, Japan,
and France. From 1970 to 1981, compensation had in­
creased about 100 percent more in Italy and about 75
percent more in the United Kingdom. Almost without
exception, both had consistently larger gains than their
competitors throughout the 1970-81 period. Hourly
compensation in Japan rose more rapidly during the
early 1970’s—by 1975, Japanese compensation had in­
creased about 35 percent more than that of competitors
— but grew less rapidly after 1975. By 1981, Japanese
compensation gains were only 8 percent higher than
competitors’. In France, hourly compensation rose at
about the same rate as in competitor countries until the
mid-1970’s, then rose more rapidly to end in 1981 with
about a 20-percent larger cumulative increase.
Canada, Belgium, and Denmark also ended the 1970—
81 period with somewhat larger compensation increases.
But in each country, the 1981 relative gains were down
from previous peaks— in Canada, 6 percent down from
10 percent in 1977; in Belgium, 5 percent down from 16
percent in 1976-77; and in Denmark, 6 percent down
from 9 percent in 1974-79. In Sweden, hourly compen­
sation generally rose at about the same rate as competi­
tor countries’ over the 1970-81 period.
Relative unit labor cost trends. Unit labor costs in
national currency increased less from 1970 to 1981 in



141

six countries— the United States, Japan, Germany, Bel­
gium, Denmark, and the Netherlands— than in their
competitor countries. The relative trend was 6 percent
lower in Denmark by 1981, and about 20 to 30 percent
lower in the other countries.
The relative change for the United States was down
because hourly compensation had fallen more than out­
put per hour. In Japan, Belgium, and Denmark, relative
productivity gains more than offset relative compensa­
tion increases; in Germany, the relative productivity
trend was about level, but relative compensation was
sharply down; and the Netherlands had both productiv­
ity and hourly compensation advantages.
The relative trend for the United States was steadily
downward from 1970 to 1977, up moderately from
1977 to 1980, and down again slightly in 1981. Relative
unit labor costs in Japan rose over 20 percent more
than those of competitors by 1974-75, then declined
steadily to 23 percent less than competitors’ by 1981.
Relative unit labor costs declined steadily in Germany
from 1973, in Belgium and the Netherlands from 1975,
and in Denmark from 1979. For the Netherlands, the
most significant relative cost declines occurred during
1980 and 1981.
Unit labor costs in national currency increased by at
least 100 percent more than competitors’ in Italy and
the United Kingdom and by about 15 to 20 percent
more in Canada, France, and Sweden, The large relative
increases in Italy and the United Kingdom are attribut­
able to hourly compensation gains as the relative pro­
ductivity trend was down in the United Kingdom and
essentially level in Italy. In Canada and France, hourly
compensation was up slightly, and the productivity
trend was down in Canada and even in France. In Swe­
den, hourly compensation trends were equal to those of
competitors, but productivity fell from 1970 relative
levels.
In U.S. dollars. AJter adjustment for the relative change
in the foreign exchange rate of the dollar, U.S. unit la­
bor costs showed a decline of nearly 30 percent versus
those of competitors from 1970 to 1981, compared with
about 20 percent in national currency. In 1980, relative
unit labor costs adjusted for the dollar exchange rate
were down almost 40 percent. However, the U.S. dollar
appreciated 10 percent against trade-weighted U.S.
competitor currencies from 1980 to 1981. This primarily
reflected the dollar’s appreciation relative to the Ger­
man mark, French franc, and British pound, because,
on a trade-weighted basis, the 2.5-percent appreciation
of the Japanese yen was balanced by a 2.5-percent de­
preciation of the Canadian dollar.
Unit labor costs adjusted for relative exchange rates
for Canada, Italy, Belgium, the Netherlands, and Den­
mark were also down— 5 to 12 percent— versus com­
petitors. For Canada, a 16-percent decline in the

exchange rate, primarily against the U.S. dollar, offset
higher increases in unit labor costs in Canadian dollars.
For Italy, the exchange rate posted a 55-percent decline
versus U.S. and German currencies. On the other hand,
trade-weighted exchange rates were up 13 and 20 per­
cent for Belgium and the Netherlands; therefore, rela­
tive unit labor costs in dollars declined less than in na­
tional currency terms.
For Germany and Japan, unit labor costs in U.S. dol­
lars increased 8 and 19 percent more than those of trade
competitors (principally the United States for Japan, and
France and the United States for Germany) even though
unit labor costs in national currency were down about 25

to 30 percent, because their relative exchange rates rose
55 to 60 percent over the 1970-81 period.
In the United Kingdom, relative unit labor costs in­
creased 100 percent in national currency terms, but 46
percent in U.S. dollars, because the British pound de­
clined 28 percent overall against competitor currencies—
primarily the dollar and the German mark. In France
and Sweden, unit labor costs in U.S. dollars posted 197081 relative increases of 7 percent, as costs in national cur­
rency rose nearly 20 percent more than those of competi­
tors, but trade-weighted exchange rates declined about
10 percent versus competitor currencies.

1The Federal Republic plus West Berlin.
2The data relate to all employed persons, including the selfemployed, in the United States and Canada, and to all wage and sala­
ry employees in the other countries. Hours refer to hours paid in the
United States, hours worked in the other countries.
Compensation includes all payments made by employers directly to
their employees (before deductions), plus employer contributions to
legally required insurance programs and to contractual and private
welfare plans for the benefit of employees. Labor costs include, in ad­
dition to compensation, employer expenditures for recruitment and
training; the cost of cafeterias, medical facilities, and other plant facil­
ities and services; and taxes (other than social security taxes, which
are part of compensation) levied on payrolls or employment rolls. An­
nual data are not available for total labor costs. As used in this arti­
cle, labor costs approximate more closely the concept of
compensation. However, compensation has been adjusted to include
all significant changes in taxes that are regarded as labor costs. For
the United States and Canada, compensation of self-employed work­
ers is measured by assuming that their hourly compensation is equal
to the average for wage and salary employees.
? Percent changes for 1960-81, 1960-73, and 1973-81 shown in the
tables are computed using the least squares method— that is, from
the least squares trend of the logarithms of index numbers— in order
to remove much of the effect of cyclical changes on the average rates
of change, and thereby estimate the underlying trends.
4To compute the series for the eight European countries and 10
foreign countries, the data have been combined by aggregating the
output, compensation, and hours figures for each year, adjusting
where necessary for compatibility of coverage and concept. Average
exchange rates for 1974-81 were used to aggregate the output and
compensation data. The use of 1974-81 exchange rates, however, does
not imply that these rates reflect the comparative real value of curren­




142

cies for manufacturing output. Moreover, the use of exchange rates
for a different period would have little effect on the combined series.
5The IMF publishes annual and quarterly indexes of relative unit
labor costs and relative normalized unit labor costs in manufacturing
— as well as relative value-added deflators, relative wholesale prices,
and relative export unit values in manufacturing— for 14 industrial
countries, in their monthly statistical publication International Finan­
cial Statistics. The OECD publishes quarterly indexes in chart form of
relative unit labor costs in manufacturing, relative export unit values
(prices) for manufactures, and relative consumer prices for 15 indus­
trial countries in their monthly statistical publication Main Economic
Indicators.
Series descriptions, data sources, and compilation methods for the
IMF measures are described in “Intercountry Cost and Price Com­
parisons,” a paper by Michael C. Deppler, Research Department, In­
ternational Monetary Fund (November 1979); the OECD measures
are described in The International Competitiveness of Selected OECD
Countries, OECD Economic Outlook Occasional Studies, July 1978.
‘ The IMF weights were derived from disaggregated 5-digit Stan­
dard International Trade Classification data (up to 1,400 individual
commodity classes) for each of the 14 countries covered by their se­
ries. The IMF weights have been simply adjusted to the 11-country
BLS comparative series by eliminating the weights for the three un­
covered countries— Austria, Norway, and Switzerland— and propor­
tionately increasing the weights for the remaining 11 countries so that
they equal 100 percent. The result should be little different from a
comprehensive reweighting based on trade data for the 11 countries
alone, because the omitted countries account for no more than 8.1
percent of the total 14-country weight for any of the 11 countries, and
for a total of only 4 percent in the case of the United States.
’ The weighting system is described in detail in Deppler, “Intercountry Cost and Price Comparisons.”

THE INTERNATIONAL
CONTEXT
Arnold Packer and
Arthur Neef

ro d u c tiv ity g ro w th is a
m ajo r d e te r m in a n t of a n a ­
tio n 's a b ility to c o m b a t
in fla tio n , to m a in ta in c o m ­
p e titiv e n e s s in in te rn a ­
tio n al tra d e , a n d to in c re a se
its s ta n d a rd of liv in g . G ro ss
d o m e s tic p ro d u c t (G D P) p e r e m p lo y e d
p e rs o n , w h ic h covers th e e c o n o m y 's
total o u tp u t of g o o d s a n d se rv ic e s,
p ro v id e s a b ro a d m e a s u re of la b o r p ro ­
d u c tiv ity . O v e r th e p e rio d 1960 to 1979,
real G D P p e r e m p lo y e d p e rs o n in ­
creased at a n a n n u a l av e ra g e rate of 1.5
p e rc e n t p e r y e a r in th e U n ite d S tates.
T his c o m p a re s w ith a n n u a l a v e ra g e
g a in s of a ro u n d 2 p e rc e n t in C a n a d a
a n d th e U n ite d K in g d o m , 4 p e rc e n t in
B elgium , F rance, G e rm a n y , th e
N e th e rla n d s , a n d Italy, a n d 7 p e rc e n t
in Jap an . R o u g h ly c o m p a ra b le fig u re s
for so m e ra p id ly d e v e lo p in g
e c o n o m ie s are 5.4 p e rc e n t for K orea,
5.8 p e rc e n t for T a iw a n , a n d 3.3 p e rc e n t
fo r M exico (1960-1977).
S in ce 1973, th e a v e ra g e ra te of g a in
in real G D P p e r e m p lo y e d p e rs o n in
th e U n ite d S tates h a s b e e n o n ly 0.3
p e rc e n t p e r y ear; in th e p e rio d fro m
1960 to 1973 it w as 2.1 p e rc e n t p e r y ear.
S in ce 1973, o th e r in d u s tria l c o u n trie s
h a v e also e x p e rie n c e d p ro d u c tiv ity

P

Arnold Packer is the assistant secretary
for policy, evaluation, and research in
the U.S. Department of Labor and
Arthur Neef is chief of foreign labor
statistics and trade in the Bureau of
Labor Statistics. Brian Brosnahan
assisted substantially in the preparation
o f this article.

slo w d o w n s — so m e as s h a rp as th e
U n ite d S ta te s' s lo w d o w n . W ith th e ex­
c e p tio n of C a n a d a , h o w e v e r, all c o n ­
tin u e d to h a v e la rg e r a n n u a l a v erag e
g a in s in G D P p e r e m p lo y e d p e rso n .

p e rs o n b e c a u se of sig n ific a n t re d u c ­
tio n s in a v e ra g e w o rk in g h o u rs . T h e
U .S . ra te of g ro w th sin c e 1973 of 1.4
p e rc e n t p e r y ear— c o m p a re d w ith 3
DP p e r e m p lo y e d
p e rc e n t p e r y e a r fro m 1960 to 1973—
p e rs o n d o e s p ro ­
e x c e e d e d th e ra te of la b o r p ro d u c tiv ity
v id e a m e a s u re of
g a in in th e U n ite d K in g d o m , b u t w as
e c o n o m y -w id e
less th a n in a n y of th e o th e r c o u n trie s
p ro d u c tiv ity , b u t it c o m p a re d .
also h a s so m e
W h ile th e U n ite d S ta te s h a s h a d th e
sh o rtc o m in g s. Let
slo w e st ra te of p ro d u c tiv ity g ro w th , it
u s b rie fly n o te th e m ostill exv a n t.s G n y o th e r c o u n try in
st re le ceed a D P
c o v ers th e area of p u b lic a d m in is tra ­
o v erall efficien cy . A s of 1979, G D P p e r
tio n , th o u g h th e U n ite d S tates a n d
e m p lo y e d p e rs o n in C a n a d a w as
m o st o th e r c o u n trie s a s s u m e z ero p r o ­ a b o u t 95 p e rc e n t of th e U .S . level; in
d u c tiv ity g ro w th fo r th is secto r b e ­
F ran ce, G e rm a n y , a n d th e B enelux
c a u se of m e a s u re m e n t d iffic u ltie s.
c o u n trie s, it w a s a ro u n d 90 p e rc e n t; in
G D P d o e s n o t tak e in to a c c o u n t
Ja p a n , tw o -th ird s of th e U .S. level;
c h a n g e s in a v e ra g e h o u rs w o rk e d . A n d a n d in Italy a n d th e U n ite d K in g d o m ,
it in c lu d e s th e effects of re so u rc e sh ifts 60 p e rc e n t. B ecau se of th e fa ste r ra te s
a m o n g se c to rs w ith v e ry d iffe re n t
of p ro d u c tiv ity g ro w th a b ro a d , th e re
lev els of p r o d u c tiv ity (all th e c o u n trie s
h a s b e e n a n a rro w in g — in so m e cases
m e n tio n e d a b o v e h a v e u n d e rg o n e
a v e ry s u b s ta n tia l n a rro w in g — of th e
sh ifts in re la tiv e e m p lo y m e n t as w o rk ­ p ro d u c tiv ity g a p . In 1960, th e c o rre ­
ers m o v e d b e tw e e n a g ric u ltu re , in d u s ­ s p o n d in g fig u re s w e re , C a n a d a , 90
try , a n d serv ices).
p e rc e n t of th e U .S . level; F ran ce, G e r­
F or so m e p u rp o s e s of a n a ly z in g in ­
m a n y , th e B en elu x c o u n trie s , a n d th e
te rn a tio n a l c o m p e titiv e n e s s , m a n u fa c ­ U n ite d K in g d o m , a b o u t 55 to 60 p e r ­
tu r in g p ro d u c tiv ity , as m e a s u re d b y
cen t; Italy , 35 p e rc e n t; a n d J a p a n , o n ly
o u tp u t p e r h o u r, is a b e tte r in d ic a to r.
25 p e rc e n t. U n fo rtu n a te ly , s im ila r level
M a n u fa c tu rin g p ro d u c tiv ity e x h ib its a c o m p a ris o n s fo r m a n u f a c tu r in g a re n o t
p a tte rn sim ila r to G D P p e r e m p lo y e d
a v a ila b le .
p e rs o n , w ith th e U .S . s h o w in g th e
T h e m u c h lo w e r re la tiv e lev els of
sm a lle st a v e ra g e ra te of g a in sin ce
p ro d u c tiv ity in Ja p a n a n d E u ro p e in
1960. A lso , all th e W e ste rn in d u s tr i­
1960 p a rtia lly e x p la in th e ir fa s te r ra te s
a liz e d c o u n trie s sh o w a m a n u fa c tu rin g of p ro d u c tiv ity g ro w th . C o m p a ra tiv e
p ro d u c tiv ity slo w d o w n sin c e 1973—
ra te s of p ro d u c tiv ity g ro w th b e tw e e n
a lth o u g h , for m o s t c o u n trie s, a less
th e U n ite d S ta te s a n d c o u n trie s a p ­
sig n ific a n t slo w d o w n th a n fo r G D P.
p ro a c h in g th e U .S . o v erall lev el o f effi­
cien c y s h o u ld , p re s u m a b ly , n a rro w .
F or n a tio n s su c h as F ran ce, W e st G e r­

C

Reprinted from Executive, Vol. 7, No. 1,
Graduate School o f Business and Public Administration,
Cornell University.




m a n y , a n d th e B en elu x c o u n trie s , o u t­
p u t p e r h o u r in m a n u fa c tu rin g slo w e d
m u c h less th a n G D P p e r e m p lo y e d

143

H o w e v e r, m a n y o th e r fa cto rs also af­
fect re la tiv e ra te s of p ro d u c tiv ity g ain .

G e rm a n y a n d th e N e th e rla n d s sh o w
th e h ig h e s t a v e ra g e ra te s of g a in —
a b o u t 9 p e rc e n t p e r y ear. In th e p e rio d
sin c e 1973, C a n a d a also h a d a sm a lle r
g a in th a n th e U n ite d S tates in u n it
la b o r co sts as m e a s u re d on a U .S. d o l­
lar b a sis. H o w e v e r, Ja p a n , G e rm a n y ,
B e lg iu m , a n d th e N e th e rla n d s , w h ic h
h a d sm a lle r n a tio n a l c u rre n c y -b a se d
ra te s of in c re a se th a n th e U .S. in c re a se
of 8 p e rc e n t p e r y e a r, sh o w a v erag e
g a in s of a b o u t 9 to 12 p e rc e n t p e r y ear.
L a b o r p ro d u c tiv ity g a in s are th e
p rin c ip a l m e a n s of ra is in g liv in g s ta n ­
d a rd s . In c re a se s in o u tp u t p e r h o u r
can b e tra n s la te d in to g re a te r o u tp u t
of g o o d s a n d se rv ices o r in to in c re a se d
le is u re tim e . C h a n g e s in real h o u rly
c o m p e n s a tio n are o n e m e a s u re of
c h a n g e s in liv in g s ta n d a rd s . For th e
e c o n o m y as a w h o le , real h o u rly c o m ­
p e n s a tio n of w o rk e rs can o n ly in c re a se
if o u tp u t p e r h o u r in c re a se s o r la b o r
re c e iv e s a n in c re a sin g p ro p o r tio n of
to ta l fa c to r in c o m e s. O f c o u rse , i n d i­
v id u a l w o rk e rs o r w o rk e rs in p a rtic u la r
in d u s tr ie s can also re c e iv e real g a in s in
h o u rly c o m p e n s a tio n at th e e x p e n se of
o th e r w o rk e rs.
R eal h o u rly c o m p e n s a tio n of U .S.
m a n u fa c tu rin g w o rk e rs, u sin g th e
c o n s u m e r p ric e in d e x as th e p rice
m e a s u re , ro se a t a n a v e ra g e ra te of 1.4
p e rc e n t p e r y e a r fro m 1960 to 1979. In
th e p e rio d u p to 1973, real h o u rly
c o m p e n s a tio n ro se 1.7 p e rc e n t p e r
y e a r; fro m 1973 to 1979, o n ly 0.8 p e r ­
c e n t p e r y e a r. W h ile th e g a in s in real
h o u rly c o m p e n s a tio n w e re less th a n
e ith e r th e to tal e c o n o m y o r m a n u fa c ­
tu r in g p ro d u c tiv ity g a in s , th e p a tte rn
w a s sim ila r— a su b s ta n tia lly re d u c e d
ra te of in c re a se in th e p e rio d o f th e
p ro d u c tiv ity slo w d o w n .
M a n u fa c tu rin g w o rk e rs in all of th e
o th e r c o u n trie s h a d m u c h la rg e r in ­
creases in real h o u rly c o m p e n s a tio n ,
ra n g in g , o v e r th e 1960 to 1979 p e rio d ,
fro m d o u b le th e U .S . a v e ra g e ra te of

e s p ite th e fact th a t
th e U .S . h a d th e
lo w e st ra te of p r o ­
d u c tiv ity g ro w th
am ong C anada,
Ja p a n , a n d all
W e s te rn E u ro p e a n
c o u n trie s from 1960 to 1979, m a n u fa c ­
tu rin g u n it la b o r co sts ro se less in th e
U n ite d S tates th a n in a n y of th e o th e r
c o u n trie s. A n d for g o o d re a so n . T h ese
c o u n trie s h a d larg e r g a in s in h o u rly
c o m p e n s a tio n — in Ja p a n a n d E u ro p e ,
ro u g h ly d o u b le th e U .S. a v e ra g e
yearly in c re a se of a b o u t 6V2 p e rc e n t.
U .S. u n it la b o r costs ro se at a n av erag e
rate of 3.7 p e rc e n t p e r y e a r, c o m p a re d
w ith a b o u t 4 p e rc e n t p e r y e a r in
C a n a d a , 4.5 p e rc e n t p e r y e a r in Ja p a n ,
B elg iu m , a n d G e rm a n y , 5.5 p e rc e n t in
th e N e th e rla n d s , 6 p e rc e n t in F ran ce,
a n d a ro u n d 9 p e rc e n t in Italy a n d th e
U n ite d K in g d o m .
W h ile re lativ e c h a n g e s in n o m in a l
u n it la b o r costs p o te n tia lly affect tra d e
c o m p e titiv e n e s s , u n it la b o r co sts a d ­
ju s te d for c h a n g e s in e x c h a n g e ra te s
are th e m o re re le v a n t m e a s u re . O v e r
th e y e a rs, p a rtic u la rly sin c e th e 1971
a n d 1973 d e v a lu a tio n s of th e U .S . d o l­
lar, th e re h av e b e e n s u b s ta n tia l sh ifts
in cu rre n c y v a lu a tio n s — w h ic h h a v e a
s ig n ific a n t im p a c t on ra te s of c h a n g e
in u n it la b o r costs. In g e n e ra l, th e
cu rre n c ie s of Ja p a n , G e rm a n y , a n d th e
sm aller E u ro p e a n c o u n trie s h a v e a p ­
p re c ia te d re la tiv e to th e U .S . d o lla r,
w h ile th e B ritish a n d Ita lia n c u rre n ­
cies h a v e d e p re c ia te d . F ro m 1960 to
1979 th e F ren ch fran c h a s also a p ­
p re c ia te d s o m e w h a t w h ile th e C a n a ­
d ia n d o llar, b e c a u se of s u b s ta n tia l
d e c lin e s in th e p a st few y e a rs, h as
d e p re c ia te d .
A fter a d ju s tm e n t fo r th e s e sh ifts in
ex ch a n g e ra te s , C a n a d a sh o w s a lo w er
ra te of g a in in u n it la b o r c o sts sin c e
g a in in C a n a d a to o v e r fo u r tim e s th e
1960 th a n th e U n ite d S ta te s w h ile
U .S. av e ra g e ra te of g a in in Ja p a n a n d




144

th e c o n tin e n ta l E u ro p e a n c o u n trie s.
B ut b e c a u se of s u b s ta n tia l re d u c tio n s
in a v e ra g e w o rk in g h o u rs in Ja p an
a n d E u ro p e , in c re a se s in real h o u rly
c o m p e n s a tio n o v e rs ta te th e in c re a se s
in real in c o m e fro m w o rk . Still, th e
in c re a se s w e re n o tic e a b ly g re a te r th a n
in th e U n ite d S tates.
W o rld W ar II left m u c h of th e in d u s ­
tria liz e d w o rld in ru in s . In m o s t of th e
m a jo r c o u n trie s , a n d in th e fo rm er
A xis p o w e rs in p a rtic u la r, th e p ro cess
of e c o n o m ic d e v e lo p m e n t w a s e ith e r
a rre ste d o r se t b a c k m a n y y ea rs. In
c o n tra st, th e w a r g re a tly a c c elerated
th e d e v e lo p m e n t of th e U .S . ec o n o m y .
S in ce th e U .S . e c o n o m y w as a lre a d y
th e m o st h ig h ly d e v e lo p e d b e fo re th e
w a r s ta rte d , in 1945 p r o d u c tiv ity a n d
p e r c a p ita in c o m e in th e U .S . w e re far
h ig h e r th a n in o th e r c o u n trie s.
In 1945, th e U n ite d S ta te s w a s a t th e
fro n tie r of te c h n o lo g y . B ut th e secrets
to h ig h e r p ro d u c tiv ity te c h n o lo g y
c o u ld b e b o u g h t fro m th e U .S.
th ro u g h p a te n t lic e n se s a n d o th e r
te c h n ic a l a g re e m e n ts .
A c o m p a riso n of th e re le v a n t c o u n ­
trie s ' im p o rts a n d e x p o rts of te c h n o l­
ogy w ill illu s tra te th e p o in t. In 1971
th e ra tio of Ja p a n e se p a y m e n ts for im ­
p o rts of te c h n o lo g y (p a te n t lic e n ses,
k n o w -h o w , a n d a ss o c ia te d e x p en ses)
to Ja p a n e se re c e ip ts fo r te c h n o lo g y
ex p o rts w as a b o u t e ig h t to o n e. In
G e rm a n y , th is ra tio w as n e a rly th re e
to o n e , a n d in F rance it w a s close to
tw o to o n e . In th e m o re m a tu re
e c o n o m ie s of G re a t B rita in a n d th e
U .S ., h o w e v e r, re c e ip ts o u tw e ig h e d
p a y m e n ts. F or B rita in , th e d ifferen ce
w a s sm all; b u t U .S . p a y m e n ts for
te c h n o lo g y sto o d at o n ly $218 m illio n
w h ile re c e ip ts w e ig h e d in a t a w h o p ­
p in g $2.5 b illio n .
In m a n y fie ld s, th e U .S . is still
o p e ra tin g on th e te c h n o lo g ic a l fr o n ­
tie r. A s in d u s tria liz e d c o u n trie s a p ­
p ro a c h te c h n o lo g ic a l p a rity , th e s p e c ­
ta c u la r g a in s a tta in e d in G e rm a n y a n d
Ja p a n w ill c o n tin u e to d im in is h .

Part V. Technology
Studies

One of the important efforts of b l s has been to ex­
amine the impact of technological change upon produc­
tivity, employment, and occupational change. This part
describes the scope of this research, including the
sources and data collection methods. A study of the
employment effects of new electronic technology is in­
cluded, together with a representative series of studies
of technological change and its effect upon labor and
other economic variables. Projections of possible future
effects of specific technologies on employment and oc­
cupational structure are also included.

Background
Studies of technological changes and their labor im­
plications have been undertaken by b l s over the years
for a variety of purposes. During the 1930’s, public in­
terest focused on the unemployed, and reports were
prepared on technological changes and displacement of
workers in various industries. During World War II,
emerging technologies were studied for purposes of im­
proving manpower utilization.
Beginning in the mid-1950’s, nationwide attention
was focused on the implications of new developments
classified under the general term “ automation.” b l s
made a series of studies on a plant basis, in the in­
surance, petroleum refining, bakery, air lines, and elec­
tronics industries, to explore the manpower implications
of various changes. Later, broader studies were under­
taken, including a survey of the manpower impact of
changeover to electronic computers in 20 large com­
panies and intensive studies of technological change in
the coal and paper industries.
These studies formed the basis, beginning in the early
1960’s, for a more systematic investigation of future
changes. R esearch now underw ay p inpoints
technologies which will become increasingly important
over the next decade in key industries and attempts to
provide advance information about their manpower im­
plications.

Analysis and Interpretation
For a better understanding of research results in this
field, it is important to keep in mind the meaning of cer­
tain key ideas and concepts. Some of the problems of in­
terpretation and analysis in this type of research are,
therefore, set forth briefly.
Definition of technological change
Technological change is defined broadly in the b l s
studies as encompassing significant changes in processes
and equipment, and product and services produced, and
materials, fuels, and energy used. The term “ automa­
tion,” which is sometimes popularly used as a synonym
for “ technological change,” designates, strictly speak­
ing, a particular type of current development. It has
been variously defined, for example, as “ automatic
operations,” “ the mechanization of sensory control
and thought processes,” and “ a concern with produc­
tion processes as a system.”
While b l s studies have been concerned with
developments in autom ation, particularly in an­
ticipating long-term trends, they are not the only
technological changes taking place that affect labor re­
quirements and industrial relations. For example, new
ways of generating power, piggybacking in transporta­
tion, use of synthetic materials in manufacturing,
mechanized methods of material handling, and faster
steelmaking processes are important technological
developments, not usually covered by technical defini­
tions of automation, but having significant manpower
implications.
Impact of productivity
Since one of the principal consequences of
technological change, so far as manpower utilization is
concerned, is an increase in productivity (output per

Description of Studies
The Bureau’s research program on technological



change involves a variety of reports and studies of dif­
ferent degrees of detail and approach. The current pro­
gram thus provides: Summary reports surveying trends
in major industries; detailed industry studies; and
studies of major technological innovations, such as
computers, that affect workers in different industries.

145

employee hour), special attention is given in b l s studies
to analyzing changes in industrial productivity. Such
trend analysis is a useful method of measuring the pace
of technological change. Changes in productivity,
however, also reflect changes in capacity utilization and
many other nontechnical factors. It is important to
recognize that the productivity trend is only a partial
measure of the rate of technological change.
In determining the Impact of a specific technology,
b l s studies try to indicate the reduction in unit labor re­
quirements that the new process is designed to achieve.
In some cases, estimates of labor savings are derived on
the basis of comparisons with the estimated average
technology of the industry under study; in others, with
the best equipment that is available; or in actual plant
studies, with the technology that is actually displaced.
It is also important to distinguish between the impact
on productivity of the operation directly affected and
on productivity of the plant as a whole. An advanced
machine tool, for example, may result in a relatively
large reduction in unit labor requirements in the
machining opeations, but would have little impact on
finishing and assembling, and may even require addi­
tional labor in engineering and maintenance work. The
impact of plant productivity, therefore, would be con­
siderably less than the effect on productivity of any
department or operation directly affected.
impaet @n employment
In assessing the impact of technological change on
employment, it is necessary to consider the implications
of plant manpower policies and the effects of economic
changes, with which technological changes interact.
Analysis of the impact of technological change purely in
terms of machinery is incomplete.
At the plant level, for example, the substitution of
machinery for labor may substantially reduce job op­
portunities in operations directly affected. If efforts are
made, however, to eliminate these jobs by not filling
vacancies created by quits, deaths, and retirement of
employees, or by transfer of affected workers to other
positions in the plant or office, labor savings could be
achieved without displacing the workers affected.
Moreover, the employment impact of technological
change is also interrelated with the effects of the
business cycle. Thus, workers whose jobs are eliminated
by technological changes may not be displaced from a
plant until a decline in demand results in layoffs—a long
time after the change has been made in some cases. In
th e ' subsequent recovery, however, they may not be




146

hired back because their jobs no longer exist.
Since many changes exert their effects on employment
through the competitive market, the employment trend
for the industry as a whole must also be examined. The
plant which reduces its unit costs through technological
improvement may be able to gain a larger share of the
market and increase its employment, but at the expense
of the less technically advanced competing plants, which
may be forced to shut down, displacing workers far
from the location of the change.
Because of the whole complex of economic factors
that operate through the market, including changes in
demand, location, foreign competition, corporate
organization, and consumer taste, it is very difficult to
isolate the expanding and displacing effects of
technological change.
Impact @ ©eeupatiomis
f
Two aspects of occupational change resulting from
technological changes are examined. Changes in job
structure—the distribution of the plant or office work
force by function or broad skill grouping—are studied
to determine the extent of upgrading or downgrading.
Since the content of jobs may be altered as a result of
changes in equipment or processes, attention also is
directed to intensive before-and-after analysis of job
duties and the knowledge and abilities required to per­
form these duties as indicated by job descriptions and
observation. The content of newly created jobs also is
studied and the qualifications required and personal
characteristics of individuals selected for these new posi­
tions are described, so far as possible.
Adjustment! t@ technological change
Technological change has important implications for
personnel management and collective bargaining within
plants. The introduction of new machinery, products,
or processes often requires movement of workers
among jobs within the plant or office by transfer or pro­
motion, the setting of wage rates, and selection of per­
sons for new jobs. Often the adjustment proceeds accor­
ding to rules established in advance through collective
bargaining. Provisions to assist workers whose jobs are
eliminated include severance pay, retraining, and early
retirement. Besides analyzing the operation of formal
provisions under collective bargaining, Bureau studies
describe informal efforts to provide training, to utilize
attrition, and to obtain, jobs for displaced workers
elsewhere. The limitations of these measures as well as
their advantages are important matters studied.

Impact of eew
electronic technology
R ic h a r d W . R ic h e

The steady stream of technological progress that has
characterized our society in America has resulted in
higher productivity, elimination of many menial and
dangerous jobs, higher wages and shorter hours, and a
continuous flow of new products and services which
have resulted in a higher standard of living. New indus­
tries employing thousands of workers have been formed
to manufacture computers, electronic products, and
technologies to provide energy and control the environ­
ment.
To be sure, innovation in industries such as longshoring, agriculture, and printing, to name a few, has
eliminated jobs and required workers to acquire the un­
familiar skills associated with new technology. For
some, the adjustment has been painful. But on balance,
there is general agreement that the benefits of new tech­
nology far outweigh the disadvantages, and that innova­
tion has led to economic progress, new job oppor­
tunities, and a more prosperous society.
At this point, early in the decade of the 1980’s, there
is widespread agreement that the pace of diffusion of
technologies which incorporate advanced electronics
will be accelerated over the next few years. The experi­
ence in the United States suggests that as long as the
economy is growing, the introduction of innovations
with potential for productivity gains can be compatible
with rising employment. When computers were first in­
troduced for office data applications, for example, fre­
quently predictions were made that large numbers of
clerical and kindred workers would be displaced and
that job opportunities for millions would be curtailed.
What actually did happen was quite different. In 1960,
clerical workers in the United States numbered about
10 million and accounted for about 15 percent of total
employment. By 1980, there were more than 18 million
clerical workers and they accounted for about 19 per­
cent of the total. Thus, instead of decreasing as had
been predicted, clerical employment increased about 85
percent. And, it is projected to grow significantly to
1990.
Why did clerical employment increase instead of de­
creasing as predicted? First, normal growth in the volRichard W. Riche is an economist in the Office of Productivity and
Technology, Bureau of Labor Statistics. This report was adapted from
his presentation at the Organization for Economic Cooperation and
Development’s Second Special Session on Information Technologies,
Productivity, and Employment, held in Paris, France, Oct. 19-21,
1981.
Reprinted from the
M onthly L abor Review, March 1982.




147

ume of clerical work exceeded jobs eliminated by the
computer. Second, computers made possible work that
was previously impractical because it would have been
too costly and too time consuming. Using computers,
managers can now prepare reports and analyses that
previously were desirable but too costly.
In addition to creating employment by expanding the
scope of activities for many industries, the computer re­
quired new occupations such as systems analysts, pro­
grammers, keypunch operators, console operators, and
tape librarians. And new industries were established to
manufacture computers and related equipment, creating
a variety of occupations and employing thousands.
Technological change can cause job displacement, es­
pecially when the industry is concentrated in a particu­
lar region or locality. Sometimes the employment
impact is direct, as in the case of agriculture. In most
cases, however, the effect is less obvious. Output does
not advance at the same rate as productivity in all in­
dustries or plants, and consequently some industries
register employment declines while others register in­
creases. Regardless of the reason, displacements are
costly for both the individual and the Nation.
This report examines four major technological chang­
es under way in the United States and discusses pros­
pects for their further diffusion. The four areas are
microelectronics, industrial robots, telecommunications,
and office automation.
The development of microprocessors and microcomput­
ers in the early 1970’s, and their widespread diffusion as
we enter the 1980’s, is a major innovation in electron­
ics. Over the past three decades, the transistor that re­
placed the bulky vacuum tube was a first step in the
development of miniaturized semiconductor integrated
circuits which provide more power and reliability in a
significantly smaller package. A microprocessor unit
contains thousands of electronic components and com­
plex circuits on a silicon chip less than one centim eter
square. The unit can be combined with memory and in­
put-output capability to build a microcomputer.
The use of microelectronics has had a significant im­
pact on American consumers, workers, and manu­
facturing operations. A vast array of products— calcu­
lators, digital watches, video games, TV sets, and mi­
crowave ovens, to name a few—incorporate micro­
processors and microcomputers. But behind the scenes
in American manufacturing plants, production technol­
ogies and manufacturing methods are undergoing equal­
ly dramatic changes. Microelectronics are being
incorporated in systems which control key production
equipment, such as industrial robots and numericallycontrolled machine tools. Moreover, microelectronic de­
vices increase the processing capability of word proces­
sors, computers, data transmission and copying devices,
automatic checkout counters, and other such equipmer’

used by banks, insurance companies, and retail and
wholesale establishments.
The industrial robot is a second major technological
innovation capturing current attention. The Robot In­
stitute of America defines a robot as “a repro­
grammable multifunctional manipulator designed to
move material, parts, tools, or specialized devices
through variable programmed motions for the perfor­
mance of a variety of tasks.” According to the institute,
about 4,000 robots are in use in U.S. establishments,
with a large share in automobile manufacturing plants.
They perform tasks such as material transfer, die cast­
ing, spot welding, spray painting, and limited assembly.
Although U.S. industry is increasing its use of robots,
Japan leads the world in robot use with more than tri­
ple the number of installations in the United States.
There is little information on the impact of robots on
productivity and employment. However, evidence sug­
gests that, following installation of robots, productivity
is increased, unit labor requirements frequently are low­
ered, and quality control is improved. At one large
manufacturer of refrigerators, for example, a robot
sprays paint on refrigerator liners twice as fast as the
two-person crew that it replaced.1The future impact of
robots on productivity and employment will depend on
the extent of development and diffusion of new genera­
tions of robots that can “see and feel.”
Technological changes in telecommunications are un­
derway in all major segments- of the industry. These
innovations are boosting productivity and changing the
type of labor required in the two basic processes of tele­
phone ^communication— call switching and signal trans­
mission. The electronic computer is used extensively in
bo$Ji processes, as. well as in other operational tasks and
in ipapagerqent and accounting functions.
Jn pall switching, electronic, switching systems use
high speed computers to handle local and long-distance
calls, iA growing share of calls is handled by electronic?
sv^itcfeing systems! Total conversion is anticipated by the
year 2000. These systems cgn handle thfee to four times
m@rq palls fhan electromechanical systems.
Sharp gains in long-distance volume have led to two
innovative and important technologies in signal trans­
mission— the millimeter waveguide and fiber optic ca­
bles. Both have far greater call-handling capabilities
than the existing coaxial cables and microwave relays.
The millimeter waveguide is essentially an underground
tube through which signal-carrying waves are transmit­
ted. It is designed for use on high density communica­
tion routes. Currently, this technology is being tested;
future diffusion will depend on call volume growth.
Fiber optic cables for signal transmission are
expected to become a major transmission medium in
the 1980’s. In this technology, glass fiber cables are
combined with semiconductor light sources for very
high capacity transmission. The fiber cables are com­



148

pact, resist electrical interference, and interface well
with digital switching and transmission techniques.
Other major changes anticipated for the telecommu­
nications industry include further expansion of satellite
communication, digital transmission, computerized sys­
tems for maintenance and testing, and automation of
switching and billing tasks. Experts also foresee
nontraditional uses of the communications network for
electronic funds transfer in banking, electronic postal
service functions, and data systems for the home which
will combine communications and data processing capa­
bilities.
Office data handling and communication is a fourth
area where major technological change has occurred. A
large segment of the Nation’s work force, including
more than 18 million clerical workers, is engaged in
producing and processing data. Historically, capital in­
vestment in the office has lagged that of other opera­
tions, with investment per office worker amounting to
less than $2,000, compared with about $25,000 per fac­
tory worker.2
This “investment gap” may be closed in the years
ahead. Investment in office technologies will likely ac­
celerate during the 1980’s, as managers turn to modern
data handling technologies to . reduce labor, material,
and related expenses.. The largest. share of office costs
are deemed to be labpr-related— a strong incentive for
further mechanization.
Specific technologies to be diffused more widely in­
clude more .powerful electronic .'Computers; advanced
model w©rd processors; new equipment and techniques
to stqre, retrieve, and transmit data on microfilm; and
electronic mail networks. Increasingly, paper will be re­
placed ; by electronic images on a screen which can Jbe
transmitted by telecommunication methods.

Genef&I impact off innovations
Following are conclusions from the Bureau’ of La&or
Statistics, research on the implications of technological
change for the work force.
© While all industries are experiencing technological
change, the pace varies among and within industries.
Each industry has its own story and it is not always
in terms of computer technology and advanced auto­
mation. But even conventional changes, such as mate­
rials handling mechanization or the installation of
larger capacity equipment or machines with faster
speeds, are often major developments requiring work­
ers to obtain new skills.
© The size of investment required, the rate of capacity
utilization, and institutional arrangements are some
of the factors that act as an “economic governor” on
the speed of diffusion of technological change and, in
turn, possible employment implications.
© Industries with greater application of technological
advances generally experience larger increases in pro­

ductivity (examples, air transportation and telephone
communication); industries lagging in application of
technological advances generally experience smaller
or negative changes in productivity (examples, foot­
wear and wood household furniture).
© The content of jobs and the qualities required of
workers are being modified by technological changes.
There is less demand for manual dexterity, physical
strength for material handling, and for traditional
craftsmanship. In contrast, employers are placing
more emphasis on formal knowledge, precision, and
perceptual aptitudes. As many manual tasks are
mechanized, unskilled workers become monitors of
very expensive equipment. The reduction in repetitive
tasks that are so dissatisfying to the industrial worker
may be welcomed, but the isolation and constant
monitoring associated with advanced technology can
create new stresses.
© Higher educational achievement of workers is becom­
ing essential. The ability to read and write at a func­




149

tional level is mandatory to interpret operating
instructions of complex equipment, and to be re­
trained for the new skills demanded by changing
technology.
© Many new occupations created by new technologies
can be filled by retraining employees. Most retraining
is accomplished in-plant and includes on-the-job and
classroom instruction.
© In general, relatively few employees have been laid off
because of technological change. This is due, in part,
to the use of various techniques by the private sector
to minimize adverse effects to the worker— tech­
niques such as providing advance notice, retraining,
and reassigning displaced employees to new jobs.

1“Robots Join the Labor Force,” Business Week, June 9, 1980, pp.
62-76.
2Philip H. Dorn, “The Automated Office— The Road to Disaster?”
Datamation, Nov. 15, 1978, pp. 154-62.

Teehn@S©gy amd lalbor in
ESeetrieal and Electronic
Equipment
Robert V. Critchlow

Summary

appliance industry experienced the highest rate of
productivity growth. Productivity gains in these
industries were most rapid during the earlier portion of
the 1960-79 period.
Expenditures for plant and equipment have been in­
creasing, with capital outlays highest for electronic com­
ponents and communication equipment. Capital spend­
ing is expected to increase as those parts of the industry
experiencing rapid technological development continue
spending for new plant and equipment.
Employment rose at a relatively low annual rate of 1.6
percent between 1960 and 1980; the growth rate was
highest in the early half of the period. Employment is
projected to increase at an average annual rate of 1.7 to
2.5 percent between 1980 and 1990.

The pace of technological change has been uneven in
the diverse group of industries that make up the electrical
and electronic equipment group. The electronic compo­
nents sector, for example, is a leader in technological
innovation and has experienced strong growth in produc­
tion, employment, and capital investment. The electrical
machinery industries, however, are experiencing less rapid
technological change. Production and employment
growth also has been slower in these industries.
A number of occupations will be affected by technolog­
ical change. Improvements in assembly procedures,
primarily in the use of automatic equipment, including
robots, are changing skill requirements and increasing
productivity for several kinds of operatives. In the largest
occupation—assemblers—work is shifting from manual
assembly toward machine monitoring, loading, and
maintenance tasks. The need for welders and painters
may decline, while more mechanics and repairers may be
needed. Engineers and technicians will make greater use
of video terminals and computer techniques in designing
machines and electronic circuits—which should improve
their productivity and reduce the need for drafting em­
ployees. Solid-state controls, which are manufactured in
this industry, will also be used in appliances and other
products manufactured in this industry. They will in­
creasingly replace mechanical controls (switches, timers,
etc.) and bundles of electrical wires, which will increase
the need for scientists, engineers, and technicians, while
reducing the amount of manual assembly and soldering
work required.
Production in the electrical and electronic equipment
industry has grown steadily during the 1960-80 period.
The electronic component sector has grown most rapidly,
due to strong demand for integrated circuits and other
semiconductors. Production has also increased in each of
the other industries within this sector.
There is no BLS index of output per employee hour
(productivity) for the broad industry group, but indexes
are available for several individual industries. Produc­
tivity has increased, at varying rates, in each of the
industries for which data are available. The household

Technology in the 1080’s
Technological changes are taking place in most sectors
of the electrical and electronic equipment industry (SIC
36). To varying degrees, these changes will affect
employment levels and occupational structures. Improve­
ments in assembly operations include more automatic
equipment to assemble printed circuit boards, and
production lines with automatic stations to manufacture
household appliances and television receivers. The trend
toward more automated operations may lower unit labor
requirements somewhat and shift job skills more toward
machine monitoring and maintenance. Computer
techniques are being developed to assist engineers in
designing solid-state (semiconductor) electronic comp­
onents and integrated circuits. Solid-state controls artd
switches, and printed circuits, important products of this
industry, are replacing mechanical controls and electrical
wires in household appliances, television and radio
receivers, communication equipment, and other products
made by this industry. Designing and installing solidstate controls generally require more engineers and
technicians, and fewer assemblers, solderers, and
machine operators than the older processes. Numerically
controlled machine tools are achieving labor and other
savings in turning out communication equipment and

Reprinted from BLS Bulletin 2104 (1982),
Technology and L abor in Four Industries.




ISO

Table 3.

E^ajor technology changes in electrical and electronic equipment
Technology

Description

1-abor implications

Diffusion

Equipment to design and fabri­
cate semiconductors and re­
lated dev ices, including micro­
processors

Computers and video display terminals can be used to
design and lay out complex integrated circuits in less time
than is necessary for older methods. This is especially appli­
cable to microprocessors, which are among the most com­
plex semiconductor devices.

Designing and fabricating semi­
conductors require relatively
more scientists, technicians, and
engineers and fewer assemblers
and machine operators than
the manufacture of electron
tubes or mechanical switches and
controls which they replace.
Computer-assisted design ( C A D )
of semiconductors could reduce
the demand for drafting employ­
ees, because engineers using
C A D can do more of the design
and layout work themselves, in­
stead of delegating this work to
drafters. If automated packaging
technology becomes more wide­
ly used, it could reduce the de­
mand for workers involved with
manual packaging operations.

Computer-assisted design and
automatic packaging equipment
are in limited use due to their
high cost. Highly automated
fabrication equipment is stand­
ard.

Fabrication of semiconductors is highly automated. Pack­
aging the semiconductors after the fabrication step can be
either very labor intensive or very automated, but the latter
requires substantial capital investment.

Increased automation in assem­
bly-line operations

Computer controlled automatic sequencing and inserting
equipment for electronic components, along with a new
assembly line where operators control the line speed while
inserting components manually, are being put into use to
manufacture television receivers. In the production ol
household appliances, larger capacity presses and a range ol
automatic assembly processes are being introduced, includ­
ing limited applications of robots; and automatic equip­
ment for some welding, fastening, material handling, and
production operations.

New assembly technology gener­
ally lowers unit labor require­
ments and modifies job duties.
New technology for assembly of
appliances, for example, requires
a higher proportion of worktime
for equipment monitoring, ma­
chine feeding and loading, and
maintenance. Manual tasks are
reduced.

This TV assembly line is the first
U.S. installation of a technology
expected to be used more widely
in the United States and already
in general use in Japan. Most of
the automated assembly equip­
ment used in appliance manufac­
turing has been introduced in the
last 5 years. Robots are likely to
increase in use.

Numerically controlled machine
tools

Numerically controlled machine tools are being used to
turn out a wide range of products which are produced in
small volume. In advanced systems, cutting sequences,
machine speed, and other operations are controlled by a
computer.

Unit labor requirements in ma­
chining operations are lower in
numerical control, and skill re­
quirements for machinists are
modified. Operators monitor the
machine tool operation rather
than directly manipulate equip­
ment. with programmer and
maintenance workers skilled in
electronics required in numerical
control installations.

Nearly 5.000 numerically con­
trolled machine tools were in use
in the industry in 1978, with in­
creased diffusion anticipated.
The communication equipment
and electrical industrial app ara­
tus sectors of the industry lead in
application of numerical control.

Advanced production
equipment

Advanced equipment is being used to turn out several key
products of the industry. For example, some portions of
electric motors are now produced on automatic equipment,
resulting in increased productivity, in the manufacture of
automobile headlights, filaments are positioned more accu­
rately and testing time reduced by use of new equipment.
Flousehold appliances can be painted on automatic lines,
with the use of electrostatic painting technology. I he elec­
trostatic spray process uses either liquid or dry powder
paint. An electrocoating process, in which metal parts are
dipped into paint, is also used lor high-volume finishing
work on household appliances.

Advanced production technolo­
gy generally reduces unit labor
requirements. New electrostatic
painting lines, for example, fea­
ture modern conveyor lines and
little or no manual painting,
l a b o r requirements are lower
with dry powder paint than with
liquid paint. Labor requirements
are also low for electrocoating
operations.

Electrostatic and electrocoating
processes are in use in a number
of plants.

other industry products. Electrodeposition painting
technology is being used in household appliances,
resulting in reduced labor requirements and materials
costs. Table 3 provides a brief overview of the major
technological changes taking place in this industry, their
labor implications, and their expected diffusion.
Produetion of electronic components
Manufacturing electronic components (SIC 367) has
traditionally been labor intensive, involving assemblers,
machine operators, inspectors, and related occupations.1
This industry’s development of semiconductors since the
mid-1960’s has changed its production methods and
labor requirements considerably. Semiconductor devices
perform most functions of electron tubes (except cathode
ray tubes), but are smaller, more reliable, and generally



151

less expensive. As a result, semiconductor devices have
replaced most electron tubes. The switch from
manufacturing electron tubes to semiconductor devices
requires more engineers, scientists, and technicians; and
fewer assemblers and operators since many manufactur­
ing operations are automated. The impact of this
technology is marked. Employment in the part of the
industry manufacturing electron tubes has declined

'Electronic components (sic 367) consists of two major product
groups. One group comprises all types of electron tubes: Television
cathode ray picture tubes (sic 3672), all other radio and television tubes
(sic 3671), and transmitting, industrial, and other special-purpose
electron tubes (sic 3673). The second group contains several products;
the most important from the standpoint of technological and
occupational changesare semiconductorsand related devices(sic 3674).

are larger and stronger than the circuit dies, and contain
the electrical connections needed for electronic appli­
ances. In general, the packaging process involves bonding
individual circuit dies onto metal stampings, then attach­
ing very fine wires to make the electrical connections
between the dies and the electrodes on the stampings.
Plastic covers are then molded around the dies, sealing
them inside the now-complete packages. Finally, each
circuit is tested to insure proper operation.
Circuits are most commonly packaged manually,
which is quite labor intensive. An alternative has been to
use automated handling equipment, although high equip­
ment costs have limited this option. However, since labor
costs are rising and packaging technology is improving,
use of automated equipment may increase.

steadily, while employment in the rest of the industry—
including semiconductors—has risen sharply.
The first steps in fabricating a semiconductor device—
circuit design and mask making—are complex and re­
quire high-level technology. Many months are needed to
design a complex integrated circuit and to make photo
masks from which the circuits will be produced. Design
and layout involve determining which electronic compo­
nents (transistors, resistors, etc.) are required to make the
circuit perform as desired; then deciding how to arrange
the circuit components in the circuit base material.
Conventional methods of circuit design and layout—
drawing circuits by hand on graph paper and assembling
“bread board” circuits for testing—are slow and require
skilled scientists and technicians. Computer-assisted
design (C A D ) is faster, more accurate, and allows the
designer flexibility in circuit design and layout.
Developing a CAD system requires complex programming
to store information in the computer and to display and
position the simulated circuits on a cathode ray tube
terminal. CAD is used by only a few manufacturers, since
equipment and software costs are high and the system
must be used intensively to be cost effective. But where it
can be used, results can be dram atic: In some applications
for thick-film integrated circuits, computer-aided design
reduced costs by 300 percent.2
As CAD technology is diffused more widely, several
occupations will be affected. Computer specialists will be
needed for initial program set-up, but this could be a one­
time operation for each CAD system. Drafters might be
largely bypassed as engineers use video terminals for
design and layout. Computer control may also be
extended to photographic and mask-making operations,
similar to automatic printing plate technology used in the
graphic arts industry. This could reduce employment of
technicians who presently do the photographic work in
mask making.
Integrated circuit fabrication is a highly automated,
batch-type process that can produce hundreds of sepa­
rate, complete circuits in each production run. Silicon
cylinders several feet long by several inches in diameter
are sliced into thin wafers, loaded onto special trays, and
put through production steps as a group. During produc­
tion, many tiny circuits are fabricated, side by side, across
the surface of each wafer.
Labor costs in integrated circuit fabrication are rela­
tively low because of extensive automation. Most labor
requirements are associated with loading and unloading
the trays of wafers, and with operating fabricating and
testing equipment.
After the wafer is cut into individual circuits, the tiny
circuits are encased in protective packages. The packages

PflicroeSeietraniie technology
Microprocessors are a fairly recent development of
semiconductor technology, and are of importance to the
electronics industry both as a product sold to others and
as a technology that can be applied to the industry’s own
design and production operations. A microprocessor
contains a complete miniature processing unit on a single
silicon chip. It can be combined with otherchips contain­
ing memory, timer, and input-output functions to build a
complete microcomputer system on a single circuit board.
Designing and fabricating microprocessor chips is a very
complex undertaking, but the range of applications is
already substantial and is growing rapidly. The largest
volume of microprocessors in use are low-powered 4-bit
devices that provide relatively simple control functions.
More powerful 8-bit devices, however, account for most
of the revenue from sales related to microprocessors.
Microprocessor-based systems are expected to dra­
matically change the function and capabilities of house­
hold appliances during the 1980’s. Estimates vary on the
rate of diffusion of microelectronics in the appliance
industry through the mid- 1980’s, but one industry source
estimates that 50 percent of all major appliances will be
controlled by microelectronic devices.3
One of the most popular applications of microcomput­
ers will continue to be microwave ovens, where sophisti­
cated controls allow a wide range of cooking sequences
and temperatures (including ovens that can be pro­
grammed by the user). Microcomputers also are being
used in cooking ranges, dishwashers, clothes washers and
dryers, and other household appliances.
The growing application of microelectronics to indus­
trial and household appliance controls has brought about
changes in design and production operations. Mechanical
engineers and industrial designers work more closely with
electrical engineers to develop electronic controls as
substitutes for mechanical and electromechanical con-

2
“Hybrid-circuit Technology Keeps Rolling Along," Electronics, July
22, 1976, p. 104.

’Donald L. Owens, “ Microelectronics: A New Horizon for Appli­
ances,” Appliance, July 1979, pp. 28-31.




152

trols. Assembly operations and labor requirements
change when solid-state controls are used. It is no longer
necessary, for instance, to route and solder large numbers
of individual wires into place; thus fewer solderers are
needed. Also, flat electrical cables and flexible printed
circuits with plug-in connectors are replacing bundles of
separate wires. There may be a secondary impact in that
fewer components and less equipment are needed to
produce electronic controls. This could reduce the labor
needed for stock control, material handling, warehous­
ing, and transportation.
Improvemonts in assembly technology
Many types of assembly operations take place in the
diverse group of industries that make up the electrical and
electronic equipment industry. Technological innovations
in assembly include increased automation and improved
manual assembly lines for TV receiver producers and, in
appliance manufacturing plants, automated assembly
operations for household appliances and in-house
assembly of printed circuit boards.
A major domestic TV manufacturer has begun oper­
ating a new assembly line that has increased productivity
and product quality. This is the first such TV receiver
assembly line in operation in the United States, although
this type of assembly line has been used for several years
in Japan with considerable success. The new line features
both computer-controlled automatic sequencing and
component inserting equipment, as well as new tech­
nology for inserting parts manually, and is achieving
productivity gains. On the new line, solid-state compo­
nents come packaged in reels from vendors. Only one type
of component is included in each reel, and each reel may
contain several hundred components. A number of these
one-of-a-kind reels are mounted on a component se­
quencing machine which—under computer control—
automatically removes individual components and de­
posits them on a conveyor in the sequence required for the
automatic inserting machine. The conveyor transports
each component through an automatic testing station to
insure that it functions properly and is in the correct
sequence. Finally, the components are automatically
taped onto a new reel for use in the inserting machine,
which automatically inserts components onto the circuit
boards and then cuts and crimps the wire leads on each
component to secure them to the board. Completed
circuit boards are transported through an automated
wave soldering machine that solders all electrical connec­
tions in one operation.
Most operators on the new line originally held assem­
bly jobs involving manual insertion of components into
circuit boards or manual assembly of parts onto the
television chassis. When the automated equipment was
brought into the plant, this group was retrained to load
and operate the machines. The machine operator posi­
tions involve higher skill and pay levels.



153

Productivity is increased with the automatic sequenc­
ing and inserting equipment in that operators can insert
more components per hour than is possible with the same
number of people manually inserting components into
circuit boards. Additionally, quality is improved since
every component in the automatic line is tested before
being inserted into the boards. A major feature of the new
line is the specially designed assembly equipment for
manually inserting components that cannot be handled
on the automatic equipment.
In the manufacture of appliances, improved tech­
nology is being installed in major production tasks.
Sheet-metal components are being fabricated by larger
capacity presses fed directly from coils of sheet metal.
These components are produced at high speeds with a
minimum of manual handling.
New technology for high-volume assembly also is being
introduced. Assembly tasks are labor intensive. When
assembly lines become automated, unit labor requirements
are lowered and job skills frequently shift to machine
monitoring, machine feeding and unloading, and machine
maintenance.
One manufacturer is using automatic assembly tech­
niques to insert a retaining pin into the spout cap of a tea
kettle, in place of a manually fastened screw-and-nut
assembly. The automatic equipment has increased the
assembly rate by 84 percent, lowered unit labor require­
ments, and decreased fastener costs.
Another appliance manufacturer has installed an
automatic line to assemble washing machine cabinets that
has a maximum output of 350 cabinets an hour. Sheetmetal blanks pass through presses that shape them into
cabinets, which then move past automatic welding
stations where gussets and brackets are attached. The
cabinets next are transported by conveyor through a
manned inspection station and then to the finishing area.
The entire line is staffed by three operators and one
inspector, who checks-cabinets as they come off the line.
In a less mechanized, conventional system with the same
volume of output, labor requirements would be higher. A
solid-state control system with a cathode ray tube (CRT)
display terminal provides data on malfunctions and
defects which facilitate repairs. The automatic assembly
line has increased output by 40 percent, raised cabinet
quality, and improved operator safety.4
Robots are being used in several assembly applications
by one large manufacturer of appliances to cut costs and
improve productivity. In one application, two robots are
used to load and unload a press that trims plastic liners
used in refrigerators. In another application elsewhere on
the assembly line, two more robots spray the interiors of
the refrigerator cabinets with an adhesive that holds a
layer of foam insulation;each robot takes the place of two
■Uames Stevens. “A New Cabinet Line Pays off for Maytag.”

Appliance, February 1976, pp. 34 35.

workers and there is a 10-percent reduction in the amount
of adhesive material used.5
One manufacturer recently built a highly mechanized
assembly line for energy-saving refrigerators that speeds
production and testing procedures, and reduces labor
requirements for machine operators, welders, and mate­
rial handlers. The cabinets are produced on a semiauto­
matic line that includes an automatic destacker (which
transfers metal cabinet blanks onto the production line)
and an automatic electric. resistance welding station.
Serpentines -the metal tubes that carry freon inside
refrigerators—are made in an off-line operation. Coils of
metal tubing are fed into a machine that automatically
straightens the tubing, then cuts and bends it into the
proper dimensions for installation in refrigerators. Solidstate controls give flexibility in programming the equip­
ment to make serpentines of varied sizes.
Foam insulation is injected into the cabinets on a 6station automatic foaming installation that utilizes solidstate controls. Only one person is needed to operate this
equipment. Cabinets are brought by conveyor belt, where
a system of photo cells and magnetic tape readers routes
the cabinets—via a turntable and runout conveyors—to
the proper foaming station. The cabinets are automatical­
ly positioned on the foaming fixtures, filled with insula­
tion, and lowered onto another conveyor to leave the
foaming operation.6
Appliance manufacturers have begun to assemble their
own printed circuit boards instead of purchasing them
from electronic component suppliers. Thus, technology
and labor are “transferred” from one sector of the
electrical and electronic equipment industry to another.
The assembly of printed circuit (PC) boards involves
inserting resistors, capacitors, integrated circuit (ic)
chips, and other components onto the boards, soldering
all connections, then cleaning and inspecting the
completed boards. Assembly-methods range from largely
manual tasks—the predominant method at this time—to
semiautomatic and fully automatic processes. There are
several “aided manual” systems that allow relatively
unskilled operators to assemble complete PC boards. One
of these systems, for example, positions the PC board in a
machine which guides the operator through a sequence of
production steps by illuminating the appropriate holes in
the PC board into which each succeeding component is to
be inserted. At the same time, a tray of parts is positioned
so that only the proper component is accessible to the
operator. If the more automatic processes become more
prominent, the employment of assemblers will be
affected.

increasingly in the electrical and electronic equipment
industry. They are used most extensively in the
communication equipment and electrical industrial
apparatus sectors of the industry to turn out a wide range
of products which are produced in small volume. More
than 1,000 numerically controlled machine tools are
being used in each of these two major sectors.7 In
advanced numerical control systems, cutting sequences,
machine speed, and other operations are controlled by a
computer with significant savings in unit labor
requirements, tooling costs, and lead time. The function
of the machine operator has changed from direct manual
manipulation of equipment to monitoring the operation
of the machine tool and loading and unloading parts. A
programmer—a new position -develops the sequence of
operations, tools to be used, and feed and speeds of the
machine tool. Maintenance workers with a knowledge of
electronics are needed to service numerical control
systems. The outlook is for furtheradvances in numerical
control technology including its use for inspection of
parts.
Advanced production equipment
Other production processes where new technology is
being used include the manufacture of portions of elec­
tric motors, testing of automobile headlamps, and paint­
ing of household appliances.
Automated stator production. Automated equipment is
being used to manufacture stators for electric motors.
Automatic presses stamp out selected parts, insulation is
inserted automatically into the stator core, and coils are
wound and inserted automatically. The new equipment
has considerably increased line speed, increasing the
output without increasing the work force.
Automobile headlamp testing. Photometric equipment
that evaluates and records headlamp performance auto­
matically is helping manufacturers test headlamps to
ensure they meet specifications set by the Federal Govern­
ment. The time required to test headlamps has declined
from 20 minutes to less than 5 minutes, with sophisticated
optical equipment available to position headlamp fila­
ments to close tolerance, thus increasing worker output
and productivity.
Advanced painting technology. New technology for the
electrical depositing of paint onto household appliances is
being introduced more widely. In electrostatic painting,
paint particles are electrically charged and sprayed onto
surfaces carrying an opposite electric charge to form a
strong bond. The painted surface is then baked to a hard
finish. The electrostatic process can be used with liquid or
“wet” paint or the increasingly popular dry powder paint.

Numerically controlled machine tools
Numerically controlled machine tools are being used
5
“Robots Join the Labor Force,” Business Week, June 9, 1980, p. 68.
h
Gene Morgan, “A Sophisticated System Speeds Production, "A p p li­
ance. September 1977, pp. 50 52.



154

7
“The 12th American Machinist Inventory of Metalworking Equip­
ment 1976 78,” American Machinist. December 1978.

Labor requirements in the new electrostatic systems are
lowered since material handling and painting tasks are
largely automatic; robots are used in some installations.
At one plant manufacturing household appliances, an
operator manning a control console on a newly installed
wet painting electrostatic line can change paint colors in
60 seconds. Quality of painting is improved and main­
tenance costs are lowered.
Dry powder paint, used in the electrostatic process, is
expected to be employed more widely during the 1980’s.
Dry powder painting systems require fewer operators
than liquid or “wet” systems since paint mixing is elimi­
nated and manual touch-up is reduced. Other factors
favorable to further diffusion include easier paint han­
dling and clean-up operations. Energy requirements also
are lower. One firm which replaced a wet system with a
porcelain enamel powder system reported labor costs
were lower by 33 percent, rejects and materials were re­
duced by 50 percent, and quality was improved.8 A dis­
advantage is the inability of dry painting systems to
handle frequent color changes. In 1976, nearly ^ e le c tro ­
static thin-film powder spraying systems for appliances
were in use. Additional installations are forecast for the
1980’s.9
Another form of electrodeposition—electrocoating—
involves immersing a metal part into a tank of coating
material. The metal and the liquid in the tank carry oppo­
site electrical charges, which form the bond, providing a
continuous, evenly deposited film on the metal part. This
process was introduced for high-volume finishing opera­
tions in the mid-1960’s, and has since gained wide accept­
ance, especially for applying primer coats on appliances.
The process is less labor intensive than conventional
painting processes, gives a uniform coating even on intri­
cately shaped objects that have hidden or recessed areas,
minimizes material costs because there is almost no
wasted paint, and causes much less air and water pollu­
tion.1
0

Output and Productivity Out0©@
lk
Output
The electrical and electronic equipment industry turns
out a variety of products for government, industry, and
consumer use. This product diversity is shown in table 4,
which presents output growth in major industry sectors.
Output in the industry as a whole has increased at a
relatively high annual rate. According to the Federal
Reserve Board production index for this industry, output
grew steadily from 1960 through 1980, averaging a
8“Plant Experiences with Porcelain Enamel Powders,” Appliance,
November 1978, pp. 49, 67.
’Gene Morgan, “Focus: Powder Coating,” Appliance, September
1976, pp. 45-49.
l0Gene Morgan, “Electrocoating," Appliance, November 1978, pp.
39-41.



155

TabB@ 4. Output growth in electrical and electronic equipment,

19SO-80
SIC
36
361,2
363
365
366
367
369

Industry sector

Average annual percent change'
1960-80

Total electrical and elec­
tronic equipment2 ...........
Electrical equipment
and parts ........................
Household appliances ..
Radio and TV receiving
equipment ....................
Communication
equipment ....................
Electronic com­
ponents ..........................
Miscellaneous electrical
equipment ....................

1960-67

1967-80

5.9

10.5

4.2

4.2
4.6

8.1
8.7

2.8
2.5

3.3

13.8

.6

3.8

7.0

2.7

12.5

22.0

8.8

5.5

7.2

5.0

' Least squares trend method.
includes data for SIC 364, electric lighting and wiring equipment, not
available separately.
SOURCE: Board of Governors of the Federal Reserve System.

growth rate of 5.9 percent a year. As shown in table 4, the
rate of growth in output was substantially higher during
1960-67 than in 1967-80. There was a slight dip in output
during 1970 and 1971, after which output climbed to a
peak in 1974, dropped rather abruptly in 1975, and then
rose sharply in 1976 through 1980.
Output in the electronic components industry (electron
tubes, semiconductors, integrated circuits, etc.) has
grown more rapidly than in any other industry in the
group, increasing at an average annual rate of 12.5
percent during 1960-80, more than double the annual
growth rate for total electrical and electronic equipment
over the same period. Most of this expansion in output
has been in integrated circuits—which include micro­
processors, introduced in the mid-1970’s and gaining
widespread acceptance. Passive components (such as
capacitors, resistors, and connectors) and discrete
semiconductors have made lesser contributions.
Output of electron tubes was fairly stable during the
1970’s; the decline in production of receiving tubes (as
solid-state components became more widely used) was
offset by slight gains in TV picture and specific-purpose
tubes.
Output growth was slowest in radio and TV and
communication equipment. Consumer discretionary
income and imports affect the level and growth of output
in the radio and TV industry—which includes a number
of consumer electronic components in addition to radio
and TV receivers. In communication equipment,
telephone and telegraph products account for almost onethird of the value of industry shipments. The remaining
two-thirds consist of electronic systems and equipment
for which the U.S. Government is the major purchaser—
especially the Departments of Defense and Transporta­
tion, and the National Aeronautics and Space Adminis­
tration. Output growth, therefore, depends heavily upon

1

the demand from new households, business communica­
tion needs, and Federal Government procurement
policies.

Table 5. Output per employee hour in selected electrical and
electronic equipment industries, 1SS0-79
Output per employee hour

The expansion of output in the major sectors of the
electrical and electronic equipment industry will continue
to depend on several factors. In electronic components,
for example, the growing commercial diffusion of a
relatively new product, microprocessors (a complex
integrated circuit finding application throughout the
economy in a growing number of products from games to
industrial robots), will require further expansion of
production facilities. In electrical equipment, demand for
electric transformers and switchgear depends heavily
upon residential, commercial, and industrial construc­
tion. In major household appliances, population growth,
family starts, and housing construction influence final
demand.

SIC

Industry

Average annual percent change
1960-79

1960-67

1967-79

3621

Motors and generators

2.1

5.6

1.2

3631,32,
33,39

Major household
appliances ..................

4.4

6.1

3.9

3641

Electric lamps ...............

1.8

2.8

2.1

3645,46,
47,48

Lighting fixtures ...........

'2.6

22.9

3
2.6

3651

Radio and TV receiving
sets ...............................

3.9

6.0

3.4

'1961-78.
21961 -67.
31967-78.
SOURCE : Bureau of Labor Statistics.

Productivity
Although a productivity measure for total electrical
and electronic equipment is not published by the b l s , the
measures available for several of the individual industries
indicate that productivity change varies significantly by
industry, and that growth rates have slowed over the past
decade.1
1
The rate of increase in output per employee hour in the
industries for which BLS publishes measures ranged
during 1960-79 from an annual rate of 1.8 percent in
electric lamps to an annual rate of 4.4 percent in major
household appliances (table 5). In household appliances,
output grew more rapidly than employee hours during
1960-68; output continued to grow slowly and employee
hours declined during 1969-79. In all five of the industries
included in table 5, the rate of increase in output per
employee hour was lower from 1967 to 1979 than from
1960 to 1967. The sharpest decline in productivity was in
motors and generators.
The extent to which technology affected the movement
of productivity cannot be measured precisely. In all five of
these industries, and in others for which BLS measures are
not available, new technology has reduced unit labor
requirements in selected production operations. The
anticipated higher levels of spending for new plant and
equipment could contribute to further productivity gains
in key production tasks.

invested substantial funds for capital improvements,
including the latest production technologies discussed in
this report. In 1976, expenditures for new plant and
equipment totaled $1.7 billion (constant 1972 dollars),
more than twice the $800 million invested in I960.1
2
Capital expenditures per production worker averaged
$1,385 in 1976, well above the average of $803 in 1960.
The pace of capital spending has been uneven. Over the
longer term 1960-76 period, expenditures for plant and
equipment rose at an annual rate of 5.3 percent. Capital
spending during 1960-67 increased at a substantially
higher annual rate of 14.0 percent. In the 1967-76 period,
however, during which expenditures fluctuated markedly,
outlays (in constant dollars) declined by an average
annual rate of 0.2 percent. Between 1973 and 1976, the
decline averaged 7.8 percent a year.
Capital spending also varied significantly among the
individual industries which make up the total electrical
and electronic equipment industry. The electronic
components industry led all industries in the group with
$478 million soent for new plant and equipment in 1976.
Expenditures in communication equipment were the
second highest, $399 million in 1976. Combined, these
two industries were the source of more than one-half of
capital spending.

Capital expenditures
The electrical and electronic equipment industry has

Research and development
The electrical and electronic equipment industry is a
leader in research and development ( r & d ) spending.
According to the National Science Foundation, R&D
expenditures by the electrical and electronic equipment

1 Productivity measures are published by the b i . s for the following
1
five industries: Motors and generators (sic 3621); major household
appliances (sic 3631, 32, 33, and 39); radio and TV receiving sets (sic
3651); electric lamps (sic 3641); and lighting fixtures (sic 3645, 46, 47,
48). See Productivity Measures fo r Selected Industries, 1954-79,
Bulletin 2093 (1981).

1 Capital expenditures data are from unpublished, deflated total
2
annual investment series developed in the b l s Office of Economic
Growth and Employment Projections. See Capital Stock Estimates fo r
Input-Output Industries: Methods and Data, Bulletin 2034 (1979).
Expenditures for 1976 are the latest available.

Investment




156

industry totaled $7.6 billion in 1979, up from the $2.9
billion allocated in 1963.'1 In 1979, this industry ranked
second only to aircraft and missiles in total funds
allocated to R&D. Federal Government R&D funds
accounted for 42 percent of the $7.6 billion spent in 1979,
and company funds, 58 percent. Since 1973, company
funds for R&D in electrical machinery and communica­
tions have exceeded Federal Government R&D funds.
The electrical and electronic equipment and communi­
cation industries employed 94,700 R&D scientists and
engineers (full-time equivalent) in 1980, leading all other
major industry groups for which the National Science
Foundation provides data.

Table 6. Average annual rates o f change in employment, electri­
cal and electronic equipment, 1SSQ-8G
SIC

Industry sector

Average annual percent change’
(all employees)
1960-80

1960-67

1967-80

Total electrical
and electronic equip­
ment .................................

1.6

4.1

0.6

Electrical transmission and
distribution equipment _
_

.1

2.3

-1.6

Electrical industrial
apparatus.............................

1.7

3.0

1.1

363

Household appliances .........

.8

2.5

-.3

364

Electrical lighting and
wiring equipment ..............

2.2

6.3

.6

Radio and TV receiving
equipment ...........................

.0

6.8

-2.6

Communication equip­
ment .....................................

.7

3.0

-.5

367

Electronic components .......

3.4

7.8

2.4

369

Miscellaneous electrical
equipment ...........................

2.9

1.0

3.3

36

361
362

Employment amd Occupational Trends
365

Employment
The industry employed slightly over 2.1 million
workers in 1980 compared to 1.4 million in 1960—a 1.6percent annual growth rate (chart 9). More than one-half
of the industry work force in 1980 was engaged in
manufacturing communication equipment and electronic
components.
The trend in employment, as in other measures for this
group of industries, varied among the major industry
sectors (table 6). Employment in electronic components
increased at the greatest annual rate (3.4 percent) during
1960-80, a period of generally strong demand for these
products, particularly integrated circuits (which include
microprocessors). The average annual employment
growth rate has been slowest in electric transmission and
distribution equipment and radio and TV receiving
equipment. Employment in these industries increased
during the 1960’s, then declined during the 1970’s, so that
by 1980 the level was about the same as in 1960.
Employment growth in the electrical and electronic
equipment industry was highest during 1960-67,
compared to the more recent 1967-80 period. As
indicated in chart 9, employment increased at an annual
rate of 4.1 percent during 1960-67, compared to an
annual rate of 0.6 percent during 1967 -80. Employment
dropped sharply (by about 14 percent) between 1974 and
1975 as demand slackened. This pattern of employment
g ro w th -a higher rate during the earlier of the two
periods discussed in this report, followed by a lower
growth rate or a decline in employment during the latter
portion—was experienced in all industry sectors except
miscellaneous electrical supplies.
''These are current-dollar data; the increase in real terms is not as
great. Before 1978, the National Science Foundation published r&d
expenditure data for both electrical and electronic equipment (sic 36)
and communications (sic 48) as one combined figure. Very little R&D
work is done in sic 48. It is largely a service industry that uses equipment
developed and manufactured insic 36. Beginning in 1978, R&Ddata for
the two industries are published separately.

http://fraser.stlouisfed.org/
157
Federal Reserve Bank of St. Louis

366

' Based on least squares trend method.
SOURCE : Bureau of Labor Statistics.

The outlook is for employment in this group of
industries to increase at an average annual rate of 1.7 to
2.5 percent between 1980 and 1990, according to b l s
projections based on three versions of economic growth.1
4
Occupations
The structure of occupations is expected to undergo
change. As shown in chart 10, all the major occupational
groups except sales workers are expected to increase
between 1978 and 1990.
Operatives, the largest occupational group in the
industry, accounting for about 45 percent of total
employment in 1978, are projected to increase by more
than one-fourth between 1978 and 1990. They will
continue to be by far the largest occupational group (47
percent of total employment in 1990). Assemblers make
1 Projections for industry employment in 1990 are based on three
4
alternative versions of economic growth for the overall economy,
developed by b l s . The low-trend version is based on a view of the
economy marked by a decline in the rate of expansion of the labor force,
continued high inflation, moderate productivity gains, and modest
increases in real output and employment. In the high-trend version I,the
economy is buoyed by higher labor force growth, much lower
unemployment rates, higher production, and greater improvements in
prices and productivity. The high-trend version II is characterized by the
highGNP growth of high-trend I, but assumes the same labor force as the
low trend. Productivity gains are quite substantial in this alternative. On
chart 9, level A is the low trend, level B is high-trend I, and level C is
high-trend II. Greater detail on assumptions is available in the August
1981 issue of the M onthly Labor Review.

up more than one-third of the operatives; they are
expected to increase in number at a slightly higher rate
than the average for all occupations in the industry.
Although new technologies applicable to assembly
operations will be diffused more widely, assembly of
household appliances and other products is expected to
continue to involve a high degree of manual tasks. In
some assembly operations, however, manual tasks are
expected to decline and job skills increasingly will involve
more equipment monitoring, machine feeding and
unloading, and equipment maintenance. In contrast to
assemblers, employment of solderers is expected to
decline by 30 percent and welders and flamecutters by 8
percent between 1978 and 1990 as automated equipment
is diffused more widely. In craft occupations, employ­
ment of mechanics, repairers, and installers is expected to
increase sharply as mechanization of production
operations continues in the 1980’s.
The rate of employment change in the major
occupational categories presented in chart 10 is expected
to vary among the industry sectors. Thus it is useful to
examine Bl.S occupational projections for three industry
groups: Household appliances (SIC 363); radio and TV
and communication equipment (SIC 365,6); and a
miscellaneous group that covers SIC 361, 2, 4, 7, and 9.
Less change in the composition of occupations is
expected in the radio, TV, and communication
equipment group than in the others. Employment of
operatives, who account for more than one-third of the
employees in this industry group, is expected to increase
by about 12 percent between 1978 and 1990. However,
fewer solderers will be needed. Professional and technical
workers are expected to increase at a greater rate, while
sales workers, service workers, and clerical workers are
projected to decline.
Strong employment growth is expected in household
appliances—numerically the smallest of the three
industry groups. Sales workers is the only major
occupational group in which a decline is expected.
Professional and technical workers should increase,
although at a lower rate than the other occupational
groups.

Employment in all occupations except sales workers is
expected to grow in the miscellaneous group. Large
increases are expected for managers, clerical workers,
craft workers, operatives, and laborers. Professional and
technical workers and service workers should experience
smaller employment increases.
Adjustment of workers to technological change
Although new technology is not expected to result in
major displacement, some collective bargainingcontracts
in the electrical and electronic equipment industry
contain specific provisions concerning technological
change. One such agreement requires that the company
provide the union (the International Brotherhood of
Electrical Workers) with at least 4 weeks’advance notice
before installing numerical-control or computer-control
equipment that will displace employees. The contract also
requires that, where reasonable and practicable, the
company will retrain displaced employees in order of
seniority. The Communications Workers of America
(CWA) and the International Brotherhood of Electrical
Workers both have contracts with one large firm that
contain a clause providing early retirement, under certain
conditions, for workers displaced by technological
change. A CWA contract recently negotiated with a large
employer contains provisions for a joint labormanagement Technological Change Committee to
establish methods to avoid adverse impacts of techno­
logical change on the work force. The CWA contract also
provides protection for employees downgraded because
of technological change.
Where no specific provision relating to technological
change is included in the contract, general provisions
pertaining to seniority, retirement, training, supplemen­
tal unemployment benefits, and related topics can
facilitate adjustment of employees to the requirements of
new technology. About two-thirds of the industry’s
production workers are estimated to be unionized. The
major unions, all AFL-ClO affiliates, are the International
Brotherhood of Electrical Workers; the International
Union of Electrical, Radio and Machine Workers; and
the Communications Workers of America.

SELECTED REFERENCES
Carnes, Richard B. “Productivity and Technology in the Electric Lamp
Industry,” Monthly Labor Review, August 1978, pp. 15-19.
Hunter, Karl. “What Microelectronics Is Doing for Your Competitor,”
Appliance, May 1978, pp. 65-68.
“Hybrid-circuit Technology Keeps Rolling Along,” Electronics, July
22, 1976, pp. 91-109.
Morgan, Gene. “Electrocoating," Appliance, November 1978, pp.
39-41.

Morgan, Gene. “A Sophisticated System Speeds Production,”
Appliance, September 1977, pp. 50-52.
Oldham, William G. “The Fabrication of Microelectronic Circuits,”
Scientific American, September 1977, pp. 111-128.
Otto, Phyllis Flohr. “The Pattern of Productivity in the Lighting
Fixtures Industry,” Monthly Labor Review, September 1978,
pp. 31-37.

Owens, Donald L. “Microelectronics: A New Horizon for Appliances,”
Morgan, Gene. “Focus: Powder Coating,” Appliance, September 1976,
Appliance, July 1979, pp. 28-31.
pp. 45-49.

http://fraser.stlouisfed.org/
158
Federal Reserve Bank of St. Louis

“The Microprocessor: A Revolution for Growth,” Business Week,
Mar. 19, 1979, pp. 42B-42X.

Phillips, Donald C., and Steve Wiseman. “Trends in Microelectronics
for Appliances,” Appliance, May 1978, pp. lb-19.
“Robots Join the Labor Force,” Business Week, June 9, 1980, pp. 62-65,
68, 73, 76.
Stevens, James. “A New Cabinet Line Pays Off for Maytag ” Appliance,
February 1976, pp. 34-35.




159

York, James, and Horst Brand. “Productivity and Technology in the
Electric Motor Industry,” Monthly Labor Review, August 1978,
pp. 20-25.

Technology and Labor in
Electric and Gas Utilities
Robert V. Critchlow

S u m m a ry

is expected to continue to increase at an average
rate of 0.7 percent a year between 1977 and 1985.
Occupational requirements may change somewhat in
response to changes in the size of electric generating
plants and the type of fuel used: Nuclear plants, for
instance, will require a larger proportion of scien­
tists, engineers, technicians, and security staff com­
pared to fossil-fuel plants. The construction and
maintenance of nuclear power plants require highly
skilled welders and other craft workers. Some con­
cern exists that possible labor shortages in some
craft and technical occupations could delay con­
struction of nuclear generating plants, and, if ex­
haust gas scrubbers become mandatory on coal-fired
plants, the number of engineers, technicians, and
maintenance personnel could increase substantially.

Technological changes in the electric power and
gas industry continue to lower labor requirements in
some occupations and raise productivity. Major in­
novations underway include the more widespread
use of computers to assist generating plant control
room operators in logging data, monitoring equip­
ment, and performing calculations; an increase in the
number of nuclear power stations, which generally
require a more highly skilled work force than con­
ventional plants of similar capacity; and the return
to coal as a major fuel source. The development of
highly mechanized vehicles for power line construc­
tion and repair has changed the size and occupation­
al makeup of power line work crews. The more
widespread use of extra-high-voltage transmission
also has brought about changes in power line repair
techniques.
Capital expenditures have increased considerably
since 1960, reaching a level of $25.8 billion in 1977.
(In real terms, however, the increase is not this great
because the price of new plant and equipment has
increased.) Electric utility companies account for
most of the industry’s expenditures—about 84 per­
cent in 1977. Capital spending is expected to in­
crease fairly steadily over the next decade. Electric
utilities cancelled or postponed part of their planned
capital expenditures for 1974 and 1975 for a combi­
nation of reasons, including unfavorable economic
conditions, forecast reductions in demand, and prob­
lems with regulatory and environmental concerns,
but expenditures rose again in 1976 and 1977.
Output per all-employee hour increased at an aver­
age annual rate of 4.6 percent from 1960 to 1977,
with the most rapid increase occurring between 1960
and 1967. Due in part to technological changes, labor
requirements for operating and maintenance employ­
ees in electric generating plants have declined since
1960, and are lower per kilowatt of capacity for
large plants than for small plants. Employment grew
at the rather slow rate of 1.2 percent a year between
1960 and 1977, reaching a peak of 684,200 workers in
1974 and declining to 673,000 in 1977. Employment

Technology in the 1@ s
7©5
Major technological changes are taking place in
the electric power and gas industry which directly
affect the industry’s work force and productivity.
These include the more widespread use of electronic
computers, nuclear power generation, and coal as a
major fuel for electric generating plants. Extra-highvoltage transmission will continue to make possible
the economical transmission of large quantities of
electric power. In constructing and maintaining
transmission lines, labor requirements are being re­
duced through the more efficient utilization of skilled
workers and fleets of mechanized vehicles by com­
puterized scheduling of work assignments. The me­
chanized fleets, however, require an increase in ve­
hicle maintenance crews. Innovations such as pro­
cess control computers, being introduced in an al­
ready highly instrumented environment, will have a
less extensive impact on employment and occupa­
tions than such changes as nuclear power installa­
tions, which require substantially more scientific and
technical staff than conventional installations of sim­
ilar capacity. Research noW underway on coal lique­
faction and gasification processes may ultimately
provide a clean-burning fuel from an abundant ener­
gy source to replace oil and natural gas.

Reprint from BLS Bulletin 2005 (1979),

Technological Change and its L abor Im pact in Five Energy Industries.




160

Electronic e@mpyt©rs
Computers are used extensively in the utilities
industry. In addition to their now commonplace use
in business operations, computers are being applied
to generating plant operations, control over transmis­
sion systems, and scheduling of work assignments
for line crews.
Process control computers in generating plants
provide assistance to control room operators in start­
up operations, data logging, monitoring, and per­
formance calculations, and they are becoming stand­
ard equipment in new plants and in many older large
plants. Of the plants sampled in a recent survey,
nearly 76 percent used automatic data collection for
computerized performance calculations, and 24 per­
cent had computers with control-function capacity.1
Fuel savings, increased safety and reliability, re­
duced chance of operating errors leading to equip­
ment damage, and improvements in equipment utili­
zation are claimed. Many large plants have opera­
tions that are so complex that a substantial amount
of automatic control is required for safety and relia­
bility.
Process control computers are commonly applied
to economic dispatch and automatic load control—
operations principally concerned with dispatching
power over transmission lines and the coordination
of power generation and interchange. These opera­
tions have become so complex that dispatching per­
sonnel have difficulty assimilating the vast amount of
data available. The solution has been the develop­
ment of automatic control systems typically con­
sisting of digital computers, local and remote cath­
ode ray tube (CRT) terminals, animated diagram
boards, and a network of telemetering devices.
These systems provide dispatchers with the informa­
tion and control necessary to supply power economi­
cally at proper voltage and frequency throughout the
power system. The’ optimization of power produc­
tion, continuous control of generating units, and
improved reliability and accuracy of the system
provide direct economic benefits. Indirect benefits
include the improved coordination of loads between
interconnected utilities.
There are some applications of process control
computers to full closed-loop control of generating
plants—although this is generally limited to hydroe­
lectric stations. Sn one such application, a 4-unit 285megawatt (Mw) hydroelectric plant can be operated
automatically, either locally or by remote control
from a central dispatching center. In another appli­
cation, a 4-unit 225-Mw hydro plant is controlled
from a location 8 miles away; the only personnel at
the plant are security guards. The extent to which

closed-loop remote control of generating plants is
used is not known, but, where used, it allows some
reduction in operating personnel.
Computers can be applied to a number of other
operations, such as plant design, long- and short-term
planning, fossil-fuel scheduling, and nuclear core
analysis. The range of computer applications will
probably grow in the future as computer hardware
and software technology continues to develop.
Many of the computer applications require the use
of sophisticated mathematical models and techniques
—which, in:turn, require programmers, systems ana­
lysts, peripheral-equipment operators, and others in
computer-related occupations. The demand for peo­
ple with computer-related job skills should increase
along with the increasing range of computer apptica<-tions. Also, utility engineers must have training in
computer techniques to use computers for transmis­
sion and distribution (T&D) systems planning and
for studies of T&D operations.
Computers ,are also being used more widely to
schedule line trew s with highly mechanized vehicles
to reduce time and cost in constructing and main­
taining transmission and distribution lines.
Mydtear p©w©r
Nuclear generation of electric power has become
increasingly important over the past several years as
costs of commercial power generation have risen
and as concern has mounted over the future availa­
bility of petroleum. Problems associated with air pol­
lution caused by conventional power plants also
have been a factor. By the end of 1977, 49 licensed
nuclear plants were in operation, with 49,881 Mw, or
9.0 percent of total generating capacity.-2 The Feder­
al Energy Regulatory Commission has estimated
that, by 1985, nuclear power plants may account for
18.6 percent of total generating capacity.-1
The increase in the number of completed nuclear
power plants over the past several years has been
less than anticipated. Inflation, combined with tight
money markets and uncertainty as to future demand
growth, has caused postponements and cancellations
in the construction of a number of nuclear plants.
Opposition to nuclear power plants based on con­
cern over safety and environmental factors, nuclear
fuel reprocessing, and waste disposal also has
caused delays and cancellations. In addition, the
lead time for bringing a nuclear plant on line has in­
creased as a result of the growing complexity and
size of the plants themselves, changing Federal regu­
lations concerning construction and operation proce­
dures, and problems in finding suitable sites. In late
2 M onthly Power Plant Reports, FPC Form 4, U.S. Department
of Energy, 1977.
2 Department of Energy estimates.

1Gordon D. Friedlander, “ 20th Steam Station Cost Survey,”
Electrical World , Nov. 15, 1977, p. 51.



162

Y afete &

K3<a|©ir S@€ta@S®g y eham g® © in ©l©cfiric a n d g a s u t ilit ie s
T echnology

D escription

L abor im plications

Diffusion

Electronic com puters

Process control com puters in
generating plants are used for
d ata logging, m onitoring, and per­
form ance calculations, providing
fuel savings, increased safety and
reliability, and im provem ents in
equipm ent and labor utilization.
Process control com puters are
com m only used in dispatching
pow er over transm ission lines and
coordinating generating and inter­
change
operations.
C om puter
scheduling of labor and vehicles
has reduced tim e and costs in
m aintenance
and
construction
operations.

N u clear pow er generation

Light-w ater reacto rs
currently
dom inate the industry and one
high-tem perature gas reacto r is in
use. Some developm ent w ork has
been done on breed er reactors.
Efforts are underw ay to standard­
ize nuclear pow er plant design in
o rd er to facilitate the increasingly
com plex licensing procedures.

G reater dem and for scientific and
technical specialists and security
personnel than conventional pow ­
e r plants. H igher skill require­
m ents for control room operators
and construction and m aintenance
crew s.

B y the end of 1977, 49 licensed
nuclear plant w ere in operation,
providing about 9 percent o f total
generating capacity.

E x h au st gas scru b b ers for solid
coalbum ing plants

Exhaust gas scrubbers rem ove
sulfur dioxide by forcing exhaust
gases through a w ater and lime­
stone slurry o r som e o th er chem i­
cal process prior to venting the
gases
into
the
atm osphere.
Scrubbers still have a num ber of
problem s that must be solved
before they can be considered
com pletely successful.

Increased labor requirem ents for
constru ctio n ,
operating,
and
m aintenance activities.

M ore than 24 scrubbers w ere in­
stalled or under construction in
1974, according to Federal Pow er
C om m ission data. The num ber of
installations is expected to in­
crease.

Extra-high-voltage (E H V ) tran s­
mission o f electric pow er

EH V technology has m ade possi­
ble the econom ical transm ission
o f large blocks of pow er, facilitat­
ing the developm ent of regional
pow er pools.

Some increase in difficulty of
w ork due to use of higher tow ers
and need to use heavier equip­
m ent on higher voltage lines. Use
of “ bareh an d ” m aintenance tech­
niques speeds repairs but requires
special training.

EH V technology now dom inates
the transm ission of electric pow ­
er.

Productivity has been increased in
the construction and m aintenance
o f transm ission and distribution
lines by the com bination of small,
highly trained line crew s w ith a
large num ber o f especially devel­
oped work vehicles.

Line crew s now handle a greater
am ount of w ork than was pre­
viously possible; consequently the
num ber of people in this occupa­
tion has not grown as rapidly as
the size of the transm ission and
distribution netw ork. D em and has
increased for vehicle m aintenance
personnel.

Presently in wide use.

M echanized vehicles fo r con­
struction and m aintenance of
pow er lines




162

R educes the tim e control room
o perators and system load dis­
patchers spend reading instru­
m ents, logging d ata, and perform ­
ing calculations. Load dispatchers
would have difficulty assim ilating
the am ount of available d ata
w ithout
com puter
assistance.
Increased dem and for people in
com puter-related
occupations.
Som e utility engineers required to
learn com puter techniques.

Seventy-six percent of generating
plants use autom atic d a ta collec­
tion for com puterized perform ­
ance calculations, and 24 percent
have com puters with control
function capacity.

1972, lead time was about 7 years;45 by mid—1977,
lead time had increased to roughly 10-12 years.5
Most nuclear plants are virtually custom built, which
is time consuming and expensive. Standardized plant
designs (perhaps based on previously approved de­
signs) that can be mass produced and approved as a
group could shorten lead times by several years. The
Nuclear Regulatory Commission is encouraging such
an approach, and standardized plants are beginning
to be constructed. To hasten the process of approv­
ing sites for nuclear power plants, the Federal Gov­
ernment is proposing that States create, in effect,
“ site banks” by approving areas for nuclear plant
construction in advance of any licensing requests by
utility companies.6
There are several types of nuclear reactors in
commercial operation or under development. Lightwater reactors (LWR’s) presently dominate the nu­
clear power industry. These reactors use enriched
uranium-235 for fuel, which is somewhat limited in
supply, and they utilize heat energy from the reactor
core to convert water into the steam that drives the
turbine-generator units. Light-water reactors with
over 1,000-Mw capacities are now in operation.
High temperature gas reactor (HTGR) technology
is well developed in Europe. Only one gas-cooled
reactor, of 330 Mw, is operating in the United
States. Gas-cooled reactors offer greater thermal
efficiency than light-water reactors (39-percent effi­
ciency for HTGR’s, compared to the 33- to 34-percent efficiency of LWR’s), reduce the effect of ther­
mal pollution, and use thorium as well as enriched
uranium for fuel. For gas-cooled reactors to be
commercially successful, their total generating costs
must be Competitive with those of light-water reac­
tors and coal-fired plants, and conclusive cost data
are not yet available.
A third type of reactor—the breeder reactor—is in
the development stage. The breeder reactor converts
uranium-238 or thorium-232 to fissionable plutonium239 or uranium-233 at a faster rate than it consumes
fuel, in effect creating more fuel than it uses. Most
of the development work has been concentrated on
the liquid-metal fast-breeder reactor, as this type of
reactor has the fastest conversion rate. The future of
breeder reactor technology is uncertain, however,
since development work is expensive and technically
difficult and requires extensive use of plutonium.
Labor requirements in nuclear plants differ from
those in fossil-fuel plants of similar capacity. Nucle­
ar plants tend to have more highly trained staffs, in­
4 “ Nuclear Survey: Lead Times Stabilizing,” Electrical World,
Oct. 15, 1972, p. 7.
5 “ Carter Seeking Speed-Up of Nuclear Plant Licensing,” The
W ashington Post, Aug. 4, 1977, p. A4.
6 Ibid.



163

eluding a larger number of scientists, engineers, and
technicians. More security personnel are required at
nuclear plants—a service which used to be contract­
ed out to private guard and detective agencies but is
now being handled to a larger extent by the utility
firms themselves. Nuclear plant operators must be
trained to work with fissionable material and must be
licensed by the Federal Government. Construction
and maintenance work in nuclear plants is done to
very exacting specifications and requires craft work­
ers with very high levels of skill. Maintenance crews
may be slightly larger at nuclear plants because
maintenance procedures are more complex. Regula­
tions concerning radiation exposure sometimes ne­
cessitate the use of protective clothing, which might
hamper working ability and decrease efficiency to
some extent.

Coal for fuel
Coal is the most abundant energy source in the
United States and was the primary fuel for steam
generating plants before 1965. Between 1965 and
1972, many utility firms switched from coal to oil.
Initially, this switch occurred because oil was less
expensive, but during the latter part of this period
pollution control also became an important consider­
ation. Much of the coal available in the United
States has a high sulfur content and is a major
source of air pollution from generating plants. Oil is
a cleaner burning fuel. By 1974, the problems inher­
ent in heavy reliance upon oil became clear; limited
domestic supplies and dependence upon foreign
sources. Coal, therefore, has become important
again to electric utilities.
The sulfur dioxide emissions that result from
burning solid coal remain a major air pollution prob­
lem. There are several possible solutions. There is
low-sulfur coal available, primarily in the western
part of the United States. This coal generally has a
lower Btu (British thermal unit) content than highsulfur coal, requiring a greater quantity to be burned
for the same energy input. The differences in quanti­
ty used would have an impact on generating plant
storage capacity and on fuel and ash handling. There
are also transportation expenses involved when us­
ing western low-sulfur coal in the eastern part of the
country.
A somewhat controversial solution is the installa­
tion in generating plants of exhaust gas scrubbers,
which are cleaning devices that remove much of the
sulfur dioxide from exhaust gases. Scrubber technol­
ogy is still developing and needs further refinement
to be fully effective. Scrubbers are expensive—they
can add up to 50 percent of the cost of a boiler-gen­
erator system. They also consume from 1.5 to 5 per­

cent of the plant’s output. 7 Reliability has also been
a problem. The solution so far has been to build in
redundant equipment or to step up maintenance op­
erations—both of which are expensive procedures.
In one of the earliest scrubber installations, the plant
maintenance force had to be increased by 50 percent
to handle equipment breakdowns and corrosion
problems.K Many scrubbers produce large amounts
of watery sludge as a waste product. The disposal of
this sludge is a major unresolved problem.
The number of scrubber installations will probably
increase because, in spite of the problems and ex­
penses involved, scrubbers do provide control over
some of the pollutants caused by generating plants.
Scrubbers are complex equipment, and, as the num­
ber of installations increases nationwide, the number
of maintenance workers needed in the industry also
will rise.
An alternative to the direct burning of coal is the
conversion of coal to a gas or a liquid. For electric
utilities, advantages include the capability to remove
sulfur and ash during the conversion, thereby reduc­
ing air pollution when the converted coal is burned.
The coal converted by at least some of the several
gasification and liquefaction procedures that have
been proposed can be transported by pipeline. At
present, however, coal gasification and liquefaction
on a large scale are not commercially available, and
the cost and reliability of the processes have yet to
be proven. Given the present technology, these al­
ternatives are more expensive than the installation of
exhaust gas scrubbers in generating plants.9
Some generating plants that were designed to burn
oil or natural gas can also burn coal in liquid or gas­
eous form. Converting these plants to burn solid
coal, however, would be prohibitively expensive
and, in some cases, where insufficient land is availa­
ble for coal storage and for coal and ash handling
equipment, technically impractical.
Labor requirements in coal-fired plants tend to be
higher than those in oil- or gas-fired plants. Using
coal requires moving it from storage areas near the
plant to furnaces in the plant and cleaning out the
ash residue after the coal is burned. This work is
performed by “ fuel and ash handlers,” a semiskilled
occupation. The future use of coal in liquid or gas­
eous form, if ultimately proven commercially attrac­
tive for U.S. utilities, would eliminate the need for
this occupation (as has occurred in gas-fired plants)
and reduce total utility industry labor requirements.

Research on conversion of coal into synthetic gas or
to liquid form has been intensified because of vast
coal resources available within the United States and
concern over future availability of oil and natural
gas.
High-voltage transmission
Extra-high-voltage (EHV) technology now domi­
nates the transmission of electric power. Develop­
ments that have facilitated the growth of EHV trans­
mission include the introduction of bundles of two
or more conductors, insulator strings set in “ V”
configurations to control swing, the use in some in­
stances of guyed structures in place of self-supporting
towers, the use of aluminum and special steels in
line structures for reduced maintenance require­
ments, and the use of helicopters to facilitate con­
struction. As of August 1977, there were almost
117,000 miles of EHV transmission lines in serv­
ice. 10 The development of EHV technology makes
possible the economical transmission of large
amounts of electric power over long distances, with
significant reductions in right-of-way requirements
and corresponding reductions in right-of-way mainte­
nance operations compared to what would have been
required using lower voltage lines. EHV inter­
connections presently cover most of the country.
The higher voltages involved in EHV transmission
have caused some changes in work techniques. Line
crews work on higher towers using longer, heavier
“ hot sticks” and the more modern “ barehand” tech­
nique. “ Barehanding” is a process in which the
worker handling an energized circuit becomes a part
of the circuit, with precautions against grounding
(such as working in an insulated fiberglass bucket or
on a fiberglass ladder suspended from the line tow­
er). Under the proper circumstances, barehand re­
pairs can be completed in a fraction of the time re­
quired by more traditional methods.
Power Bine construction and maintenance
Construction and maintenance techniques continue
to improve, with crew size and productivity un­
dergoing change. The use of helicopters in rough ter­
rain, chemicals to control brush on rights-of-way,
and lighter metals in structures are among changes
that have reduced construction time and mainte­
nance requirements for line crews.
The vehicles used in constructing and maintaining
transmission and distribution (T&D) lines have un-

1 Department of Energy, Federal Energy Regulatory Commis­
0
sion.
1 “ Mechanization Revolutionizes Construction,” Electrical
1
World, June I. 1974, p. 164.
1 Ibid.
2

7 Paul H. Weaver, “ Behind (he Great Scrubber Fracas,” For­
tune , Feb. 1975, p. 112.

8 Ibid.
9 Lawrence H. Weiss, “ Clean Fuel and Scrubbing Compared,”
Electrical W orld , Oct. I, 1976, pp. 70-73.



164

dergone considerable technological change over the
past 10-15 years—a change that has had quite an
impact on T&D workers. These vehicles (mostly
truck chassis weighing 22,(X ) to 40,(X ) lbs.) carry
X
X
hydraulically operated equipment, such as- 360-de­
gree rotating derricks and pole hole diggers, or aerial
lifts with booms that can range from 20 to 150 feet
high, or plows, backhoes, earth augers, cable pull­
ers, etc. This mechanization of mobile equipment
was well underway by the early 1960’s and has con­
tinued to grow rapidly, as illustrated by the in­
creased use of aerial lifts: The average utility used
10 aerial lifts in 1962 and 97 lifts in 1974. M
Vehicle mechanization grew so rapidly because
utilities needed to keep up with increasing construc­
tion demands with minimum increases in cost and in
the size of construction work crews. Additionally,
the cost of labor was increasing more rapidly than
the cost of construction equipment. In the mid1960’s, for example, the price of a 1/2-ton pickup
truck was equal to a top line crew worker’s pay for
455 hours of work. In 1974, the cost of a new pickup
truck was equivalent to a top line crew worker’s pay
for only 325 hours.12
Mechanized mobile equipment has made possible
a reduction in the size of construction work crews
and T&D line crews. Large, all-purpose line trucks
are used where work is concentrated in one area—
but generally with crews of 6 people rather than the
traditional 8- to 9-person line crews. A fleet of small­
er special-purpose vehicles with 2 or 3 crew mem­
bers each, equipped with 2-way radios and backed
by computerized scheduling of work assignments,
can generally provide the greatest productivity for
work scattered over large areas. A modern transmis­
sion line crew might typically consist of 4 aerial lifts,
an earth auger, and a digger/derrick truck with 2
crew members each, and a pickup truck for the su­
pervisor—7 specialized vehicles and a crew of 13
highly skilled workers.
The growing mobile fleet requires an increasing
commitment of resources—labor, equipment, and
managerial skill—for maintenance and repair opera­
tions. Over 90 percent of the utilities that own their
vehicle fleets operate service and repair facilities (al­
though some major repair work may be contracted
out). 12 Scheduled maintenance programs are neces­
sary to maximize vehicle availability and minimize
fleet operating costs. Managerial ability, sometimes
combined with computerized scheduling and record­
keeping, is necessary to operate such programs.
Maintenance personnel need to be familiar with both
automotive and hydraulic servicing and repair.1
*
3

Investment
Capital expenditures
Expenditures for new plant and equipment in the
major industry group electric, gas, and sanitary serv­
ices (SIC 49)<4 rose from $5.2 billion in 1960 to $25.8
billion in 1977, an average annual increase of 11.6
percent. (In real terms, however, the increase is not
as great since the price of plant and equipment has
risen over this period.) Most of the growth occurred
after 1967, with expenditures increasing at an aver­
age rate of 10.8 percent a year between 1967 and
1977. The average rate of growth between 1960 and
1967 was 7.8 percent a year.
Capital expenditures per nonsupervisory worker in
the industry have grown almost fivefold over the
past 17 years, from $10,143 per worker in 1960 to
$46,638 per worker in 1977—an average increase of
10.8 percent a year. The average annual growth rate
was 7.9 percent during 1960— and 10.0 percent
67
during 1967-77.
Electric utilities account for the largest portion of
the industry’s capital expenditures, with 69.1 percent
of 1960 expenditures and 83.7 percent of 1977 ex­
penditures. Electric utilities spent $3.6 .billion in
1960 and $21.6 billion in 1977—an increase averaging
13.2 percent a year. The average annual growth in
spending during 1960 — was 8.9 percent; the rate
67
during 1967-77 was 12.1 percent.
The industry went through a period of economic
uncertainty during the mid-1970’s which had an Im­
pact upon its capital spending activities. This is a
highly capital-intensive industry which for more than
15 years had a steady, predictable growth in demand
averaging 7.4 percent a year15 —a situation that al­
lowed an orderly growth in capital expenditures.
However, in 1974 and to a lesser extent in 1975,
construction and fuel costs rose rapidly while the
growth in demand was well below the historical rate.
High interest rates and low stock market prices lim­
ited the ability of utilities to raise funds in the mo­
ney market. Problems with regulatory and environ­
mental concerns continued. In response to this situa­
tion, electric utility firms cancelled or postponed a
considerable part of their planned capital expendi­
tures. According to Business Week, 170,000 Mw or

1 Data are available from the Department of Commerce only
4
for this broader SIC 49 industry grouping, which, in addition to
including establishments which generate, transmit and/or distrib­
ute electricity, gas, or steam (SIC 491, 492, and 493), also in­
cludes establishments which distribute water, provide sanitary
services, supply steam, and operate water supply systems for irri­
gation.

1 Carol J. Loomis, “ For the Utilities It’s a Fight for Surviv­
5
1 Michael G. McGraw, “ Fleet Management Becomes More
3
al,” Fortune, Mar. 1975, p. 97.
Sophisticated,” Electrical World, Aug. 1, 1975, p. 38.

http://fraser.stlouisfed.org/
165
Federal Reserve Bank of St. Louis

47.2 percent of a planned 360,090-Mw generating
capacity were cancelled or significantly delayed in
1974.161 Electrical World noted that in 1975 capital
8
7
spending declined for the first time in the industry’s
history. 17
Expenditures turned upward again in 1976 and
1977. This resumption of capita! spending reflected
the general improvement in economic conditions af­
ter 1975, the inability of utility companies to further
postpone to a significant degree generating plant
construction in the face of growing demand, and
concern over power shortages and service reliability.
The outlook over the next several years is for a
continued increase in expenditures. McGraw-Hill’s
1977 annual survey of business plans for capital
spending^ indicated that the electric utility industry
planned to spend $25.2 billion for new plant and
equipment in 1978, $27.5 billion in 1979, and $29.2
billion in 1980. Approximately 87-88 percent of
these funds were to be for machinery and equip­
ment; the balance was for buildings and vehicles.
A slower rate of growth in demand could ease the
pressure on generating capacity. Demand (kilowatthour sales) actually dropped slightly in 1974—a
short-run response to conservation efforts, the eco­
nomic downturn, and unexpectedly large increases in
the price of all energy sources, including electric
power. After a period of adjustment to higher ener­
gy costs, demand began to grow again, but at less
than the historical rate of 7.4 percent a year. The
Federal Energy Regulatory Commission’s Bureau of
Power considers a growth rate of 5.7 percent a year
to be likely between 1977 and 1986.
Funds for research and development
There are several sources of research itnd devel­
opment funds in the electric power industry: Equip­
ment manufacturers, the Federal Government, and
the utility companies themselves. Equipment manu­
facturers perform much of the basic research and
development (R&D) work applicable to the electric
power industry, recouping their costs by selling the
equipment they develop to the power companies.
Federal R&D funds have been largely concentrated
in the development of nuclear power.
According to the Federal Energy Regulatory
Commission, annual R&D expenditures for class A

1 “ Utilities: Weak Point in the Energy Future,”
6

and class B electric utilities19 were in the range of
$37 million to $47 million between 1966 and 1970,
rising to $239 million in 1973. Expenditures declined
slightly to $234 million in 1974, but rose again to
$290 million in 1976. Only about 20 percent of theSe
funds were spent directly by utility companies. The
majority of the funds went to organizations such as
the Edison Electric Institute, the Electric Power
Research Institute, and the Battelle Memorial Insti­
tute.

Production and Productivity Outlook
Output
Output per employee hour increased at an average an­
nual rate of 4.6 percent from 1960 to 1977. The growth
rate was higher during the 1960-67 period (6.3 percent
per year) than between 1967 and 1977 (3.0 percent per
year).
Output has grown steadily for many years. In
1974, however, demand for electricity declined in
response to price increases, economic conditions,
and conservation efforts. This contributed signifi­
cantly to the first drop in this industry’s output since
at least 1947. In 1975, output returned roughly to the
1973 level, and increased again in 1976 and 1977.
Output will probably continue to increase through
the coming decade for the industry as a whole.
Demand for electricity, as discussed earlier, is ex­
pected to increase steadily. For gas utilities,, howev­
er, the outlook is not so positive. The supply of
domestic natural gas is declining and synthetic gas is
not expected to be available in significant quantity
until the late 1980’s. Use of imported natural gas can
be increased to some extent. The net result is a pro­
jected slight decline in the gas supply through
1985.20
Productivity
Output in electric power and gas (BLS weighted in­
dex) increased at an averaged annual rate of 5.9 percent
between 1960 and 1977. During the 1960-67 period,
growth in output averaged 6.9 percent a year, while the
1967-77 period experienced a lower average annual
growth rate of 4.2 percent.
There may be a long-term decline occurring in the
rate of productivity growth. Although output contin­
ues to rise at a faster rate than employee hours, the
rate at which output is growing peaked in 1970 and
declined through 1977, while the rate of change for

Business

Week, Jan. 20, 1975, p. 46. J

1 “ 27th Annual Electrical Industry Forecast,” Electrical
7
World, Sept. 15, 1976, p. 58.
18 Business Plans for N ew Plants and Equipment, 1977— 30th
80,

19Class A and class B electric utilities have accounted for
i roughly 80 percent of total kilowatt-hour sales over the past de­

cade.
Annual McGraw-Hill Survey (New York, McGraw-Hill Publica­
20 United States Energy Through the Year 2000 (R evised) (U.S.
tions Co., Economics Department) May 6, 1977.
Department of the Interior, Bureau of Mines, Dec. 1^75), p. 65.

http://fraser.stlouisfed.org/
166
Federal Reserve Bank of St. Louis

in 1960 and 1967 but then grew between 1967 and
1977 at an average annual rate of 0.8 percent.

employee hours continued to grow steadily through
1974 and was only slightly lower in 1975, 1976, and
1977. Hence, output p e r‘employee hour is growing,
but the average annua! rate of growth has been grad­
ually declining since reaching a peak in 1964. The
productivity growth rate for nonsupervisory workers
has been higher, and the increase in employment has
been lower, than for all employees.
Electrical World publishes a continuing survey of
generating costs for electric utility steam plants that
includes data bn' the humber of operating and
maintenance employees' per Mw of net output. In
1960, 0.306 employees were required per Mw of
net output.21 By 1976, hoWever, labor requirements
had declined by 60 percent to 0.122 employees per
M w .2 2 The survey indicates that labor requirements
tend to? be lower for larger generating plants. The
survey also indicates that labor requirements vary by
type of generating plant. Nuclear plants have the
greatest labor requirements per Mw, needing more
people in all occupations (except fuel and ash han­
dlers) than the other types of generating plants.
Coal-fired plants have the- second highest level of
labor requirements* oil-fired plants the next, and gasfired plants the lowest.
‘ ~
The- size of generating units is not likely to in­
crease as rapidly in the future as over the past 20
years, and nuclear and coal-fired plants are expected
to be the main sources of electric power in the fu­
ture. Labor requirements per Mw, therefore, may
not continue to decline, as much as they have over
the past decade.

Oscypati@ns
Technological and other factors are altering to
some extent the occupational structure in the electric
power and gas industry. One area of change is in the
balance of supervisory and nonsupervisory workers:
Nonsupervisory workers have declined from 89 per­
cent of total employment in 1960 to 83 percent in
1976.
•;
A comparison was made of labor costs for various
ipccupations between a group of large generating
plants (averaging 2,626 Mw) and a .group of smaller
plants (340 Mw) in 1975.23 Labor costs5 per net Mw
for the smaller plants were approximately 35 percent
higher for supervisors, 315 percent higher for operat­
ing personnel, 48 percent higher for maintenance
personnel, 200 percent higher for fuel and ash han­
dlers, and 188 percent higher for clerks.
The types of fuel used by generating plants also
affect occupational requirements. Fuel and .ash
handlers are not required for plants using natural gas
but are needed in plants that burn coal and, to some
extent, in plants that burn oil. Also, nuclear plants '
require more specialists than any type of fossil-fuel
plant. As nuclear plants and coal-fired plants are
expected to become the dominant types of power
plants over the next decade, the occupations of spe­
cialist and fuel and ash handler should become
more important.
;
H j

Employment and Occupational Trends
Employment
Employment in electric power and gas, according
to BLS data (SIC 491, 492, 493), increased rather
slowly from 582,300 in I960 to a peak of 684,200 in
1974 and then declined to 673,000 in 1977. The aver­
age annual growth fate over' the 1960 — period was
77
1.2 percent, with most of the growth occurring after
1967. The average annual rates of change for 1960—
67 and 1967— were 0.4 percent and 1.3 percent,
77
respectively. BLS projections to 1985 indicate that
growth in employment may average 0.7 percent a
year between 1977 and 1985.
Employment growth for nonsupervisory employ­
ees has been slower than for all employees; nonsu­
pervisory workers increased at an average rate of
0.8 percent a year between I960 and 1977. The num­
ber of nonsupervisory workers was about the same
2! Leonard fyj. Olmsted, “ 14thj Steam Station Cost Survey,”
E lectrical W orld , Oct. !8, 1965, p. 104.

22‘Fried|lander, “ 20th Steam Station Cost Survey,” E lectrical
W orld , Nov. 15, 1977, p. 4 4.’




167

Employment is projected to increase in six of the •
eight major occupational groups, with the largest ■
increases expected to occur among professional apd
technical workers, managers and administrators, and
craft workers. Specific occupations in which increases
are expected include electrical engineers, electronic
technicians, computer specialists, computer peripheral
equipment operators, construction electricians,
plumbers and pipefitters, boiler-truckdrivers. Some of
the occupations for which declining employment is pro­
jected are keypunch operators, furnace tenders and
stokers, cleaning service workers, and construction
laborers (except carpenters).

23 Results from the survey of steam generating plants by E lec­
trical W orld indicate that the cost of operating and maintenance
employees per Mw of net output declined steadily from I960 to
1970, then rose somewhat in 1972, and declined again in 1974 (al­
though not returning to the 1970 level). In this study, 1960 data
are from the 14th Steam Station Cost Survey, E lectrical W orld.
Data for 1962-72 are from Leonard M. Olmsted, “ 19th Steam
Station Cost Survey,” E lectrica l W orld , Nov. 15, 1975, p. 44. The
20th Cost Survey, in 1977, did not have such detailed information
for labor cost

Some decline in the number of power plant opera­
tors is anticipated. Larger and more efficient equip­
ment is expected to create increases in output with
little or no increase in labor requirements. The same
number of people, for instance, can operate a large
generator or a small orid.
The occupational structure at a fossil-fuel generat­
ing plant visited by BLS staff tends to support this
projection. This plant utilizes three generating units:
Two small units (175 Mw each) operated from one
centralized control room and one large unit (850
Mw) that has its own control room. Both control
rooms are run by fouL-person operating crews, al­
though the skill requirements are higher for the larger
generating unit.
However, a nuclear generating plant also visited
by BLS staff has somewhat different occupational
requirements. This plant uses larger and more highly
skilled control room operating crews—seven to eight
people, including a minimum of five operators li­
censed to work with fissionable fuel. Additionally,
there is an ongoing training/retraining program at the
plant to which operators are assigned on a rotating
basis.. If this plant is representative of nuclear plants
in general, then an increase in the number of nuclear
plants could reduce the projected decline in the
number of power plant operators.
There has been some concern in the electric pow­
er industry about possible shortages of skilled con­
struction and operating personnel during the coming
decade. Such shortages would have greater impact
upon nuclear generating plants because of the many
special skills involved. Among the occupations criti­
cal for constructing and operating nuclear plants,
where shortages are possible, are nuclear, mechani­
cal, and electrical engineers, reactor operators,
health physics/radiation monitor technicians, mill­
wrights, and nuclear-qualified welders (most of
whom come from the ranks of steam/pipe fitters and
boilermakers). 24
A new labor demand model that forecasts power
plant construction employment has been developed
by the Departments of Labor and Energy and the
Tennessee Valley Authority. 25 The model covers

1978— and breaks employment estimates down by
81
region, occupation,. and type of generating plant.
This model could be a useful tool for utility compa­
nies in estimating their employment needs.
Some increase is expected in occupations Con­
cerned with the transmission and distribution of
electric power. The number of line and cable work­
ers should increase. Increased use of automatic
equipment in substations—allowing more remote
control operations—may cause a decline in regular
substation operators but an increase in the more
highly skilled mobile substation operators, who trav­
el from one remote-controlled substation to another.

Adjustment of workers to technological change
Training programs are being, established to facili­
tate adjustment of employees to. thei^equirement^ of
new technology. Control room operators in nuclear
generating plants, for example, are licensed by the
Nuclear Regulatory Commission (NR(|) to manipu­
late the controls of a nuclear reactor. The training
program used in the plant visited by BLS staff to
prepare operators for the NRC licensing test re­
quires between 6 months and a year to complete and
includes'extensive training in nuclear physics, radia­
tion protection, and .power plant operations. The
NRC operator’s license must be renewed every 2
years; since nuclear power generation is a rapidly
evolving technology, the power company maintains
an ongoing retraining program for its operators.
Some utilities are installing simulators that will be
used to train nuclear operators. Control room super­
visors are required to hold a senior operator’s li­
cense which, in the company visited, requires an
additional 6 months of training.
About one-half of the workers in electric and gas
utilities are unionized.. Of the several unions repre­
senting utility industry employees, the largest are the
International Brotherhood of Electrical Workers and
the Utility Workers Union of America.
Specific provisions relating to technological change
are not commonly found in collective bargaining
contracts for this industry. There are, however, con­
tract provisions pertaining to seniority, layoffs, job
training, and promotions that could be applied to job
losses resulting from technological change.

24 Project Independence (Federal Energy Administration, Nov.
1974), pp. 6 1 -7 2 .
2<
iWillis J. Nordlund and John Murrtford, “ Estimating Employ­
ment Potential in U.S. Energy Industry” , Monthly U ihor R eview ,
May 1978, pp. 10—13.




168

SELECTED REFERENCES
Comar, C. L. “ Putting Plutonium in Perspective,”
World, December 1, 1976, pp. 39—41.

Electrical

“ Mobile Equipment Paces System Growth and Slows
Spread,” Electrical World, June 1, 1974, pp. 2 5 2 -5 3 .

Federal Energy Administration. National Energy Outlook — 1976. ,

Cost

“ Nuclear Power Claims Major Capacity Role,” Electrical World,
June I, 1974, pp. 83—
89.

“ From Fossil to Fusion:,A Milestone Century of Technological
Progress,” Electrical Wo^ld, June 1, 1974, pp. 76-78.

U.S. Department of Energy. Electric Power Supply and Demand
1978—1987 for the Contiguous United States. July 1978.

Graham, John. “The New; Coal Age, Utility N eeds will Bring
Unprecedented D em ands’ Electrical World. June 1, 1975, pp.

U.S. Department of Energy, Energy Information Administration.

3 7 -4 4 .

Projections of Energy Supply and Demand and Their Impacts:
Annbal Report to Congress, Vol. II, 1977. Chapter 10, pp. 2 0 5 -

“ Flow to Enhance Productivity,” Electrical World, November 1,

26.

1976, pp. 3 9 -4 1 .

Electrical

McGraw, Michael G. “.Fleet Management Becomes More Sophis­
ticated,” Electrical World, August 1, 1975, pp. 35—
50.

“ Utilities .Weigh Economics of Nuclear vs C oal,”
World, January 1 1976, pp. 21 —
23-

“ Mechanization RevolutioaiaeS Construction,” Electrical World,
June I, 1974, pp. 164—
6&.

Weaver, Paul H ., “ Behind the Great Scrubber Fracas,” Fortune,
February 1975, pp. 106-14.




16 9

Technology and Labor in
Insurance
Gustav A. Sallas

Summary

Employment in the insurance industry increased at an
annual rate of 2.0 percent during 1960-78 as personal
consumption expenditures for insurance reached higher
levels. A total of 1.2 million persons were employed in the
industry in 1978. The number of persons working in
clerical occupations is expected to rise at a slower rate
than total employment during the period 1978-85 as
computers and related technologies increasingly reduce
labor requirements in data processing tasks.

Insurance carriers (SIC 63) were the first industry to
apply computers to business office procedures on a wide
scale with the advent of electronic data processing (EDP)
in the early 1950’s.1 Since then, the industry’s substantial
accounting and statistical requirements, its vast data
storage and retrieval operations, and the mass of
paperwork it produces have pushed it to the forefront of
EDP utilization. Practically every function of an
insurance carrier has been computerized, and almost all
firms have applied EDP and related technology to at least
part of their operations. By 1985, the industry may
employ more than 30,000 persons in computer positions.2
The spread of EDP and the growth of the industry have
been mutually supporting. A rising population in a
growing economy, with concomitant increases in per­
sonal income and expenditures for insurance, have made
it nearly impossible for firms to function without EDP.
Moreover, computers have made it easier for carriers to
take advantage of statutes allowing the underwriting of
different lines of insurance by the same company. The
main problems created by this rapid growth have been
gigantic and cumbersome files, interminable transcrip­
tion of the same data from one form to another, and a
myriad of tedious and repetitive operations; for all of
these, EDP and related technology is ideally suited.
New technology has brought about substantial
improvements in productivity in a wide range of laborintensive insurance processing operations. The more
widespread use of computers, along with optical
character recognition equipment, remote computer
terminals, microfilm technology, and related in­
novations, has reduced unit labor requirements in data
storage and retrieval, computations, billing functions,
and processing of claims, bids, and proposals.

Technology in the 1980’s
Technological developments in the insurance industry
have centered on the application of EDP to an increasing
number of the industry’s functions, particularly to the
management of information. A summary of the major
technological changes is presented in table 4.
To cope with the huge volume of paperwork,
calculations, and records generated by rapid industry
growth, an increasing number of firms are using EDP.
Computer programs have been developed for a wide
range of insurance industry functions, ranging from
actuarial research to claims processing. In addition, EDP
is used extensively by the industry to perform basic office
tasks such as word processing (typing, copying, printing),
mail handling, and check writing.
Technological advances in the industry have led to the
merger of a vast electronic data base and computation
capability with on-line communications networks, output
devices, and office operations equipment. Utilizing this
technology, many of the major operating functions for all
insurance lines may be performed by a central computer
on demand from a terminal in the home office or in any
field office (including overseas offices), with the results
made available either through visual display devices or on
hard copy. More complex operations are constantly
being devised, and worldwide electronic communications
and operations networks are being created. In addition to
EDP, more extensive use of equipment such as closedcircuit TV and other audiovisual equipment for training,
new and improved copying devices, text-management
equipment, and electronic calculators also have improved
efficiency in insurance industry operations.

■Standard Industrial Classification 63 comprises stock and mutual
enterprises which underwrite all types of insurance, primarily life,
accident and health, property, casualty, surety (financial responsibility),
and title insurance. This study does not include independent agents and
brokers (SIC 64) who sell insurance underwritten by others or who
render services to insurance carriers or policyholders.
2Unpublished BLS data which include both insurance carriers (SIC
63) and insurance agents and brokers (SIC 64).

Reprinted from BLS Bulletin 2033 (1979),
Technology and Labor in Five Industries.




170

Table 4.

Major technological changes in the insurance industry
D escription

Labor implications

Diffusion

Electronic data processing
(ED P)

E D P is being applied to an increasing num ber of insur­
ance industry functions including billing and collection,
actuarial research, underw riting, and claims processing.
M inicom puters are being introduced more widely in
various departm ents within an insurance firm and in
some instances are being joined with a central computer.
These small com puters edit, correct, and preprocess data
at the departm ent level before the data are transm itted
to the main com puter installation. They also retrieve and
m anipulate data already stored in integrated electronic
d ata b ase/d ata com m unications systems and carry out
other functions.

E D P technology has reduced
unit labor requirem ents for file
clerks, typists, and other cleri­
cal staff. Em ploym ent of pro ­
grammers, com puter operators,
and related staff has increased
along with the num ber of in­
stallations.

E D P technology will be dif­
fused more widely over the
next decade as new, less costly,
and technologically improved
small com puters become avail­
able.

Integrated electronic data base/
data com m unication systems

All d ata are recorded on magnetic tape, discs, or drums
and processed through a central com puter installation.
Inform ation can be retrieved instantly at base or remote
locations anywhere in the country and displayed on video
screens or in printed form. The data may be updated, de­
leted, o r segmented for specific functions. The central
processing unit can activate various electronic devices
which will utilize the stored d ata to perform functions
such as accounting, billing, internal and statutory reports
and statistical tables, correspondence, and prom otional
brochures.

Reduces the num ber of file
clerks required.

The num ber of systems in use
is increasing. M ost of the m ajor
carriers have already accom ­
plished the conversion to the
system; others are using the
building block approach, com ­
puterizing their files and opera­
tions systems function by func­
tion, with a view to eventual
integration.

Photoelectric devices are being used to scan encoded
docum ents, interpret them, and produce a magnetic tape
which triggers operating systems and also constitutes a
perm anent record of each transaction. They can produce
outp u t as disks or video display devices.

Reduces the num ber of keypunchers and mechanical sys­
tems required for EDP.

W idespread, particularly in pre­
mium billing and collection
operations, prom otional cam ­
paigns, and data bank input.

M ark sense input system

Special writing tools which trace symbols (lines, dots, or
curves) on ordinary paper are being used to produce in­
form ation readable by computers.

Reduces the num ber of keypunchers required for ED P.

N ot widely used as yet. Ex­
pected to gain acceptance in
the near future.

Portable com puter
terminal

M icrocom puters in briefcases, complete with keyboard,
video display unit and line printer, are being joined to
integrated electronic d ata b ase/d ata com m unications
systems by ordinary telephone.

As yet minimal.

Expected to be used more ex­
tensively by salespersons.

Technology

D ata input devices:
Optical character
recognition

stalled increasingly larger and more complex EDP
hardware. This compelled the user to attempt to process
through a large computer a growing number of minor
operations (as for example, the management of office
supplies). The minicomputer will be used more extensive­
ly to perform such tasks as payment of policy dividends
and claims and, through a terminal, to edit and process
data prior to input into the central system. A minicom­
puter programmed for an increasing number of insurance
applications ultimately may replace the desktop cal­
culator in an insurance firm’s head office and
branches.3
Billing and collection. A common EDP application is
premium billing and collection. Although some carriers
prefer to collect premiums through their agencies, the
trend is toward home office collection, often through the
“turnaround” billing system.
In turnaround billing, a machine readable notice of
premium due produced by the computer is returned by

Planning and implementing EDP applications are
preceded by intensive analysis of existing procedures and
practices in order to develop the appropriate program,
and conversion to EDP (including personnel retraining)
is accomplished gradually to' avoid disrupting current
operations. With the advent of the new generation of
small computers and the corollary drop in computer
system operating costs, EDP and integrated systems will
be widely applied in the industry during the 1980’s.
EDP applications

The number of insurance industry functions which are
being converted from manual to EDP operation is
growing. Computers are being applied to a wide range of
major activities including billing and collection, actuarial
research, underwriting, and claims. The industry consen­
sus is that during the balance of the decade and through
the 1980’s computer technology will be extended to other
areas, with unit labor requirements in clerical operations
expected to continue to decline. Employment of com­
puter programmers, systems analysts, and other com­
puter specialists is expected to continue to increase as
computer use grows.
The further introduction of minicomputers is expected
to bring about additional changes in insurance industry
operations. Until recently, the insurance industry in­



3At present, branch offices have little or no processing equipment and
often no EDP systems whatever. Most branch equipment consists of
key-entry devices for input to home office computers. This is
particularly the case among property and liability insurance carriers.
Current technology permits the home office to transmit computer
programs from the central EDP installation, thereby eliminating the
need for programmers in field offices.

171

the policyholder with the remittance. The computer puts
the payment data on a magnetic tape or disk for the
accounting department, calculates the agent’s commis­
sion on each premium, and credits the appropriate
account. These systems are currently operating at the rate
of about 5,000 remittances an hour, but equipment
already available is capable of handling over 40,000
remittances an hour.
Actuarial research. Computers also are reducing unit
labor requirements in actuarial tasks. In life insurance,
actuarial research yields studies and projections based on
mortality and morbidity (the proportion of disease cases
to population) rates as well as on records of premium
earnings and of policy lapses and maturity. An actuarial
department calculates each carrier’s premium and
dividend rates and provides risk selection guidelines,
among other things. Actuarial work is well suited for
EDP since it requires the quick retrieval of large amounts
of statistical data, the performance of sophisticated
mathematical analyses and complex computations, and
the production of statistical tables to meet the company’s
reporting requirements.
Underwriting.
Underwriters
review
all
policy
applications, pass upon endorsements (changes to policy
conditions), and determine the amount and type of
reinsurance required. Although the decision to accept or
reject each risk must still be made by a trained under­
writer, EDP experts and underwriters are working
together to expand the computer’s participation in under­
writing. Operations already performed electronically
include logging for transaction control, file search for
possible related material, and coding, rating, and policy
issuance.4 Electronic processing can control and monitor
a policy application through the entire underwriting and
policy issuance cycle, producing daily reports on case
status at each work station, thereby saving time and
labor—an important factor since 10,000 new and renewal
applications may be processed daily in the underwriting
department of a major carrier. One property and liability
company, for example, reduced the personnel in its
policy-issuing operation alone by one-third after conver­
sion to EDP.5

involved. The widest variation in claims procedures
occurs within the life and health insurance companies.
Claims that call only for death benefit payments require
only eligibility verification and check issuance. In
integrated electronic systems, the search capability of the
system identifies all policies for a given claimant, and the
data base will show the current status of the policies,
compute the amounts payable, activate the check-issuing
devices, make the appropriate entries into the general
ledger, and block any other transaction from occurring.
Health and disability insurance claims involve more
intricate processes which require complex EDP
programs. These policies provide for payments for
services as well as income protection and indemnity.
Health and disability claims require large numbers of
examiners, file clerks, and typists. In an EDP system, the
examiner can view the claim history file on a video screen,
determine coverage, adjudicate the claim, calculate the
benefits, issue payment checks, and generate the required
correspondence.
In the property and liability field, the growing
complexity of the coverage has placed a strain on claims
processing and encouraged more extensive application of
EDP. For example, no-fault automobile insurance in
many cases sets a time limit for the insurer to respond to
the claim. To achieve growth with a stable work force will
require improvements attainable only through electronic
automation—a significant challenge in light of the
substantial manual processing of claims that occurs in
many property and liability companies. The potential of
EDP to improve efficiency in claim processing is
illustrated by the following example of a system in
operation. An appraiser involved in an auto damage
claim first checks eligibility and coverage through the
home office computer, assesses the damage, and then
writes a check to settle the claim. A form filled out at
settlement is fed into a transceiver which sends an image
by telephone line to the home office (usually at night) for
accounting and central files. At the receiving end, the
unattended device produces a photograph of the forms
and turns itself off when the transmission is completed.
Integrated ©©mmyrnication interns
Electronic data base/data communication systems are
being applied to the substantial data handling re­
quirements associated with the insurance industry. These
data generally must be maintained in current form for
several decades and are periodically required for
immediate use at various locations in the home office and
in the field. As many as 12 separate records are
maintained on a given policy, spread among various
departments and often duplicated in field offices. Inm ost
instances, these functional files include duplicate infor­
mation. The data overlap often results in one file being
current while its duplicates remain static and become out­
dated.
Integrated electronic data base/data communication
systems have several advantages, including the capability

Claims. The application of EDP to claims processing
varies widely with the type of insurance involved; even
within the same line, the time frame over which a claim
must be serviced greatly affects the automated processes
insurance policies usually consist of three parts: (1) A jacket, or
printed section, which contains only constants applicable to a given
carrier or kind of policy; (2) the declarations—a printed form bearing a
typed description of the variables involved, such as premium or
particular risks; and (3) endorsements, which usually are extensions of
either the jacket or the declarations. (Both the printing and typing of
policies have been computerized.) Surveys in the property and liability
lines have shown that electronic processing of the declarations
eliminates an error factor of approximately 8 percent which plagues the
manual operation.
. 5LOM A
Resource,
March/April, 1976.




Life

Office

Management

Association,

S 72

to consolidate all data related to a single policy in the
company’s central computer installation. The systems
search capability permits instant identification of all
policies connected with one name and of all names
connected with one policy; review of all documents which
meet a given specification; and browsing through entire
files by flashing their contents on a video screen document
by document. In advanced installations, the information
is stored on magnetic tape or in disk packs which can hold
up to about 30,000 pages of data. The files may be
updated or deleted at will, and may easily be segmented
for specific purposes and the various segments protected
from unauthorized use. To protect the data from
accidental destruction, duplicate disks or tapes are kept,
usually off premises. Through advances in communica­
tion technology, data are accessible rapidly at all sta­
tions—immediate and remote—at terminals equipped
with video display devices and line printers. Each system
is linked by a communication and teleprocessing network
of ordinary telephone circuits and leased telephone or
telegraph lines. The integrated electronic data base/data
communication system produces internal statistical
statements on demand as well as the periodic reports
required by regulatory agencies. Because of their
capability to improve data handling, electronic data
base/data communication systems will be used more
extensively over the next decade.
Data Input devices
Electronic devices which feed into a computer directly
from documents are increasing accuracy and displacing
keypunch operators. Optical character recognition
technology is expected to assist the industry to cope with
the massive increase in paperwork to accompany future
growth. Without such devices it would already be
impossible for the large carriers to handle expeditiously
the huge number of inquiries, applications, remittances,
and claims they receive every day.
Optical character recognition (OCR) is the instant
interpretation and transmission of the symbols and
alphanumeric characters which constitute a coded entry,
by a photoelectric device known as an optical scanner. In
the insurance industry, the optical scanner is used
primarily for premium billing and collection. Additional
uses are being developed in the processing of other
documents which can be preprinted or encoded in
machine readable form, such as policy applications,
surrender notices, and reinstatement requests, as well as
loan applications in life insurance and loss statements in
property and casualty claims.
Coded machine readable documents can be produced
by the computer or independently by special purpose
typewriters and by the “Mark Sense” system, which
involves the use of graphic input devices. These consist of
a form listing a number of specific pieces of information, a
writing tool for touching the items to be entered, and a
photoelectric device which converts the data to digital



173

form for the computer. Data from graphic input devices
can be fed onto magnetic tape through a multiplex unit
which can support up to 15 stations, each capable of
carrying the data stream for- a different insurance
operation—such as underwriting, claims, or investments.
These devices generally cut data entry time by 30 percent.
In one instance, a carrier which processed about 400 bids
and proposals a month with a staff of 11 was able to
process nearly twice as many with a staff of 9 after
switching from keypunch to an electronic pen.
Increasing utilization also is seen for the portable
input-output device, or “briefcase terminal.” The latest of
these devices, which weighs under 15 pounds, is equipped
with a keyboard, a microprocessor, a small video screen,
and a compact line printer. Connected to an ordinary
telephone (by placing the telephone receiver on a cradle in
the device) and plugged into an electrical outlet, this
device links an agent anywhere in the world with the
company’s central EDP installation. Hard copy is
produced at the rate of 1,800 characters per minute. With
instant access to the company’s central computer 24 hours
a day the year round, the agent can make a presentation
and place the order for a policy.
Userofslnt technology
Advances in microfilm technology are resulting in
space savings in records storage, faster retrieval of policy­
holder information, and lower labor requirements for file
clerks, typists, and other clerical staff. Microfilm
processes are particularly advantageous to the insurance
industry since large quantities of records are produced,
such as policy and loan applications, paid drafts, and
Medicare reports, that cannot be stored in the computer
in digital form because of their format or statutory
requirements.
The original microfilming process (which insurance
companies pioneered in the 1940’s) is giving way to micro­
fiche, in which the roll film is replaced by 4" * 6" cards
each holding about 100 frames (microphotographs) of
legal size documents. (Already under development is a
high-reduction technique known as ultrafiche, which will
increase the capacity of the card substantially.) Micro­
fiche expedites handling of hard copy documents and
practically eliminates the storage space required for these
items: The contents of about 140 file cabinet drawers can
be stored in an 18-inch card tray. When microfiche is
coupled with computer-output-micrographic (COM)
devices, data are transferred directly from the computer
to microfiche without need for intermediate hard copy.
Multiple hard copies can be produced rapidly from the
microfiche through a nonimpact printer.6Thus, the COM
system can produce in a few seconds, from data in the
computer, a microfiche frame for the record, available for
instant viewing on a screen, and a number of hard copies
for distribution. One life insurance carrier made two hard
copies each of 230,000 documents from computer tape
through microfiche in 15 hours—an operation that would

have required 200 hours with an impact printer. Another
carrier, which produced nearly 300 million microfiche
frames in 1974, has plans to double this volume within the
decade. The COM process is particularly advantageous
for use in property and casualty underwriting where there
are frequent changes in policy data (as when an
automobile is traded in) many of which require
transferring or reprogramming data needed for historical
purposes.

Industry Outlook
The outlook is for continued expansion in the three
major segments of the insurance industry—life, health,
and property/liability. About two-thirds of the Nation’s
total population are covered by life insurance policies
issued by the 1,750 life insurance companies with
headquarters in the United States.7 The face value of
policies in force totaled $586 billion in 1960, $2,583 billion
in 1977, and was projected by the U.S. Department of
Commerce to reach $3,097 billion in 1979. The number of
life insurance policies in force totaled 390 million in 1977,
38 percent more than in 1960. Purchase of new life
insurance totaled $367.3 billion in 1977, compared to
$74.4 billion in 1960, with the average amount of life
insurance in force per insured family rising steadily and
amounting to $36,900 in 1977.
Group life insurance, in particular, is gaining rapidly as
life insurance benefits increasingly are being included in
employee benefit programs, with coverage extended to
dependents of group insurance certificate holders. Over
89 percent of all group life insurance contracts covered
employer-employee groups in 1973 (most recent year for
which data are available), and the average coverage per
employee for these contracts was about $12,000. In
addition, millions of employed and retired people are
participants in retirement plans operated by life insurance
companies. Even greater growth is anticipated in
individual retirement plans under the Employee Retirement Income Security Act of 1974, which permits persons
employed in firms with no private retirement plan to
purchase an individual plan of their own.
Health insurance has expanded substantially. Expan­
sion has been accelerated by the increasing inclusion of
health benefits in labor contracts and by insurance
6Several printing techniques are available which do not involve
striking paper with a mechanical hammer, but use photographic,
chemical, or magnetic ink processes and specially treated paper. In one
of the most sophisticated of these, the Ink-jet Printer, a stream of
magnetic ink droplets is shot towards the paper and deflected by elec­
trostatic plates to form the desired character. Speeds in excess of 75,000
lines per minute are possible with the Ink-jet Printer, although such
speeds are not yet available in commercial versions.
7The sources of statistics on amounts of insurance in force, number of
policies in force, premiums written, number of persons covered, and
related data included in this section for these three segments of the
insurance industry are as follows: Life insurance, American Council of
Life Insurance; health insurance, Health Insurance Institute; and
property and liability insurance, Insurance Information Institute.




174

company participation in the Medicare program enacted
in 1966. At the end of 1976 (latest year for which data are
available), 177 million persons were covered under
provisions of one or more forms of private health
insurance. Individuals under age 65 accounted for 93
percent of the total. The older groups were insured under
private plans to supplement Medicare benefits. Premiums
received by the industry for health insurance coverage
rose to $24.3 billion in 1976, nearly double the premium
income received in 1971.
The property and liability segment of the insurance
industry also has recorded substantial growth. Between
1960 and 1977, net premiums written for property and
liability insurance increased by almost 400 percent, from
$15 billion to $74 billion. Auto insurance in 1977
accounted for about 42 percent of total property / liability
premiums.
Accompanying the premium growth in the major
categories of insurance, as discussed above, has been
increasing diversification in the types of insurance under­
written by insurance carriers. Insurance companies also
are merchandising mutual fund shares and other
investment plans, and giant department store chains have
formed insurance subsidiaries and are selling policies
through in-store booths at their hundreds of outlets.
Already many carriers have formed holding companies
encompassing estate planning and equity investment, real
estate, and data processing.

Employment and Occupational Trends
Employment
Insurance is a major industry that employed 1.2 million
persons in 1978, or about 1 out of every 4 persons working
in the finance, insurance, and real estate sector.
Employment in insurance has been rising steadily as gains
in personal consumption expenditures for insurance
resulted in a steady rise in policies issued, premiums
written, and insurance in force per insured family.
Total employment in insurance increased at an average
annual rate of 2.0 percent during 1960-78. The annual
employment growth rate was 1.8 percent during 1967-78
and 1.7 percent during the earlier 1960-67 period.
Employment in the three largest components of
the insurance industry—which combined accounted for
about 94 percent of the total industry work force—fluc­
tuated over the period 1967-78. Establishments primarily
engaged in underwriting life insurance (SIC 631)
continued to employ the largest number of employees,
522,500 in 1978, but their share of total employment
declined from over 50 percent in 1967 to about 44 percent
in 1978.
Second in size of work force is the fire, marine, and
casualty insurance segment of the industry (SIC 633)
which employed 460,600 people in 1978, a gain of 44
percent since 1966. Its share of total employment in the
insurance industry rose slowly, reaching 39 percent in
1978.

The accident and health segment of the industry (SIC
632) is third in terms of employment, but it is the fastest
growing and accounted for about 12 percent of the work
force in 1978. Employment in establishments which
underwrite accident and health insurance reached
138,900 persons in 1978, more than double their 1966
employment.
The rise in total employment in the industry has been
accompanied by increased use of computer and related
technology. Despite the fact that in many applications
unit labor requirements have declined substantially, with
the computer doing in a few hours the work dozens of
employees formerly turned out in one day, the adverse
effect on employment has been mitigated by several
factors. The high rate of attrition which prevailed among
the clerical work force required to perform the industry’s
massive, repetitive data processing tasks has facilitated
conversion with minimum dislocation. Moreover, rapid
business expansion and advance planning of work force
changes appear to have obviated the need for any
significant layoffs, and conversion has been carried out
mainly by retraining and relocating personnel. Some
companies however, have imposed a total freeze on the
hiring of clerical workers for protracted periods after
computerization to facilitate reassignment of existing
staff.
Occupations
A wide range of occupations in the insurance industry
has been affected by the application of EDP and related
technologies. Because of these innovations, employment
in clerical occupations is expected to increase at a slower
rate than that envisioned for all employment in the
industry. Consequently, clerical workers, who constitute
nearly half of the insurance carrier work force, will make
up an increasingly smaller proportion of total insurance
industry employment by 1985. The demand for and
relative importance of a number of clerical positions,
including secretaries and typists, keypunch and other
office machine operators, bookkeepers, and file clerks,
is expected to decline.
However, employment in managerial and technical
positions including those related to the planning,
installation, and operation of electronic computer
systems is expected to increase. Demand will rise for
executives and other officials to determine EDP policy,
manage data base and teleprocessing networks, and
supervise program development and the recruiting and
training of computer applications personnel. Employ­
ment of systems analysts, programmers, and computer
console operators and related computer operating staff is
expected to continue to increase. The insurance industry
will continue to be receptive to further application of
technology to insurance operations since the industry
ranks second only to commercial banks in the proportion
of its work force engaged in data processing activities.
Employment of persons engaged full time in selling all
lines of insurance directly for the carriers (excluding




175

general agents and brokers) is not expected to keep pace
with the rapidly expanding volume of sales, primarily
because of the increasing number of policies sold to
groups and the increasing sales of policies which cover
several perils previously covered by separate policies.8
As indicated earlier, occupations in clerical fields have
been those principally affected by technological change;
these include file clerks, keypunch operators, and typists.
Although the insurance industry will continue to employ
substantial numbers of young high school graduates,
many of them women, computerization of practically
every mechanical operation is expected to reduce the
availability of low-skilled, entry-level clerical positions.
Adjustment ©f workers to technotogical change
The further diffusion of computer and related tech­
nology in the insurance industry is not expected to bring
about major displacement. In most cases, the initial
impact of the transition to new technology in data
processing operations has been slight because of the
increased workload entailed in the conversion. Often, a
special department was established during the early stage
of the changeover to handle the corollary personnel
problems. Moreover, as already indicated, a high degree
of attrition of clerical workers—the occupational group
most affected by new technology—has eased the
transition from manual operations to electronic data
pfocessing and made layoffs unnecessary. Continuous
industry growth has permitted absorption of displaced
personnel in all categories, and coordinated retraining
and relocation by the employers have reduced dis­
locations to a minimum.
Relatively few employees affected by new technology in
the insurance industry are union members. The principal
union in the insurance industry, the Insurance Workers
International Union (IWIU), thus far has concentrated
on organizing insurance agents (salespersons), and this
has been the category least affected by technological
innovation. In late 1976, the IWIU negotiated to organize
the clerks of one of the largest insurance carriers. The
Office and Professional Employees International Union
also has organized in the industry.
The very few union contracts in effect in the industry
incorporate specific provisions relating to technological
displacement. One such provision which provides job
security for employees whose jobs have been affected by
new technology reads as follows: “... It is further agreed
by the parties that no persons filling jobs within the
presently existing collective bargaining unit will be
subject to layoff in the event that jobs are abolished or
altered by the introduction of data processing equipment,
computers or other automated equipment. . . ” For
displaced workers, provisions are in effect which provide
retraining and preference for any jobs resulting from
automation or conversion to electronic data processing.
8Occupational Outlook Handbook, 1978-79 Edition, Bulletin 1955
(Bureau of Labor Statistics, 1978), p. 763.

No information is available concerning the effect of
reorganizations, mergers, and consolidation of branch

offices on employment. There is, however, no evidence
that these trends have caused any widespread dislocation.

SELECTED REFERENCES
Bowers, Dan M. “Intelligent Terminals and Distributed Processing,”
The Office, September 1976, pp. 86 ff.

International Business Machines Corporation. Property I Liability Field
Office Application Systems— Executive Overview, May 1975.

Cantrell, Gary L. “Remote Job Processing as an Alternative,” Best’s
Review, Life/Health Edition, July 1976, pp. 68-70.

Life Office Management Association. Insurance Information Process­
ing: A Look at Our Future, LOMA Systems and Procedures Report
28, 1975.

Fischer, Robert A. “Insurance Tomorrow: The Data Processing
Picture,” Best’s Review, Property/Liability Edition, May 1975, pp.
104-09.

Roach, Thomas. “The Data Processing Organization of the 1980’s,”
LOM A Resource, November 1975, pp. 29-30.

Fromm, Erwin F. “The Mechanical Underwriter,” Best’s Review,
Property/Liability Edition, June 1975, pp. 16-18.

“Telecommunications Speed Settlement of Auto Claims,” Best’s
Review, Property/Liability Edition, November 1975, p. 102.

Goldbeck, George. “Mini-Computers—A Big Part of the Future,”
Best’s Review, Property/Liability Edition, January 1975, pp. 78-81.

U.S. Department of Labor, Bureau of Labor Statistics. Impact o f Office
Automation in the Insurance Industry, Bulletin 1468, 1966.

Goldbeck, George. “Information Processing in the P/ C Business,” The
National Underwriter, Sept. 5, 1975, p. 2.

Valovic, Stefan. “Survey Shows Risk Managers Make More Use of
Computers,” Business Insurance, Dec. 1, 1975, pp. 19-20.

International Business Machines Corporation. Group InsuranceApplication Description and System Planning Guide. White Plains,
N.Y., May 1975.

Vanderbeek, Robert E., and H. Thomas Verdonk. “Conputer System
Offers Personalized Customer Service,” Best’s Review, Life/Health
Edition, February 1976, pp. 66-67.

International Business Machines Corporation. Individual Life In­
surance: D B / D C Information Systems Design Concepts, January
1975.

Wray, Theodore S. “Field Office Video Units Improve Policyholder
Service,” Best’s Review, Life/Health Edition, June 1975, pp. 81-84.




176

Technology and Labor in
Metalworking Machinery
A. Harvey Belitsky

Employment in the industry stood at the relatively high
level of 371,500 persons in 1980. The average annual
increase from 1960 to 1980 was 1.4 percent (about the
same rate as for all durable goods manufacturing). The
outlook for employment growth from 1980 to 1990 is in
the range of 0.8 to 3.8 percent (average annual rates) as
projected by BLS, based on alternative versions of
economic growth. Increases are projected for virtually all
of the industry’s occupational groups, but the number of
craft workers is expected to grow only half as rapidly as
the number of operatives. A shortage of skilled workers
could remain a principal obstacle to expansion in the
metal-cutting machine sector for at least the immediate
future.

Summary
The metalworking machinery industry is rapidly
increasing the application of the numerically controlled
(N C ) machine tool. NC machines accounted for an
estimated 30 percent of the value of machine tools
installed in the metal-cutting machine tool sector in 1979.
Increased diffusion of NC is expected in the 1980’s in
response to a host of economic conditions, including the
advanced age and low productivity of the machine tools
in use, the need to meet increasingly precise and complex
requirements for machined parts, and the shortage of
skilled workers.
While utilization of NC machine tools is likely to
increase steadily among the industry’s large and mediumsize plants, the many small shops which manufacture
simple parts will still rely heavily upon manually operated
machine tools. Other, more sophisticated technologies
such as machining centers (multifunction NC machines),
controls utilizing sensors, and NC by computer are also
economically feasible, principally for the larger plants.
Complex microprocessors, which have fallen sharply in
price, permit firms of all sizes to use various intermediate
technologies which are not as sophisticated as NC. In
general, these technologies increase output per employee
hour, improve quality, and reduce occupational skill
requirements.
Although definitive measurements of productivity for
the industry as a whole are not available, very small
productivity improvement in 1960-78 is suggested by an
average annual rise in output of about 2 percent and in
employee hours of less than 1 percent. Wide swings in
output—with associated lags in adjustment of hours—
and aging equipment are major reasons for the industry’s
comparatively low productivity growth rate.
While the industry’s dollar outlays for new plant and
equipment rose by nearly two-thirds from 1966 to 1978,
expenditures in real terms did not surpass the 1966 peak.
Real expenditures rose to a comparatively high level in
1974, declined in 1975, and then rose in the succeeding 3
years. They have generally continued to increase, and this
may enable manufacturers to compete more effectively
with imports for the large tooling requirements of
automobile, commercial aircraft, and defense-related
manufactures.

Industry Structure
This study examines' the metalworking machinery
industry as a whole (SIC 354) and three major sectors
within the industry: Metal-cutting machine tools (SIC
3541); special dies and tools,die sets,jigsand fixtures,and
industrial molds (SIC 3544); and machine tool accessories
and measuring devices (SIC 3545). Reference is also made
to metal-forming machine tools (sic 3542). The other
metalworking sectors not covered in this study are:
Power-driven handtools; rolling mill machinery and
equipment; and such machinery as gas cutting and
welding equipment.
Some characteristics of this industry tend to limit
productivity growth. An estimated three-fourths of the
industry’s output is in batches of less than 50 pieces. This
holds true particularly for the machine tool makers, who
often produce special machines for their customers, and
for the tool-and-die firms, which frequently produce oneof-a-kind parts. Moreover, establishments in metalwork­
ing are comparatively small and highly specialized. Such
firms often find it economically unfeasible to invest in
new equipment. For example, the tool-and-die sector of
the industry (SIC 3544) is made up of 7,100
establishments, averaging only 15 employees. In
comparison, the average size of all manufacturing
establishments is 55 employees. Additionally, the sharp
fluctuation in industry output over the course of the
business cycle tends to reduce productivity growth.

Reprinted from BLS Bulletin 2104 (1982),
T echnology a n d L a b o r in F ou r Industries.




177

T@
e!hiT)® gy in tb® 19®05
S©
s
Numerical control (NC) of machine tools is the most
significant new technology introduced in the metalwork­
ing industry in the past 25 years. It has experienced
substantial and frequent changes in concept and/or
design. However, some metalworking firms are investing
in intermediate technologies, i.e., technologies which are
not as sophisticated or expensive as NC, but which
nevertheless improve productivity. These include digital
readouts and manual-data-input controls which are
applied largely to conventional machine tools. Innova­
tions have also taken place in management techniques
and cutting-tool materials. The major technologies, their
diffusion, and their labor impact are discussed in more
detail below and are presented in table 2.1
WumeroeaSly ©ontrolledl machine tools
Numerical control (NC) involves the automatic control
of a machine tool’s movement by an electronic controller
or special computer which reads instructions in digital
form. NC tools are more productive than manually
operated tools. They reduce setup time; consequently a
higher proportion of working time is spent on cutting.
The need for costly jigs, templates, and other tooling
devices is eliminated. NC tools can produce parts with
greater precision and uniformity, thereby further saving
machining time and minimizing scrap losses. NC may
make possible the production of complex parts that could
otherwise not be turned out, or only at great cost; and the
process permits engineering changes on a part by merely
changing portions of the input program.
NC enhances managerial control by predetermining
and coding every stage of machining onto a control tape.
It becomes possible for managers to plan more accurately
such operations as machine loading and shop scheduling,
and > is much easier to predict labor and machine
t
requirements.
NC prc\ ides the opportunity to attain some automation
in the small batch production which characterizes this
industry. The innovation may be more fully appreciated
by characterizing NC as a manufacturing system, and not
merely a means to control a machine.
Despite the advantages of NC, only about 3 percent of
all metal-cutting tools in the metalworking industry in
1976-78 were NC.2 However, they accounted for a much
larger proportion of output. Another indication of the
importance of NC is the value of recently installed NC
machines. In the metal-cutting machine tool sector, value
'This study does not include the more than 20 technologies that
are identified as nontraditional machining processes, although
production and application of some of these processes are
increasing. Two of these are electrochemical machining and
electrical discharge machining. Electrical discharge machining
is used widely in tool-and-die shops.
2
“The 12th American Machinist Inventory of Metalworking
Equipment 1976-78,” A m erican Machinist, December 1978, p.
136.




178

of shipments was estimated to be 30 percent of all
machine tools installed in 1979.
The reason for the lack of diffusion of NC throughout
the metalworking machinery industry is the small size of
the majority of the firms; they have limited funds for
investment in this comparatively expensive technology.
Although NC is intended for small batch operations, the
risk of investment may be too great because of the
volatility of demand for the industry’s products.
Moreover, firms producing simple parts are unlikely to
utilize NC. In addition, NC is not feasible, technically, for
some machining methods, such as broaching. It is also
interesting to note that surveys disclose that only a few of
the general managers in firms without NC fully
understood its workings.3
Currently, except for a small number of comparatively
large firms which use advanced sophisticated NC, most
machine tool shops still rely heavily upon skilled workers
working on conventional tools. Nonetheless, numerous
modest-sized contract tool-and-die shops—also referred
to as contract tooling and machining shops—have
adopted NC because they do a significant amount of
precision machining.
However, if, as expected, NC replaces a large
proportion of conventional tools in the 1980’s, it could
have considerable impact on the metalworking industry.
The application of NC should be accelerated by shortages
of skilled workers and by the growing need for parts of
greater precision. Adding urgency is the steadily
increasing demand for variety and versatility in products.
Firms which introduce or expand their use of NC can
experience pronounced savings in labor and material. A
study of over 350 companies4disclosed direct and indirect
savings of NC over manually operated tools. Reduced
machining time ranged from 35 to 50 percent. Indirect
savings of 25 percent or higher were found for material
handling, scrap, and inspection. A majority of firms did
not even have higher outlays for NC programming if
“process planning” on conventional machines is taken
into account.
With the introduction of NC, the occupational
composition of the work force generally changes. The
number of machine operators is likely to decline for a
given level of production, since one person can often
operate two NC machines. In many cases, skill
requirements are reduced. For example, operators no
longer need to interpret a blueprint in selecting machine
settings. On the other hand, they must be perceptive to a
malfunction. Some firms try to enhance the duties of the
operator of a very expensive NC machine to make the job
’Edwin Mansfield and others, Research an d Innovation in the
C orporation (New York, Norton and Company, Inc.,
1971), pp. 201-202; George P. Putnam, “Why More NC Isn’t Being
Used,” M achine a n d Tool Blue Book, September 1978, pp.
100- 101.
M odern

4Donald N. Smith and Lary Evans, M anagem ent S ta n dards f o r
N um erical C ontrols (Ann Arbor, University of
Michigan, 1977), pp. 185, 192, 212, 214, and 222.
C om pu ter a n d

attractive to a skilled machinist. The new position of
programmer required with NC is being filled in a growing
number of firms by skilled machinists who have received
supplementary training. A somewhat larger number of
maintenance personnel may be required, and their skill
requirements are higher, calling for special training. In
general, NC machinery is operated in two or three
workshifts, without comparable labor additions on the
later shifts.
Numerical control by coimiputer
According to some experts, the combination of the
computer with NC in small batch production compares in
impact to the introduction of the assembly line and
interchangeable parts. Software advances in the late
1960’s made it possible for a computer to convey
numerical data directly to a machine control unit, thereby
eliminating the need for a special control system to
operate a machine tool by tape commands. Computer NC
(CNC), first with minicomputers and later with
microcomputers, made it possible to operate one or more
machines and even to connect to larger computers. CNC
eliminates the problem of the constant redoing of tape.
Also, workpiece program data can be changed in the
control system without the necessity for reading an entire
program tape. A computer can also keep track of the. time
each machine tool is in use.
A significant proportion of the NC units in the
metalworking industry are the CNC type. Costs of
electronic controls have declined so sharply in recent
years that some CNC units are competitive with tapedriven control units and are economically feasible for
small and medium-size firms. Whereas minicomputers
accounted for 50 percent of the cost of a CNC tool in the
late 1960’s, they now account for less than 20 percent. All
the reasons for the lack of diffusion of NC apply even
more to CNC. And direct numerical control (D N C ), in
which a central computer may control up to 100 or more
machine tools, is currently used by only a few of the larger
toolbuilders in the metalworking industry.
The principal new job classifications arising from the
use of CNC are computer programmers, electronic
maintenance personnel, and, in some firms, systems
analysts. Personnel skilled in preventive maintenance
assume great importance with CNC, and machine
problems require a multiskill approach.
H/taefolnlng c©nt®r§
Machining centers are more elaborate and costly than
basic, single-purpose NC. The centers may have automatic
tool-changing systems for selecting among 20 to 100 tools
that bore, drill, mill, and tap. With rotary heads and
tables, a center can work on many surfaces of a part in a
single setup. The centers substantially improve manager­
ial flexibility. Moreover, they raise productivity because
tool changing is less than on basic NC; one operator may
control several machines. While NC may be most suitable
for low-volume metalworking shops, a machining center




can be justified whether the volume of parts is small or
large.
Nevertheless, this technology has only been minimally
adopted by the metalworking machinery industry. There
are many reasons for this slow acceptance, including high
initial cost, simple products which are unsuitable for
machining centers, and lack of managerial knowhow.
Adaptive controls
Adaptive controls utilize sensors that automatically
control such factors as vibration, tool wear, tool or
workpiece deflection, and cutting temperatures. Such
sensors can be integrated within an NC controller, or they
can operate with conventional machine tools. Although
the technology has been available for about 20 years, the
application of these controls in metalworking machinery
is limited to large shops. Utilization by small shops will
depend upon development of improved sensors, their cost
effectiveness relative to the availability of skilled workers,
and the type of work performed.
Reported improvements in productivity in currently
available controls range from 20 to 40 percent, with
largest gains when a part requires diverse cutting condi­
tions and the material is hard to machine. Resides im­
proved machining time, scrap and cutter breakage are
reduced.5
Sensory devices tend to reduce worker skill require­
ments because they assume many functions traditionally
performed by the operator. These devices help to
standardize unit labor time. While machining time on an
identical part can vary by over 30 percent for different
operators, adaptive controls virtually eliminate the
differences.
Computer-aided design and manufacture
Computer-aided design and computer-aided manufact­
ure (C A D /C A M ) constitute a system which utilizes
computer-controlled methods to unite several technologies.
Computers are used to assist in developing designs for
products to be manufactured (C A D ). CAM, among other
things, directs numerically controlled machines and
automatically guides workpieces among machines on
computer-controlled material handling systems. CAD/CAM
is influenced by continuing improvements and applica­
tions in various phases of manufacturing, including:
Assembly with industrial robots; adaptive control;
systems to monitor maintenance; and systems to inspect
parts automatically. The total impact on productivity and
the work force is substantially greater than that of any
single technology.
Currently, only the largest machine tool manufacturers
utilize CAD/CAM . Its diffusion to medium-size firms will
continue to be severely limited by cost and lack of
technical expertise.
Higher productivity resulting from adoption of CAD/CAM
is associated with the shift of workers to more skilled jobs
5D.N. Smith and L. Evans, op. cit., pp. 123, 129, 151.

179

and the reduction in the number of lesser skilled jobs.
Skilled machinists continue to be needed. The number of
drafting personnel is reduced because of less need for
extensive lettering and layouts. Some workers have to be
retrained for new tasks associated with computer
terminals. In addition, worker involvement in decisionmaking will increase although it will be informal, unlike
the practice in some other industries.6
Ipterm ecfete S©<gta@togs<gg
There are other technologies being introduced into the
industry which are not as sophisticated or expensive as
NC, but which improve productivity.7
Digital readout (dro ). This device enables a machine
operator to position the moving portion of a machine tool
more rapidly and accurately. A major change in the d r o
followed the introduction of an electronic display panel
separate from the metering component. A measure of
automatic control is added to almost any manually
operated machine, although readouts can be used for
verification on NC tools, too. The DRO also enables an
operator to change a cutting machine from inch to metric
measure by flipping a switch; this is the most economical
way of providing certain machines with metric capability.
Increases in efficiency result from fewer operator errors
and faster machining cycles. Shop efficiency is also raised
because less time is required for setups, repetitive tasks,
and inspection. The devices decrease positioning times by
up to 80 percent.8 Although machine setups require
highly skilled machinists, operators need less training
than previously to carry out a job with the aid of a DRO.
Also, operator fatigue is reduced.
The use of DRO ’s is expanding among machine tool
builders and is already widespread among tool-and-die
shops. DRO manufacturers anticipate 25-percent yearly
growth in sales to the metalworking industry for several
years. Major improvements on some DRO ’s since their
introduction about 15 years ago can make programming
easier for operators. DRO ’s can be installed on existing
machines and are also relatively reasonable; the payback
period is usually considered to be less than 1 year.
Consequently, they are feasible for the job shop which
cannot afford NC
Manual-data-input control. Manual-data-input (MDl)
control of machine tools is more sophisticated than the
DRO. Both types of controls inform an operator of
6Proceedings, Eighth A n n u al Tri-Service M anufacturing Tech­
n o lo g y C oordination Conference, Arlington, Texas, Nov. 8-12,

1976, p. 170.
’Programmable controllers, normally associated with massproduction industries, and programmable hand calculators are
also intermediate technologies which have had successful appli­
cations in small batch manufacturing. See National Center for
Productivity and Quality of Working Life, New Technologies an d
Training in M etalw orking, Summer 1978, pp. 4-7.
8George Schaffer, “Digital Readout Systems,” Am erican M a­
chinist, May 1979, p. SR-4.




180

machine position, but the M Dl also enables an operator to
change the machine’s position automatically, reducing
further the chance of error. Many M D l’s can be
transformed into NC systems by inserting a tape reader;
most current NC systems contain editing capabilities to
accept programs manually. M Dl usually controls simpler
machines and turns out simpler workpieces than does NC.
Further, some firms do not want or may be unable to add
NC machines. A major difference is that an M Dl machinist
(and not a programmer, as with NC) usually plans and
enters the program for a part.
Increased use of M Dl in recent years by the
metalworking industry has been stimulated by substantial
improvements, in microcircuit technology and declining
costs compared with other controls. A separate
programming department is not needed with M Dl, and it
can be applied in large as well as small shops. While
precise data on the utilization of MDl are unavailable, its
use is spreading among machine tool builders, and even
more rapidly in the contract tool-and-die shops.
Higher labor productivity and improved product
quality are credited to M Dl. In addition, M D l can help
alleviate the shortage of skilled workers because a
machine operator may be able to tend two machine tools.
Also, some M D l’s utilize “shop language” for programming
so that a moderately skilled machinist can use the
programming language after brief training.
Gutting-tool materials
Improved cutting-tool materials can play a substantial
role in the application of highly productive, advanced
technologies. The performance of a $250,000 NC machine
tool depends on the cutting capability of a $30 end mill.
Firms which can utilize the improved tool materials can
more efficiently satisfy material and quality specifications
of customers. The new cutting-tool materials improve
productivity because, unlike older materials, they do not
wear out as fast and thus do not have to be changed as
often.
Such new materials as coated carbides, polycrystalline
diamonds, and special ceramics are being used in place of
tungsten carbide. Applications of the relatively longlasting coated carbides will continue to increase because
of sizable price increases for tungsten. According to an
industry analyst, the coated carbides will increase from a
current application of some 15 percent to at least 25
percent of cutting-tool materials used by the metalwork­
ing machinery industry in 1985.

Group technology
Group technology (GT), a management technique, can
be as important to productivity as new machines. It
involves the grouping of parts on the basis of similar
shapes and/or processing requirements. GT revises the
belief that small batch manufacturing consists of making
distinctive parts from design to end-product. Marked
savings are attributed to GT as a result of improved

production scheduling, reduced inventory, and greater
efficiency in machine loading. Design rationalization and
reduced, as well as more efficient, setup and tooling are
also credited to the process.
GT also could reduce skill specializations which often
exist in medium-size and large machine tool shops.
Unlike the usual practice in Europe and Japan, only a
small proportion of the U.S. metalworking firms which
Table 2.

have introduced GT have broadened the skill require­
ments of their w ork forces.

Small firms cannot afford the sophisticated effort
needed to install GT, and it is used by less than 20 percent
of all toolbuilders. Tool-and-die shops do not utilize GT as
such, but elements of it are present in shop layout
procedures among firms engaged in precision machining
and the manufacture of machine tools.

Ma|®r technology changes in metalworking machinery
Technology

Numerically controlled machine
tool (NC)

Description

Labor implications

•
Tool is controlled by instructions received from tape,
punched cards, plugs, or other media. Allows rapid change
to new product designs; permits stricter tolerances for parts;
and reduces setup time. Useful for small batch production,
because parts can be machined by merely changing tapes
and resetting tool.

Diffusion

Estimated reduction in machin­
ing time of 35-50 percent; typi­
cally used two or three shifts.
Reduces unit requirements for
machine operators; requires less
skill than manually operated
tools, creates new job of pro­
grammer; requires more broadly
trained maintenance personnel.

Three percent of all machine tools
in metalworking are NC, but they
account for a much larger propor­
tion of total output. In metal­
cutting sector, the value of NC
machines is estimated at 30 per­
cent of all machine tools installed
in 1979. Considerable growth in
the number of NC tools and their
share of total output is expected
in the I980’s.
Significant proportion of NC
machines; mainly limited to larg­
er and medium-size machine tool
shops. Expected to increase sub­
stantially in the 1980’s as result
of cost reductions in electric con­
trols; some smaller shops will
introduce it.

Numerical control by com puter
(CNC)

O n-board com puter stores and conveys inform ation direct­
ly to NC control unit; utilizes latest microprocessor tech­
nology.

Same labor implications as NC
but, unlike NC, may require
com puter personnel. Saves time
in reprogram m ing to remove
errors or make design changes.
Requires maintenance personnel
with electronic skills.

Machining center

An autom atic tool changer makes the center a multifunc­
tion NC machine. Each center is equivalent to several ma­
chines, each having a specific function.

Raises productivity by perm it­ Accounts for a small but grow­
ting operations on many surfaces ing percentage of machine tools
of a part in a single setup. O pera­ in larger plants, but a dispropor­
tor may control several ma­ tionately large share of the in­
chines.
dustry’s output.

Adaptive control

A utomatically controls feed rate to reduce or eliminate such
factors as vibration, tool wear, and cutting temperatures,
and alerts operator. Can be used with conventional tools or
with NC.

Raises productivity in m achin­
ing through substitution of sen­
sors for workers’ own percep­
tions. Reduces skill require­
ments.

Used by large plants. Utilization
by small shops will depend upon
development of improved sen­
sors and their cost effectiveness
relative to the availability of
skilled workers and also the type
of work performed.

Computer-aided design/com ­
puter-aided manufacture
(CAD/CAM)

C om puters are used to develop designs for products to be
m anufactured (CAD). CAM directs numerically controlled
machines and autom atically guides workpieces among
machines on com puter-controlled handling systems.

Reduces need for low-skilled op­
erators; increases requirements
for higher skilled workers.

Used by large machine tool m an­
ufacturers only; diffusion to
medium-size firms will be severe­
ly limited by the technology’s
cost.

Digital readout (DRO)

A device is applied to movable portion of a machine tool
to measure its actual movement; can provide some a u to ­
matic control; measurement appears on a display unit.

O perator efficiency and accuracy
are enhanced during the position­
ing phase of the machine cycle.
O perators are trained in less time
and fatigue is reduced.

Use is limited but increasing
am ong machine tool builders; al­
ready widespread am ong tooland-die shops. Producers of
DRO's expect a 25-percent an­
nual grow th in their sales to the
m etalworking machinery indus­
try in the next several years.

M anual-data-input control
(MD1)

Enables an operator to change the position of a machine
autom atically; also identified as “operator-program m ed
NC.”

M achinist can plan and enter
part programs; possible in “shop
language." Training period short­
er than for NC programm ing.

Precise data on utilization are
unavailable, but its use is spread­
ing am ong machine tool builders
and even more rapidly in the
contract tool-and-die shops.

Cutting-tool materials

D urable new materials, such as coated carbides, polycrys­
talline diam onds, and special ceramics more efficiently
meet continued increases in machining speed.

Reduce
labor
requirements
somewhat because tools do not
have to be changed as often.

Tungsten carbide expected to re­
main the m ajor m aterial, but
coated carbides may increase
from current 15 percent to 25
percent of all cutting-tool ma­
terials in m etalworking machin­
ery in 1985.

G roup technology (GT)

M anagement skills used to reduce small batch operations.
Involves the grouping of parts on the basis of similar shapes
a n d /o r processing requirements. W orkers may perform a
wide range of tasks.
|

Improves efficiency and quality
of output. Workers may broaden
skills and replace narrow spe­
cializations.

Used by some large machine tool
builders; elements of GT likely
to spread slowly to smaller build­
ers and tool-and-die shops.




181

Output and Productivity Outlook
©ytpyt
The metalworking industry has undergone sharp
fluctuations in output in response to the business cycle,
typical of the experience of other capital goods industries.
Although reliable output data are not available for the
total metalworking machinery industry, Census value of
shipments data adjusted for price changes (used as a crude
measure of output) suggest an average annual growth rate
of about 2 percent in 1960-78. Output grew very rapidly
(8-9 percent a year) in the period 1960-67. However, in
1967-73, average output declined by about 2 percent and,
in 1973-78, the average rate of decline was still about 1
percent, reflecting the 1970 and 1975 recessions.
Although later data are not available, there is evidence
that some sectors of the industry are recovering in
response to the large tooling requirements of the
automobile, aircraft, and defense industries.
While all sectors had impressive rates of growth in
1960-67, the pattern of the sectors varied considerably in
the next decade.9 Metal-cutting tool output advanced at a
double-digit rate annually (12.0 percent) in 1960-67, and
then experienced a very steep decline (8.9 percent
annually) in 1967-73. However, in 1973-80 a moderate
rate of recovery (2.7 percent) occurred, as output grew
strongly after the 1975-76 recession. A similar rise and
fall, but not as pronounced, occurred in metal-forming
output in 1960-67 and 1967-73. However, in the
succeeding period, 1973-80, metal-forming output
declined—more steeply than in 1967-73. As in the
recession years of 1975 and 1976, output again fell sharply
in 1980. Nevertheless, in certain sectors, e.g., metal
cutting, delivery time for some orders was as much as 2
years.
One explanation for the decline in metal-cutting output
over the longer period of 1967-79 is the tardiness of
customer industries in buying new tools. The aging of the
machines in the U.S. economy has been a long-term
problem. The percentage of metal-cutting machines that
are less than 10 years old has been declining steadily since
the end of World War II; an estimate of 31 percent for
1976-78 approaches the rate at the end of the depression
in 1940. Moreover, when all U.S. metal-cutting and
metal-forming tools are combined and compared with
tools in six other major industrial nations, the United
States had the smallest percentage of machines less than
10 years old and the highest percentage of machines over
20 years old.1
0
’Output data for the metal-cutting (sic 3541) and metal-forming
sectors (sic 3542) are weighted output measures developed by b l s for
1958 to 1980. The data for all other sectors, as well as the entire industry
(sic 354), are deflated Census value of shipments data; latest year, 1978.
l0“The 32th American Machinist Inventory of Metalworking
Equipment 1976-78,” American Machinist, op. cit., pp. 133, 135,
and 137. The data for the six foreign industrial nations are based
on the most recent studies in each country, ranging from 1973 to
1978.




Of major importance is the change in the mix of output
as the proportion of numerically controlled tools has
increased. It is estimated that, in 1973, about 40 percent of
the value of shipments of NC milling machines and
machining centers plus comparable conventional machine
tools was accounted for by NC machines; the percentage
rose to 66 percent by 1977, according to estimates of a
consultant to the industry. In the case of NC lathes and
conventional lathes, the estimated rise in NC shipments
was from 32 percent of the total value in 1973 to 52
percent in 1977. Because NC is considerably more
productive, fewer machines are required for a given
amount of production.
To some extent, output growth in the industry has been
held back by the lack of skilled workers to accommodate
to periods of higher demand. While the industry has
undertaken considerable worker training in strong
growth periods, numerous trained skilled workers had to
be laid off when slumps in production took place. Some
of these workers left the industry for more stable jobs in
other industries; this was especially the case for workers
who experienced more than one long spell of unem­
ployment.
■Foreign trade. Traditionally, the United States has
enjoyed a comfortable advantage in the export of metal­
cutting machines. In 1958-65, the value of exports was at
least three times as large as the value of imports. Although
a trade advantage was maintained in the years 1966-76, it
was no longer as large as in earlier years, and in only 2
years (1970 and 1971) was the trade advantage as high as 2
to 1.
in 1977, imports of metal-cutting machines exceeded
exports for the first time. Although exports of cutting
machines continued to average more in the 1970’s than in
the 1960’s (83.5 percent higher), the rise in imports was
considerably more rapid (4.2 times as high in the 1970’s).
In 1978, the continued deterioration in the trade position
of metal-cutting machines- offset the favorable trade
balance of the metal-forming sector for the first time, and
the trade imbalance rose even higher in 1979.
U.S. machine tool manufacturers have not been
competing successfully with foreign producers in
domestic markets, as is evident by the increase in the
import penetration ratio, that is, imports as a percent of
apparent consumption (imports plus domestic production,
excluding exports). This ratio more than doubled
between 1972 and 1979, rising from 9 percent to 22
percent, with the biggest jump in 1978. It rose again in
1980. U.S. manufacturers may be unable to recapture a
large portion of the imports of conventional machine
tools from Japan and Taiwan. As a result, U.S. firms will
have to become more competitive with Japan and also
West Germany in the production and sale of NC
machines, machining centers, and specialized machines.
In addition to the very rapid expansion in imports, the
unfavorable balance of trade in metal cutting was also
affected by the way in which foreign trading relationships
182

in these machines are established. During several
prosperous years in the 1960’s, machine tool builders did
not have much incentive to expand their exports to new
customers. Yet, machine exports depend upon the
establishment of a long-term relationship. Machine
buyers often rely upon the machine makers to service and
ultimately even rebuild their machines after several years
of usage. In the absence of more commercial ties abroad,
U.S. machine tool firms could not take advantage of their
excess capacity during the recessions of the 1970’s to
increase their exports. More recently, the sizable upswing
in domestic orders has again been reducing interest in
exports, as numerous firms are operating at or near
capacity levels, and skilled workers are in short supply.
Prospective exporters are also handicapped because their
lead times for delivery of new machines are relatively
longer than in some other countries.
Productivity
Although a reliable measure of productivity is unavail­
able for the metalworking machinery industry as a whole,
trends can be estimated from available output and
employee-hours data. Deflated value of shipments, used
as a crude measure of output, increased by about 2 per­
cent annually in 1960-78, while the rate of change in
employee hours was less than 1 percent for the same
period (chart 6). These data suggest that the productivity
growth rate averaged only 1-2 percent annually in the
1960-78 period. While productivity (estimated as above)
grew moderately from 1960-67, it rose more slowly in
1967-73 and then edged down in 3973-78.
In the metal-cutting sector, for which a BLS weighted
measure is available, productivity growth averaged only
1.2 percent in the 1960-80 period, approximately the rate
of growth for the whole industry. While productivity grew
at an average of 4.1 percent annually from 1960 to 1967, it
only edged up at less than 1 percent from 1967 to
1973, when output dropped about 9 percent. In the
succeeding 7 years, 1973-80, productivity showed no
growth. While growth was relatively strong after 1976, it
just offset the recession declines.
Traditionally in capital goods manufacturing, which is
highly cyclical, there is a lag in the adjustment of hours to
output changes. In this industry, there is often consider­
able reliance on changes in the overtime component of
employee hours in order to adapt to the wide shifts in
output. This is especially evident in the metal-cutting
sector. Metal-cutting production has characteristics that
are not present to the same degree in other industries,
particularly labor intensity, high skill needs, and the
considerable time and cost to train workers. Moreover,
there is a shortage of skilled workers. Consequently,
during an upturn, overtime hours are increased; in a
downturn, firms keep as many of their employees on the
payroll for as long a period as they regard practicable, but
reduce overtime.
In contrast, the products of the accessory sector, which
include perishable tools and various attachments and




383

accessories for machine tools and other metalworking
machinery, enable these manufacturers to benefit from a
production process of typically shorter time frames and
larger batches than in machine tools. The accessories
firms stock numerous standardized products to accom­
modate anticipated industry demand. Moreover, while
the accessory sector needs skilled instrumentmakers to
produce measuring devices, the bulk of its output requires
a relatively less skilled work force than does metal cutting.
Therefore a downturn in the accessory sector output is
more likely to be matched by a comparable decrease in
employment than in the metal-cutting sector.
The machine tool builders have less flexibility in the
adjustment of employment. This can be attributed to the
length of time required to complete orders for metal­
cutting machinery and the limited standardization that is
possible in its production. Such machinery is usually only
manufactured when purchase orders are in hand and
requires several months to make. The special machines
that are manufactured in this sector are necessarily made
in small production runs. Even many of the so-called
universal machines are in some way modified to meet
individual buyer specifications, thereby ruling out some
of the economies associated with longer production runs.
The industry has made some attempt to overcome the
obstacle of small batch production to improve
productivity in the machine tool sectors. For instance,
construction standards have been developed over a 20year period by NC committees of the Electronic Industries
Association to enable greater output of standard
components and interchangeable subassemblies. In
addition, standards have been revised to accommodate
advances in technology, such as the capability of
computer NC systems to handle manual data input.
Aging machinery is a factor which limits productivity
growth in firms throughout the metalworking machinery
industry. It has been stated that many of the rather old
manually operated tools in use actually cut metal for
much less than 10 percent of the time a workpiece is in a
batch production shop. Considerable time is involved in
setting up to make a part, or parts are being loaded or
unloaded, or tools are being changed. As noted in the next
section, an upswing in investment in new plant and
equipment was underway in the metal-cutting sector.
Future data may better reflect the installation of newer,
more productive machines, because, in general, optimiza­
tion of new plant and equipment can be a lengthy process.

investment
Capital ©Kp©ndstyr®s
Real capital expenditures1 by the metalworking ma­
1
chinery industry in 1978 were less than 80 percent of peak
levels in 1966, although in current dollars they rose by
nearly two-thirds. The cyclical volatility of capital outlays
"Deflated
chinery.

by implicit

price deflator for metalworking ma­

in this industry is pronounced. Expenditures (in constant
dollars) rose almost steadily to their highest level in 1966,
moved down rapidly in 1970, and plummeted in 1971 to
reach the lowest level since 1963. After rising to near peak
levels in 1974, expenditures fell sharply again in 1975.
Although they recovered in succeeding years, the peak
outlays of 1966 were not surpassed.
Approximately the same pattern is reflected in the three
major sectors of the metalworking machinery industry. In
metal-cutting machinery, peak expenditures (real) oc­
curred in 1967. In 1978, over a decade later, real expendi­
tures were less than 65 percent of the peak. Machine tool
accessories and measuring devices attained their summit
in real capital expenditures in 1966-67; by 1978 they were
only about 85 percent of the level for those 2 years.
The third major sector—special dies and tools, die sets,
jigs and fixtures, and industrial molds—which typically
accounts for well over one-third of capital expenditures in
metalworking machinery, surpassed its 1966 peak in
expenditures in 1974. More pronounced, however, than
in the other sectors, was the decline in outlays in 1975,
to half of those of the previous year. By 1978, real outlays
were about 80 percent of the peak.
No more recent Census data are available, but,
according to industry reports, outlays by the machine tool
builders were rising sharply in response to increasing
demand, noted earlier, by the automobile, aircraft, and




184

defense industries. Capacity of this sector may be greatly
enlarged in the early 1980’s. The more productive,
automated machinery being installed may improve the
industry’s competitive position vis-a-vis foreign imports.
Research m d development
A few of the relatively large machine tool builders have
undertaken research as well as development. However, in
general, most of these firms are in development only. For
example, large machine tool makers are involved in the
development of improved computer controls for machine
tools. To some extent, large companies outside the
industry have undertaken R&D at least in part to improve
machining standards in their own work. For instance, a
firm pioneered in the recent development of a cutting tool
with polycrystalline diamond material.
Some joint interindustry development has taken place,
including the pooling of development costs by several
machine tool builders and an automobile manufacturer in
order to increase the speed of lathes. However, there is
currently relatively little in the way of joint efforts by
government, industry, and labor compared with those in
some countries, notably Japan.1
2
1Comptroller General of the United States, Manufacturing
Technology—A Changing Challenge to Improved Productivity,
June 3,1976, pp. 74, 88 89.

growth through 1967 when peak levels were attained.
Subsequently, the industry reflected the economy’s
recessions with deep employment declines in 1971 and
again, but not as steeply, in 1975 and 1976. Since then,
employment has moved up sharply.
As mentioned earlier, overtime hours play an impor­
tant role in this industry. By expanding and contracting
overtime hours of production workers in response to
changes in output, employers tend to moderate short­
term hirings and layoffs. While this is true for the entire
metalworking machinery industry, it is more marked in
the metal-cutting sector. Overtime hours in metal cutting
during the sector’s cyclical peaks and troughs of 1960-61,
1969-70, and 1973-75 ranged, respectively, from 5.5 to
1.9 hours, 6.5 to 1.8 hours, and 7.8 to 1.8 hours (average
weekly data). The proportion of production workers to
all employees in the industry has not changed significant­
ly in the last two decades. Production workers accounted
for 75 percent of all employees in 1960 and 73 percent in
1980. The comparable figures for all durable goods manu­
facturing industries were 74 percent and 69 percent.
The three largest industrial sectors in metalworking
machinery—namely, tool and die, metal cutting, and
machine tool accessories—accounted for slightly over
three-fourths of the industry’s total employment in the
years 1960 through 1980, but the metal-cutting sector
declined in relative importance. They all exhibited growth
from 1960 to 1967, but the tool and die and machine tool
accessory sectors grew at faster average annual rates than
did metal cutting. In spite of its recent sharp rise, employ­
ment in metal cutting has not recovered fully from its low
levels of 1971 and 1972, while the other two sectors re­
covered more rapidly and attained their peaks in 1979 and
1980 (chart 7). Tool and die and machine tool accessories
firms increased their share of employment within the
metalworking industry from 50 percent in 1960 to 55
percent in 1980. The share of metal-cutting employment
declined from 27 percent in 1960 to 21.5 percent in 1980.

Considerable RAD effort has been directed toward
standardization. Whereas once there were more than 30
adapters in use by machine tool builders, the industry is
moving toward more universal use of an adapter
developed by a large machinery manufacturer. The
adapter makes possible substantial reduction in cuttingtool inventories and is considered particularly useful for
small firms with NC machines.
The National Bureau of Standards is also advancing
the pace of innovation within the industry. The Bureau
has research underway to improve the ability of NC and
industrial robots to work together. The Bureau is also
trying to improve the adaptability of robots. According to
a Bureau research analyst, the largest share of robot
applications during the 1980’s will be in loading-and
unloading machine tools.

Employment and ©©(g&jp®ts®G Trsondli
ii!!
The industry’s employment increased at an average
annual rate of 1.4 percent in 1960-80, about the same as
for all durable goods manufacturing industries. While
employment growth averaged 5.1 percent in 1960-67, it
declined 3.0 percent annually during 1967-73, but rose
again at 2.3 percent annually in 1973-80. For the period
1980-90, three employment projections by the Bureau of
Labor Statistics, based on alternative versions of
economic growth, fall in the range of 0.8 percent annually
(only about half the 1960-80 rate) to 3.8 percent (296 times
the 1960-80 rate).1 The low-trend estimate for
3
metalworking machinery is only half the growth rate
expected for all durable goods by 1990, while the hightrend estimate (Level B on chart 7) exceeds the projected
growth rate for durable goods manufacturing.
At 371,500 persons, employment in 1980 was exceeded
only during 2 World War II years. After dropping to a
postwar trough in 1949, employment rose steadily to
314,000 in 1953, and was not surpassed until the second
half of the 1960’s. Over the years 1960-80, there were
three periods of sharp cyclical fluctuations affecting every
major sector of the metalworking industry. Employment
hit a low in 1961 and then exhibited strong continuous

©©gypiitteois
B l s projects an employment increase from 1978 to 1990
for all but the smallest occupational group (sales workers)
in the metalworking machinery industry. Craft workers
and operatives, the two largest of the blue-collar groups,
each constituted nearly one-third of all employees in
metalworking in 1978. While operatives are expected to
grow 36 percent by 1990, the increase for craft workers is
expected to be about half that rate (chart 8). By 1990,
operatives will account for a somewhat larger percent of
total employment than in 1978, while the proportion of
craft workers is expected to decline slightly.
A major influence on occupational skills and
responsibilities in the past decade has been the use of
numerically controlled machines. For example, the more
rapid growth in employment of operatives is at least
partly attributable to the recent expansion of NC

'^Projections for industry employment'in 1990 are based on three
alternative versions of economic growth for the overall economy
developed by bls. The low-trend version is based on a view of the
economy marked by a decline in the rate of expansion of the labor force,
continued high inflation, moderate productivity gains, and modest
increases in real output and employment, in the high-trend version I, the
economy is buoyed by higher labor force growth, much lower
unemployment rates, higher production, and greater improvements in
prices and productivity. The high-trend version II is characterized by the
high gnp growth of high-trend I, but assumes the same labor force as the
low trend. Productivity gains are quite substantial in this alternative. On
chart 7, Level A is the low-trend, Level B is high-trend 1, and Level C is
high-trend II. Greater detail on assumptions is available in the August
1981 issue of the Monthly Labor Review.




185

Chart 7. Employment in metalworking machinery and selected
industry sectors, 1960-80, and projections for 1980-90
Employees (thousands)

560

Note: See text footnote 13 for explanation of alternative projections.
Source: Bureau of Labor Statistics.




186

Chart 8. Projected changes in employment in metalworking
machinery by occupational group, 1978-90

Occupational
group

Percent of
industry
employment
in 1978

Professional and
technical workers

20

8.4

Sales workers

10

9.0

Managers, officials,
and proprietors

Percent change

1.9

Clerical workers

12.1

Craft workers

32.3

Operatives

32.5

Service workers

1.8

Laborers

2.0

Source: Bureau of Labor Statistics.

machines and various intermediate technologies. A
shortage of skilled workers is a major reason certain firms
have turned to NC. Over 800 member firms of the National
Tooling and Machining Association indicated that they
needed on average a 26-percent increase in skilled
toolmakers and machinists.1 Similarly, the great
4
majority of nonelectrical machinery manufacturing
plants responding to an Industry Week survey reported
shortages of skilled workers; the most pressing needs were
for machine operators, mechanics, electricians, and tooland-die makers.1
5
A study which compared the skills of machinists on NC
with those on manually operated machine tools revealed
that NC machines are associated with a decline in demand
for motor skills and decisionmaking abilities.1 According
6
to a BLS study, NC operators need less knowledge because
tapes are programmed to control speed, feed, and width

u National Tooling and Machining Association, Record, Vol. 2,
No. 5, May 1979, p. 4.
l5Daniel D. Cook and John S. McClenahen, “Skilled Worker
Nears Extinction,” Industry Week, Aug. 29, 1977, p. 46; also see
Michael Marley, “If They Could Clone Skilled Workers,” Iron Age,
Vol. 221, No. 37, Sept. 11, 1978, pp. 36-38.
I6R.J. Hazlehurst, R.J. Bradbury, and E.N. Corlett, “A Comparison
of the Skills of Machinists on Numerically Controlled and
Conventional Machines,” Occupational Psychology, Vol. 43, Nos. 3
and 4, 1969, p. 177.




187

and depth of cut.1 At the same time, the study referred to
7
the need for greater conceptual skills on NC machines.
There will, however, be continued need for highly
skilled machine operators on the most advanced NC
machines. Moreover, the costliness of NC machines and
the intricacy of their control systems increase the need for
preventive maintenance mechanics trained in electronics
with practical knowledge of hydraulics and pneumatics.
Bls projects that employment of mechanics, repairers,
and installers, a subdivision of the craft worker group,
will expand five times as fast as all craft employment.
While employment of professional and technical
workers is projected to grow by 22 percent from 1978 to
1990, the projected increase for managers, officials, and
proprietors will be only at one-fourth that rate. The
former’s share of total industry employment will be
virtually unchanged by 1990, while the latter’s share will
decline. Changed skills and responsibilities, also largely
related to NC equipment, are occurring for these
occupational groups. The competitive structure of the
industry and complexity of NC equipment and other
technologies require not only knowledge of the new
machines but also the capability to organize a shop’s
production so that the machines are utilized optimally,

17Technological Change and Manpower Trends in Five Indus­
tries, Bulletin 1856 (Bureau of Labor Statistics, 1975), p. 42.

Engineers will remain the dominant occupation for the
professional and technical worker group in 1990, with
about half the engineers still in the mechanical field.
Drafters will remain, by far, the largest single occupation
in the technician group. While computer specialists are
expected to increase at only half the rate for total industry
employment, the number of numerical tool programmers
will more than double by 1990, but will still account for
less than one-half of 1 percent of the industry’s employees.
The programmer position on advanced NC tools requires
mathematics, the ability to visualize objects and motions
in dimensions, and an understanding of cutting and
tooling principles.
Adjustment © workers t® tediinologicali sHissiig®
If

Programs to protect the worker from the adverse affects
of changes in machinery and methods may be incorpora­
ted into union contracts or they may be informal arrange­
ments between workers and management. In general,
such programs are more prevalent and detailed in formal
contracts. Both formal and informal labor-management
arrangements are influenced by the state of the economy
and the availability of labor.
Training may be the major factor in the adjustment of
workers to technological change in this industry. Officers
of leading machine tool manufacturing firms refer to
shortages of trained machinists and other technical and
skilled workers as a principal obstacle to maintaining
high levels of production or increasing them.
Provision for an adequate level of training is
complicated by demographic factors. An aging work
force is making it steadily harder to maintain a nucleus of
skilled workers as many, including tool-and-die makers,
continue to retire. To cover the skill shortages, NC tool
builders have provided short, intensive training program s
in the fundamentals of maintenance to electricians and
other skilled workers. Machinists and even experienced
machine operators are being trained in programming.
The extent of training provided by employers is not
precisely known. In some cases, small firms have taken a
multiemployer approach to apprenticeship and other
qualifying training. While some form of training is
provided by most employers, local union bargaining
agreements typically do not refer specifically to training.
A BLS survey of structured training in the nonelectrical
machinery industries disclosed that only 18 percent of the
surveyed establishments provided training in one or more
skilled occupations; and only about one-third of the
training was for skill improvement, while two-thirds was
training to qualify for the job. The survey excluded the
most common forms of training, namely, learning
through experience and informal training. The survey
included metalworking machinery establishments but
data for them were not available separately.
Since 1966, the U.S. Department of Labor (D O L) has
provided training funds which have been distributed by



188

the National Machine Tool Builders Association. As of
fiscal year 1979, trainees hired by the machine tool
manufacturers for the DOL program have to be
economically disadvantaged persons. The training
(including classroom instruction) is conducted on the job
site. Typically, the training is in such fields as machine
operation, assembly, and machine repair, and the
programs run 13 to 16 weeks. More than 14,000 graduates
have proven a good screening source for apprentices who
can later qualify for more skilled, higher paying jobs
requiring further training.
The National Tooling and Machining Association has
enrolled over 15,000 persons in their preemployment
training program funded by the DOL. The program, which
previously consisted of 16 weeks of institutional training
followed by 36 weeks of on-the-job training, now is a 12week program of institutional training only for
economically disadvantaged persons. An industry
spokesman believes that this program alone is not
providing a sufficient number of persons who are
qualified for further training in more highly skilled
occupations.
Since skilled workers are in short supply, some firms
have sought foreign workers. However, the firms do this
reluctantly because of the time involved in completing
paperwork and securing approval for immigration. A
survey in Milwaukee disclosed that the careers of
machinist or machine operator ranked rather low with
high school students, even though the city is a major
machine tool producer.1 However, some metalworking
8
firms are making greater efforts to attract young people to
the industry by enrolling high school students in
cooperative programs (involving morning school attend­
ance and afternoon work), much as they have done
successfully with engineering personnel.
The International Association of Machinists and
Aerospace Workers (1AM) is the major union in this
industry. The United Automobile Workers (UAW ) and the
United Steelworkers of America (U SA ) are the other two
leading unions. Overall, these unions plus several others
have organized about one-third of the workers in the
industry.
Contract provisions for nine metalworking firms
studied by BLS which each employ at least 1,000 workers
appear to be representative of the bargaining agreements
negotiated by the three leading unions. In general, the
prevalence of seniority provisions acts as a measure of job
security when technological change takes place. Agree­
ments provide for seniority rights in the event of layoff
and for purposes of rehiring. Interplant transfers are quite
uncommon. A provision requiring advance notice of
layoff is present in a majority of the agreements studied,
IS
J.G. Udell and others, Skilled Labor in the Milwaukee Area: The
Supply, Education, Problems and Opportunities, Wisconsin Economy
Study No. 15 (Madison, University of Wisconsin, Graduate School of
Business, July 1977).

but such notices are generally unrelated to technological
change. Nevertheless, some agreements specifically refer
to “new equipment” and most u a w agreements deal with
“new jobs.” Companies in these instances are normally
required to consult with the union regarding changes in
job description or occupational assignment of the job,
and provisions exist for resolving grievances.
Considerable effort by management to improve job
security is related to the shortage of skilled workers. The

problem is complicated by the cyclical nature of the
industry. A Connecticut machine tool builder visited by
BLS made the following arrangement: During a 9-month
slack period, the firm employed its work force for 3-week
periods, and unemployment insurance (ui) payments
were secured for the fourth week of each month. (In
Connecticut, no waiting period is required for Ul
payments.)

SELECTED REFERENCES
Ashburn, Anderson. “The 1980 Machine-Tool Standings,” American
Machinist, February 1981, p. 93.

National Center for Productivity and Quality of Working Life. New
Technologies and Training in Metalworking, Washington, U.S. Gov­
ernment Printing Office, 1978.

Ashburn, Anderson, and others. “The Machine Tool Task Force Re­
ports on Metalcutting-Machine-Tool Technology,” American Ma­
chinist, October 1980.

National Machine Tool Builders’ Association. 1980-1981 Economic
Handbook o f the Machine Tool Industry. McLean, Virginia, 1980.

Bellows, Guy. Nontraditional Machining Guide—26 Newcomers for
Production. Cincinnati, Metcut Research Associates, Inc., 1976.

Putnam, George P. “Why More NC Isn’t Being Used,” Machine and
Tool Blue Book, September 1978, pp. 98-107.

Beman, Lewis, and Steven E. Prokesch, “Foreign Competition Stirs
U.S. Toolmakers,” Business Week, Sept. 1, 1980, pp. 68-70.

Schaffer, George. “Digital Readout Systems,” American Machinist,
May 1979, pp. SR-2-4.

Comptroller General of the United States. Manufacturing Technolo­
gy—A Changing Challenge to Improved Productivity, Report to the
Congress, lcd-75-436, June 3, 1976.

Smith, Donald N., and Lary Evans. Management Standards for Com­
puter and Numerical Controls. Ann Arbor, University of Michigan,
1977.

Cook, Daniel D., and John S. McClenahen. “Skilled Worker Nears
Extinction,” Industry Week, Aug. 29, 1977, pp. 38-48.
Dallas, Daniel B. “Machining Outlook for 1978,” Manufacturing Engi­
neering. January 1978, pp. 46-50.

“The 12th American Machinist Inventory of Metalworking Equipment
1976-78,” American Machinist, December 1978, pp. 133-48.

Gettelman, Ken. “Numerical Control’s Tech Explosion,” Modern
Machine Shop, July 1979, pp. 79-88.

Udell, J.G.,and others. Skilled Labor in the Milwaukee Area: The Sup­
ply, Education, Problems and Opportunities. Wisconsin Economy
Study No. 15, Madison, University of Wisconsin, July 1977,

Golembe, Stanley, “Application, Justification and Selection of Digital
Readouts,” Modern Machine Shop, May 1977, pp. 88-96.

U.S. Department of Labor, Bureau of Labor Statistics. Outlook for
Numerical Control o f Machine Tools, by John Macut. Bulletin 1437,
March 1965.

Hatschek, R.L. “Manual-Data-lnput Controls,” American Machinist,
May 1978, pp. SR -18-SR-19.
Macut, John. “New Technology in Metalworking,” Occupational Out­
look Quarterly, February 1965.
Mansfield, Edwin, and others. Research and Innovation in the Modern
Corporation. New York, Norton and Company, Inc., 1971, pp. 186—
205.




189

U.S. Department of Labor, Bureau of Labor Statistics and Employ­
ment and Training Administration. Occupational Training in Se­
lected Metalworking Industries, 1974. B ls Bulletin 1976, ETA R&D
Monograph 53, 1977.

Technology and Labor in
Motor Vehicles and
Equipment
Robert V. Critchlow

S u m m a ry

managers, sales workers, and semiskilled operatives will in-,
crease while declines are expected ip the other major occu­
pational groups. Although new technology will reduce unit,
labor requirements in some operations, industry growth will
result in higher long-term employment levels for computer
specialists, assemblers, and others who work with new tech­
nology. Semiskilled workers will continue to constitute the
largest occupational category. These workers are engaged in
production operations which generally are the most labor
intensive and have potential for further technological
change.

New equipment and manufacturing methods are ex­
pected to continue to be introduced in the motor vehicle
and equipment industry. Specific innovations which may be
applied more widely include electronic computers, im­
proved equipment for automatic assembly, use of plastic
and powdered metal materials, numerical control, and im­
proved transfer lines. New technology in some instances is
expected to improve quality and achieve productivity gains.
A total of $2.1 billion was spent by the motor vehicle
industry for new plant and equipment in 1975, and an
estimated $2.4 billion was spent in 1976. These amounts
are about three times as much as the 1960 expenditure of
$790 million, although the increase would be less in real
terms due to increases in plant and equipment prices over
the period. The average annual rate of increase in spending
was lower during 1967-75 than during the 1960-67 period.
Capital expenditures are expected to increase considerably
over the next several years in order to produce cars that can
meet Federal Government standards for safety, exhaust pol­
lution levels, and fuel economy.
Output per employee-hour in the motor vehicle and
equipment industry (BLS data) increased at an annual rate
of 3.2 percent between 1960 and 1975. The productivity
growth rate in the motor vehicle industry was 3.6 percent
annually during 1960-67, slightly above the 3.2-percent
average annual rate achieved during 1967-75. Growth in
output per employee-hour was particularly strong during
1971 and 1972 as output rose sharply from the 197f)
strike-year level in response to very strong demand for cars
and trucks. Further productivity growth occurred in 1975
when employment fell more rapidly than output. Produc­
tivity gains in assembly, machining, and other production
operations are expected as new technology is introduced.
Industry employment rose from 724,000 in 1960 to a
peak of 955,300 in 1973, then dropped during the eco­
nomic downturn of late 1974-75 to 774,100 in 1975.1 As
sales and production improved in 1976, employment rose
to 850,600. BLS projections indicate that employment may
decline to 808,000 by 1985.
Technological and other changes will continue to alter
the structure of occupations in this industry. Demand for

Technology in the 1®7®?
s
Technological changes in the motor vehicle and equip­
ment industry are underway in major phases of production,
with productivity gains and laborsavings anticipated. These
changes include more extensive use of electronic com­
puters, improved equipment for automatic assembly and
inspection, more widespread use of plastics and other light­
weight materials, more widespread application of numerical
control, and improvements in transfer lines. (See table
5.) Modifications of automobile engines also are underway
to meet stricter emission standards and to raise fuel econ­
omy.
Electronic computersComputers are a jkey technology in the automobile in­
dustry, initially applied to business operations such as pay­
roll and bookkeeping records and subsequently extended to
an increasing number of research and production opera­
tions. According to International Data Corporation, more
than 400 computers are in use in the industry, with fur­
ther growth in computer use expected. Examples ,qf com­
puter applications gaining prominence, and their labor im­
plications, are presented below.
Auto styling and design. Mathematical information, repre­
senting automobile body surfaces can be store4 in %com­
puter memory system. A designer, working with a graphic
display terminal, can use this information to design auto

Reprinted from BLS Bulletin 1961 (1977),

Technological Change and its Labor Impact in Five Industries.



190

TabS® 5.

Major technology changes in the motor vehicle and equipment

industry

D e s c rip tio n

L a b o r im p lic a tio n s

D if fu s io n

E le c tr o n ic c o m p u te rs

T h e use o f g ra p h ic d is p la y t e r ­
m in a ls can in te g ra te and speed
w o r k f lo w
b e tw e e n d e sig n ,
to o lin g , a nd p r o d u c tio n . T im e
re q u ire m e n ts f o r R & D w o r k
are lo w e re d . N u m e ro u s a p p li­
c a tio n s in q u a lit y c o n t r o l in ­
crease p r o d u c t iv it y o f in s p e c ­
tio n p e rs o n n e l. C o m p u te r c o n ­
t r o l o f m a c h in in g a nd assem ­
b ly o p e ra tio n s m a y incre a se
p r o d u c tio n ra te s a nd re d u c e
la b o r re q u ire m e n ts .

E m p lo y m e n t increases in c o m ­
p u te r -r e la te d o c c u p a tio n s su c h as
s y s te m s a n a ly s ts , p ro g ra m m e rs ,
a nd p e rip h e ra l e q u ip m e n t o p e ra ­
to rs .
D e c lin e s
e x p e c te d
fo r
d r a fte r s a n d k e y p u n c h o p e ra to rs .

M o re th a n 4 0 0 c o m p u te r s are e s ti­
m a te d to be in use. C o n tin u e d
g r o w th e x p e c te d in th e n u m b e r o f
c o m p u te r s a n d ty p e s o f a p p lic a ­
tio n s .

M a c h in e a s s e m b ly o p e ra tio n s
(a u to m a te d a s s e m b ly lin e s)

A u to m a te d a s s e m b ly a p p lic a ­
tio n s range fr o m tig h te n in g
b o lts to w e ld in g ca r b o d ie s t o ­
g e th e r . A u to m a tic a sse m b ly
s ta tio n s are f r e q u e n tly in t e r ­
m ix e d w it h m a n u a l s ta tio n s ,
d e p e n d in g u p o n th e n a tu re o f
th e jo b .

R e d u c e d la b o r r e q u ire m e n ts in
s e m is k ille d a s s e m b ly o p e r a tio n s ,
a nd in c re a s e d need f o r m a c h in e
m a in te n a n c e p e rs o n n e l.

M a c h in e a s s e m b ly has e x p e rie n c e d
c o n s id e ra b le d e v e lo p m e n t o v e r th e
p a s t d e c a d e , a n d is e x p e c te d to c o n ­
tin u e to g r o w in use.

N e w m a te ria ls

P la stic m a te ria ls o f f e r a d v a n ­
tages o v e r ste e l a n d ca st m e ta l
m a te r ia ls in w e ig h t savings
a n d , o fte n , fe w e r p ro ce s s in g
o p e ra tio n s . P a rts can be f a b r i ­
c a te d fr o m m e ta l p o w d e r in to
c o m p le x
s h a p e s and w it h
fe w e r m a c h in in g o p e ra tio n s .
M o re w id e s p re a d use o f a lu m i­
n u m a nd sp e cia l steels also is
a n tic ip a te d .

S o m e re d u c tio n
in s e m is k ille d
s h e e t-m e ta l w o r k e r s a nd m a c h in e
t o o l o p e ra to rs .

Use o f p la s tic s a nd a lu m in u m
c o n tin u e to g ro w .

N u m e ric a l c o n t r o l

A u to m a tic o p e r a tio n o f m a ­
c h in e to o ls b y e le c tr o n ic c o n ­
t r o l d e vice s a n d c o d e d ta p e in ­
s tru c tio n s can re d u c e m a c h in ­
in g tim e and la b o r costs.

D e c lin e in th e n u m b e r o f m a c h in e
t o o l o p e ra to rs n e e d e d , a nd p o s s i­
b ly s o m e incre a se in m a c h in e
m a in te n a n c e p e rs o n n e l.

N u m e ri'c a l c o n t r o l in lim ite d use a t
p re s e n t. A p p lic a tio n s e x p e c te d to
g ro w in fu t u r e , w it h e m p h a s is o n
n e w s o lid - s ta te p ro g r a m m a b le c o n ­
tr o lle r s a n d d ir e c t c o m p u te r c o n ­
t r o l.

Im p r o v e d tra n s fe r lin e s

T ra n s fe r lin e p r o d u c t iv it y and
f l e x ib i li t y have been incre a se d
b y th e in t r o d u c t io n o f m u lt i- ;
p u rp o s e
m a c h in e s ,
in t e r ­
ch a n g e a b le m a c h in e m o d u le s ,
sto ra g e b a n k s f o r p a rts a t in ­
te rv a ls in th e m a c h in e lin e ,
and an incre a se in th e n u m b e r
o f a u to m a tic o p e ra tio n s .

R e d u c tio n in th e n u m b e r o f m a ­
c h in e to o l o p e ra to rs a n d in s p e c ­
to rs .

E q u ip m e n t d e s ig n e d t o
in cre a se
tr a n s fe r m a c h in e f l e x i b i l i t y is in
lim ite d use, a n d s h o u ld in cre a se as
tra n s fe r lin e s are m o d ifie d o r re ­
p la c e d in th e fu tu r e .

T e c h n o lo g y

body parts. The .computer translates the design into mathe­
matical coordinates thaf can operate automatic drafting ma­
chines and numerically controlled (N/C) machine tools. En­
gineering, drafting work, and tool production operations
can be more closely integrated, thereby speeding pp the
work flow. Computer use in design may affect labor re­
quirements in tfye industry in several ways. First, th§ com­
plex programming required may increase the need for com­
puter programmers. The need for drafters however, should
decline. Such computer-aided design is expected to increase
in the years ahead.
Engineering research and product development. Using com­
puters, engineers can analyze large quantities of data—
some­
times from computer simulation of real situations— solve
to
design or production problems in remarkably short periods
of time. An auto parts manufacturer, for example, saved 9
months to 1 year of development time by using its com­



191

w ill

puter capacities to perform preliminary design calculations
on a new long-life piston ring.2 Computer application t o '
research and development operations is not yet common­
place, but it is a frequently used tool that will probably
become commonplace in the future.
Computer control. The application of computers to the
control of production operations is a major step in the
evolution of computer technology. Computers are being
used to keep track of parts and production materials and to
forecast potential shortages that could disrupt production.
Computer control can aid in attaining uniformity and qual­
ity control in machining. It also can aid in work scheduling
and production line balancing to increase productivity by
directing the proper materials to the worker at his place on
the assembly line. Computers can be applied to a group of
such operations, tying them together in such a manner as to
provide computerized control over an entire manufacturing

or assembly process. The computer system also can make
available large quantities of current data to management to
aid decisionmaking. The major auto manufacturing firms
are using most of these applications but there are no data
on the extent of their use.
Several machine tool manufacturers market computer­
ized control systems for machine operations. One system
links a small computer directly (no tapes are used) to four
machine tool controllers. The system requires only one op­
erator to load stock and oversee the operation of the sys­
tem, and it can perform the work of ten conventional ma­
chines and operations.3 Another system uses a small multi­
purpose computer, memory drum, and a teletypewriter input/output unit to operate simultaneously a combination
(up to 16 units) of N/C machines, special purpose ma­
chines, and transfer machines. One operator can control the
entire system.4
Numerical control

Numerical control is a process of operating machine
tools through a series of electronic control devices and
coded tape instructions. It is a process that is particularly
suitable for the manufacture of metal parts in small volume
because it eliminates the many expensive fixtures, jigs, and
templates otherwise necessary. As such, numerical control
techniques are in limited,'but increasing, use for the fabrica­
tion of the tools and dies needed to operate the industry’s
many high-volume production machines. Extensive use is
being made of numerical control in fabricating sheet-metal
parts— development which ranks among the major applica­
a
tions of numerical control techniques in the United States.
Increased use of numerical control techniques should, as
has occurred in other industries, reduce the need for ma­
chine tool operators.
Applications of numerical control and direct computer
control (discussed elsewhere in this chapter) can be ex­
pected to grow. This is one of several methods the auto
industry can use to improve the flexibility and utilization
of its basic production machines.

Transfer lines

Transfer lines—
highly mechanized production lines—
are
becoming more flexible. Traditionally, transfer lines have
been custom built to do one job. Any significant change in
the job to be done has generally necessitated a significant
change in the construction of the transfer line itself—
an
expensive and time-consuming process.
Flexibility is being increased by the use of “building
block” , or “modular” , transfer lines, constructed from ma­
chinery and equipment consisting of interchangeable, stan­
dardized units. These lines can accommodate changes in
parts design or retooling for new car models with delays
and retooling costs minimized.



192

The inclusion of storage banks for parts at intervals
along a transfer line provides a further increase in flexi­
bility. These storage banks allow a line to continue in oper­
ation even if a station in the line stops. Although not a new
concept, the use of storage banks has yet to be fully imple­
mented. Computer simulation is being used by at least one
manufacturer to predict optimum locations and sizes for
storage banks within the transfer lines. The number of auto­
matic operations performed on transfer lines also is increas­
ing, especially time-consuming gaging and inspection opera­
tions, which allows a reduction in labor requirements and
an improvement in quality control. The new transfer lines
are mechanically more complex, requiring more highly
skilled maintenance crews.
The development of solid-state programmable machine
controllers also contributes to transfer line flexibility.
These controllers operate faster and more reliably than the
older magnetic-relay controllers they are replacing. It is
their programmability that makes them important. Chang­
ing the application of a conventional magnetic-relay con­
troller involves changing the physical wiring in the con­
troller, and each such change can take an hour or more to
make. The programmable controller needs only to be repro­
grammed, which can be accomplished in minutes, rather
than hours. Furthermore, it is possible that the use of pro­
grammable controllers will lead to more widespread use of
computer control.
Machine assembly

Machine assembly (where it can be used) reduces the
high labor content of assembly operations, which may in
turn lower manufacturing costs. In addition, stricter safety
standards and increased emphasis on product performance
and quality can often be better met by machine assembly
than by manual methods.
The potential impact of automatic assembly operations
on labor requirements is considerable because assembly op­
erations are the most labor intensive in the manufacture of
autos. There are many simple, repetitive, and monotonous
assembly operations that are candidates for machine assem­
bly. Similarly, machine assembly can be applied to some
operations that are physically difficult and fatiguing. Job
skills for assemblers tend to shift toward machine monitor­
ing and materials handling. The demand and skill require­
ments for machine maintenance personnel could increase
considerably; these can be met by retraining machinists
who might otherwise be displaced by the new process.
The diversity and productivity potential of automatic
assembly machines are illustrated by the following example
obtained by BLS staff during plant visits: One manufac­
turer uses both an automatic and a manual line to assemble
and test torque converters used in automatic transmissions.
When the automatic line is in full operation, a crew of 8
people per shift is expected to produce as much as is pres­
ently done by a crew of 13 people on the manual line. One

part of the automatic line already in operation inserts
blades into slots in the body of the torque converter—
a
process in which two people per shift (one attendant and
one parts loader) on the automatic line can do as much
work as four people per shift inserting blades by hand.
Several major automakers utilize industrial robots to per­
form many of the welding operations required on a pas­
senger car body, including those that are the most difficult
for employees to accomplish. The robots are programmed
to make a particular type of weld on a specific body style.
The first robot in the line is supplied computer data on the
sequence of body styles forthcoming on the assembly line.
The first robot also contains a master program for control­
ling the succeeding robots on the welding line. Each robot
reportedly can do work equal to U4 welders, thereby reduc­
ing the number of welders needed. This is, however, some­
what counterbalanced by the need for a larger and more
highly trained maintenance crew. Although there may be
little or no labor savings, the quality of the weld is more
consistent than is possible with manual welding.
New materials and processes

The use of plastic materials has grown considerably as
improvements in both the plastic materials and the plastic­
working technology have become available. Advantages of
plastics over steel (in those cases where plastics meet rigid­
ity and strength requirements) include lower weight and
generally lower tooling costs. Increased use of plastics may
reduce labor requirements because plastic parts often re­
quire fewer finishing operations than comparable metal
parts and large, one-piece molded plastic panels (such as
dash panels or front-end body panels) can often replace an
assemblage of sheet-metal parts, reducing assembly time.
Plastics (especially fiber-reinforced composites using glass
or other filaments) are expected to grow considerably in
use because of the increased emphasis on lowering vehicle
weight to improve fuel economy. Aluminum and special
steels also will be used more widely for a growing number
of auto components to reduce weight.
The fabrication of metal parts from metal powder is
more widely used in the automotive field than in any other
industry, and may become even more important due to
recent improvements in materials and manufacturing pro­
cesses. Powder metallurgy parts can be made in complex
shapes, of high strength, and to such close tolerances that
many secondary machining operations and inspection pro­
cedures can be reduced or eliminated, thereby reducing
labor requirements.

©ytpyft sod

©ytS@ k
@

O utp u t

Industry output increased at an average annual rate of 4.8
percent between 1960 and 1975. The growth rate was higher



193

during 1960-67 period—averaging 7.9 percent a year—than it
was during the more recent 1967-75 period, when it averaged
3.2 percent a year. The lower growth rate of recent years
reflects several negative economic factors: There was a
moderate recession and a major industry strike in 1970,
followed—from late 1973 to 1975—by an oil embargo, a
period of high inflation, and a severe recession. What tends to
be obscured in this growth rate figure is that output rose to
record levels in 1971, 1972, and 1973. Auto sales began to rise
in late 1975. and continued strong during 1976.
Historically, “regular” size passenger cars have been the
mainstay of U.S. auto manufacturers. During the late
1960’s, however, smaller passenger cars—
intermediates,
compacts, and subcompacts, both domestic and foreign—
became more important in the marketplace at the expense
of regular size and large cars. According to Ward’s Automo­
tive Yearbook, intermediate and small cars accounted for
almost 40 percent of new car registrations in 1966. By
1975 this figure had grown to 77 percent.
The trend toward smaller cars will continue in response
to the present Federal Government regulations for fuel
economy (27.5 miles per gallon by 1985) set in the Energy
Policy and Conservation Act of 1975. To meet such a fuel
economy goal with current automotive technology will re­
quire a rather large shift to small cars. The popularity of
such a shift among car buyers remains to be seen. During
the energy crisis from late 1973 to early 1974 the demand
for small, fuel-efficient cars was strong. But as fears of gaso­
line shortages declined, so did some of the enthusiasm for
the smallest cars. The strongest sales for 1975-76, according
to industry sources, were of the intermediate and larger
autos, although sales of the smaller cars did not decline. As
of late 1976, however, some dealers were offering discounts
on some subcompact models in an effort to improve their
sales.
The most likely market structure over the next 5 to 10
years will be a general reduction in size in all categories.
Passenger cars presently considered to be of “intermediate”
size may well become the standard size. A demand for “full
size” cars is expected to continue if production of such cars
remains possible under the fuel economy regulations. Sev­
eral domestic manufacturers have expressed concern that
the various Federal regulations on fuel economy, exhaust
pollution, and safety standards could affect the size, perfor­
mance, and general desirability of future passenger cars.
Demand for light trucks and vans (less than 14,000
pounds gross vehicle weight) has been strong since the late
1960’s as recreational vehicles gained in popularity. During
1973, demand for heavy trucks (which had increased stead­
ily after a 1970 slump) also grew sharply. Truck production
peaked at a record level in 1973, then dropped slightly, but
surged to a new record in the 1976 model year. Truck
trailer production dropped sharply in 1975, in part because
of heavy purchases in late 1974 as customers sought to
avoid purchasing 1975 units that were required by law to
have expensive anti-skid braking equipment.

Productivity

Output per employee-hour increased at an average an­
nual rate of 3.2 percent from 1960 to 1975. The in­
crease averaged 3.6 percent annually during 1960-67,
slightly higher than the 3.2-percent productivity growth
rate achieved during 1967-75.
Growth in output per employee-hour was particularly
strong in 1971 and 1972 as output rose sharply from 1970
in response to a very strong demand for new cars and
trucks. Productivity continued to grow in 1973 as manu­
facturers reported a third year of record new car and truck
sales; however, by the fourth quarter of 1973, retail sales
had begun to fall, causing a final-quarter decline in both
output and productivity levels. The decline in productivity
continued through 1974, during which there was a sharp
drop in output and in employee hours, but a considerably
smaller drop in the number of people employed. Appar­
ently the manufacturers chose to cut working hours (espe­
cially overtime) and keep their work force intact.
The year 1975 was unusual for the industry in terms of
output and productivity. Output continued to decline (for
the second year in a row) due in part to the recession and in
part to higher auto prices. Although output declined, pro­
ductivity increased substantially. In this instance, both em­
ployee hours and total employment declined at about the
same rate—
and both dropped considerably more than did
output. Thus, the productivity increase resulted, for the
year as a whole, from a large drop in the industry’s work
force and a much smaller drop in output. In fact, output
actually increased in two quarters during the year, while
employee hours remained at low levels during all four quar­
ters.

Investment
Capital expenditures

Expenditures for new plant and equipment, in current
dollars, increased from $790 million in 1960 to $2.1 billion
in 1975, an average of 7.4 percent per year. An estimated
$2.4 billion was spent in 1976. Since current-dollar figures
do not take into account price increases over the years, real
capital outlays were less than these figures indicate. The
rate of increase in capital expenditures was significantly
higher between 1960 and 1967, when,the industry was ex­
panding its productive capacity, than during the more re­
cent 1967-75 period. The average annual rates of growth
were 16.0 percent in 1960-67 and 6.9 percent in 1967-75.
As shown in table 6, the rate of increase in capital ex­
penditures per production worker was also greater during
the first half of the 1960-75 period. Plant and equipment
expenditures per production worker in 1974 reached a peak
of $4,266, or triple the 1960 total of approximately $1,400
per production worker, and then declined to $3,460 per
production worker in’ 1975.



194

Capital spending is expected to increase strongly over
the next several years. A recent McGraw-Hill survey of capi­
tal spending plans5 indicates that planned expenditures for
1977 will jump to $4.15 billion, followed by an increase to
$4.36 billion in 1978. One manufacturer plans to invest
$15 billion by 1980 for new, redesigned, smaller passenger
cars, while another manufacturer plans to spend almost $2
billion (worldwide) in 1977, and over $2 billion a year in
1978, 1979, and 1980.6
This high level of capital spending is necessary to design
and produce car models that will meet Federal Government
standards for safety requirements, exhaust pollution levels,
and—
most especially—
fuel economy. While funds will be
invested in all of the production phases, the emphasis will
be on new tooling for updated car models.
The increasing importance of capital relative to labor is
reflected in a decline in the ratio of payroll to value added,
from 0.451 in 1960 to 0.419 in 1972, an annual average
rate of decline of 0.1 percent. (See table 6.)
Funds for research and development

Expenditures for research and development (R&D) in
the industry group of motor vehicles and other transporta­
tion equipment except aircraft7 increased from $884 mil­
lion in 1960 to a planned level of $2.4 billion in 1974, or at
an average rate of 7.2 percent a year. Company R&D ex­
penditures were 2.3 percent of net sales in 1960, increasing
to a planned level of 2.8 percent in 1974. R&D expendi­
tures are expected to rise to $3.1 billion by 1977.8
Research is underway to develop new automobile power
plants that meet exhaust emission standards and provide
improved fuel economy. Alternative types of power plants
being considered range from modified conventional piston
engines to alternative engine concepts including the rotary
engine, diesel engine, and turbine, Stirling cycle, and electric
engines. The approach found most feasible by most major
Table 6. Indicators of change in the motor vehicle and equipment
industry, 1960-75
In d ic a t o r

C a p ita l e x p e n d itu r e s p er
p r o d u c tio n w o r k e r ..............
P a y ro ll p e r u n it o f va lu e
a d d e d .........................................
R esearch a n d d e v e lo p m e n t
e x p e n d itu r e s 3 ......................

A v e ra g e a n n u a l ra te o f c h a n g e 1
1 9 6 0 -7 5
1 9 6 0 -6 7
1 9 6 7 -7 5

5 .9

1 1 .8

7 .4

2- 0 . 1

- 0 .5

2- 0 . 5

4 7 .2

6 .8

4 9.1

1 L in e a r lea st sq ua re s tr e n d s m e th o d .
2 F in a l y e a r = 1 9 7 2 .
3 D a ta are f o r m o to r v e h ic le s a n d a ll o th e r t r a n s p o r t a t io n
e q u ip m e n t e x c e p t a ir c r a ft , a n d a re based o n e x p e n d itu r e s o f
m a n u fa c tu r in g c o m p a n ie s in th e t r a n s p o r t a t io n in d u s tr y (e x c e p t
a ir c r a ft c o m p a n ie s ) th a t have re sea rch a n d d e v e lo p m e n t p ro g ra m s .
1 9 7 4 fig u re s are based o n c o r p o r a te s p e n d in g p la n s as r e p o r te d b y
M c G r a w - H ill.
4 F in a l y e a r = 1 9 7 4 .
S O U R C E : B u re a u o f L a b o r S ta tis tic s , B u re a u o f E c o n o m ic
A n a ly s is , B u re a u o f th e C ensus, N a tio n a l S c ie n c e F o u n d a tio n , a n d
M c G r a w - H ill.

automobile manufacturers starting with 1975 models is
modification of the piston engine through application of
catalytic converters— device attached to the exhaust sys­
a
tem which uses platinum and palladium as catalytic agents
to convert noxious auto exhaust emissions into water
vapor and carbon dioxide. The “stratified charge” engine, a
conventional piston engine with an unconventional cylinder
head, reportedly has the capability to meet most of the
strict emission standards to be implemented after 1978 and
may, according to some experts, become more widely used
in the early 1980’s.
While some improvements in fuel economy may result
from refinements in engine design, reducing automobile
weight is probably the best way to improve fuel economy.
Building smaller cars and substituting lightweight materials
(such as aluminum and plastic) are two of the more obvious
ways to reduce weight. One manufacturer has already intro­
duced some new car models that are smaller and lighter
than the corresponding models of previous years-and this
trend will continue.

Employment and Oasypationa! Trends
Employment

Employment in this industry rose from 724,100 in 1960
to a peak of 955,300 in 1973 and then dropped sharply as
economic conditions turned downward and auto sales fell,
to 774,100 in 1975. This pattern represents an average
growth rate of only 1.6 percent a year between 1960
and 1975. During the first half of this period, 1960 to
1967, employment grew at an average annual rate of 3.6
percent. Between 1967 and 1975, however, employment
declined by an average of 0.2 percent a year. As sales and
production rose again in 1976, employment increased to
850,600.
The long-term trend, however, is for a decline in employ­
ment. The BLS projections for 1973-85, indicate a parti­
cularly sharp decline from 1973, when employment was at an
all-time high.
Employment in the motor vehicle and equipment indus­
try is concentrated in two industry sectors: Motor vehicles
(SIC 3711), and parts and accessories (SIC 3714). The
motor vehicles sector employed 41 percent of the indus­
try’s work force in 1960 and 42 percent in 1975. Employ­
ment in the parts and accessories component of the indus^
try accounted for 43 percent of the work force in 1960 and
45 percent in 1975.
The ratio of production workers to total employment
has remained fairly stable; production workers accounted
for 78 percent of total employment in 1960 and 77 percent
in 1975. The rate of employment growth for production
workers during 1960-75 was 2 percent-about the same as
the all-employee growth rate indicated earlier. The rates of
growth in employment of production workers during the



195

shorter term 1960-67 and 1967-75 periods closely parallel
trends for total employment.

Occupations

Technological and other changes are expected to alter
the occupational structure of the motor vehicle industry by
1985. Employment is expected to increase in only three of
the eight major occupational groups—managers, officials, and
proprietors; sales workers; and operatives. In the other major
occupational groups employment is expected to decline,
ment is expected to decline.
Increased use of computers in design, engineering, and
production applications should bring about several changes
among professional and technical workers and clerical
workers. The number of computer specialists (primarily sys­
tems analysts and programmers) is expected to increase by
8 percent. Greater use of computer terminals should in­
crease the productivity of drafting technicians and engi­
neers, although the effect of this on employment is unclear.
If the volume of work were to remain unchanged, employ­
ment might decline. But there is a strong possibility that
computer techniques will be used more intensively to im­
prove vehicle design and weight optimization—
new analyti­
cal work which could absorb people who might otherwise
not be needed. An increase of 34 percent is expected for
computer peripheral equipment operators. Keypunch oper­
ators are expected to decline*by 58 percent as punchcard
data entry is supplanted by more sophisticated forms of
data entry.
Operatives (semiskilled workers) will continue to be the
largest occupational category in the motor vehicle industry,
making up about 50 percent of the work force. Many of
these workers are engaged in production operations that are
relatively labor intensive and have potential for further
automation. Semiskilled metal workers (drill press opera­
tors, lathe operators, welders, etc.) are expected to decline
by 20 percent in response to more widespread use of nu­
merically controlled machines, industrial robots for welding
and inspection operations, and more automatic transfer
lines.
Although some advances are anticipated in automatic (or
machine) assembly operations, the job category of assem­
blers is expected to grow by 34 percent to employ almost
168,000 people by 1985— far the largest single occupa­
by
tion in the industry. The general increase in automated pro­
duction and inspection operations should serve to limit any
increase in the number of inspectors needed. Training for
many of the semiskilled jobs is relatively brief, consisting
primarily of on-the-job instruction for periods of several
days to several weeks. Hence, shifting semiskilled workers
from one position to another generally should not cause
great dislocations.
The impact of advanced production machines on occu­
pational skills was discussed with officials from several auto

manufacturers visited by BLS staff. In general, a shift to­
ward skilled workers is expected—
especially in computerrelated occupations—
with a decline in unskilled workers
and semiskilled machine operators. Maintenance workers
would be the occupation most greatly affected, with de­
mand for these workers rising in step with increases in the
use of N/C machines, industrial robots, and other auto­
mated machines. Skilled machinists who are displaced by
automated machines can be retrained to maintain the new
equipment.

Adjustment of workers to technological change

The impact of technology on jobs is probably not as
critical in the auto industry as it is in many other industries.
A substantial proportion of blue-collar jobs are in semi­
skilled occupations, and operators displaced from one job
can be retrained for other jobs more easily than in indus­
tries with high skill level requirements. Also, there are areas
in auto production (such as final assembly) that are fairly
labor intensive, and will continue to be so in the foreseeable
future.
Approximately two-thirds of the industry’s employees
are covered by collective bargaining contracts. All of the

contracts contain general provisions pertaining to seniority,
layoffs, grievances, retirement, and supplementary unem­
ployment benefits that could be applied to job losses result­
ing from technological change. Additionally, contracts
with two manufacturers contain specific statements con­
cerning technological change. In both cases, the contracts
have provisions that require advance notice to the union of
planned technological changes, create training programs for
qualified employees within the bargaining unit, and allow
problems not otherwise resolved to be submitted through
the regular grievance procedures.
The recession of 1974 and 1975 caused considerable tur­
moil in the auto industry. Employment dropped substan­
tially and some plants were shut down sufficiently long for
a number of laid-off employees to exhaust their unemploy­
ment benefits. By the time new labor contracts were due to
be negotiated in late 1976, production and employment
had returned to healthy levels— the recession probably
but
left its imprint on the contract negotiations. In a 4-week
strike at one manufacturer, the United Auto Workers won a
shorter work year. Employees will receive a total of 13
additional days off over the 3-year contract period, which
will serve to create new jobs over the short run and preserve
job security in the future. The other manufacturers have
since agreed to this pattern.

5Preliminary Plans for Capital Spending in 1977-78, McGrawHill Fall Survey, Fall 1976.

1These data exclude employees in a number o f industries which
produce components for the motor vehicle industry. According to
estimates o f the Motor Vehicle Manufacturers Association, more
than 517,000 workers are engaged in producing motor vehicle com­
ponents and thus are classified in industries other than SIC 371,
motor vehicles and equipment.

6 “Capital Spending to Set Record in ’77,” Autom otive. Indus­
tries, October 1, 1976, pp. 14-15.
7 Motor vehicles and other transportation equipment except air­
craft consists o f SIC’s 371, 373, 374, 375, and 379. Separate data
for the motor vehicle industry, SIC 371, were not available until
1972. The importance o f motor vehicles within this industry group
is illustrated by the fact that the motor vehicle segment accounted
for over 98 percent o f the industry group’s R&D funds in 1972 and
1973.

2 “Computer Speeds Design Production of Piston Rings,” A u to ­
m otive Industries, November 15, 1968, pp. 79-85.
3 “N/C and C/C, New Keys to Productivity,’'A u to m o tive Indus­
’
tries, O ctober 15, 1972, pp. 33-36.
4 “Computer Controlled Machining,” A u tom otive Industries,
July 15, 1970, pp. 51-52.

8 R&D expenditures for 1960, National Science Foundation;
planned R&D expenditures for 1974 and 1977, McGraw-Hill.

SELECTED
“Computer Controlled M a c h in in g A u tom otive Industries, July 15,
1970, pp. 51-52.

“ Lordstown Plant: GM’s New Mark o f Excellence?,” Iron Age,
March 11, 1971, pp. 39-40.

“Computer Speeds Design Production o f Piston Rings ''A u to m o tive
Industries, November 15, 1968, pp. 79-85.

“ Machine Assembly . . . Industry’s Last Change for Increasing Pro­
ductivity,” A u tom otive Industries, April l , 1972, pp. 35^4-1.

“ Detroit’s Frantic Hunt for a Cleaner Engine,” Business Week,
December 9, 1972, pp. 60-70.

“ Materials: New Marriages in Design,” A u tom otive Industries,
December 15, 1973, pp. 37-47.

“Gage-Assemble-Test Warms Up Again.” A utom otive Industries, C)ctober 15, 1976, pp. 24-27.

“N/C and, C /C -N ew Keys to Productivity,” A u tom otive Industries,
October 15, 1972, pp. 33-36.

“ How Computers Unify Manufacturing,” A u tom otive Industries,
June 1, 1974, pp. 31-36.

“ Powder Metallurgy: Phase II,” A u tom otive Industries,. July J ,
1972, pp. 25-28.




196

T@
(g[h[n)® gf and Labor in
S@
Petroleum (Refining
Rose N. Zeisel and Micheal D. Dymmel

S y m m a ry
Technological changes in the refining industry are
being made in response to shifts in crude oil supply,
changing demand for petroleum products, and envi­
ronmental and energy considerations, in addition to
the usual incentives of greater productivity and low­
er costs. These changes are primarily in the areas of
cracking, hydrotreating, and reforming, in associa­
tion with advanced instrumentation and computer
control. The outlook is for greater emphasis on pro­
cesses for desulfurization and octane improvement.
Because the industry is capital intensive, the shortrun effects on labor are likely to be minimal, but in
the longer run they will alter job content and may
reduce employment growth.
Productivity rose sharply from 1960 to 1977, at an
average rate of 4.3 percent annually compared with
2.6 percent in ail manufacturing. From 1967 to 1977,
the rate was 3.0 percent. The outlook to 1985 is for
productivity to rise but at a slower rate than in the
last decade. Many uncertain variables affect the out­
look including crude imports, gas supplies, and gov­
ernment environmental and energy policies. But for
the most pari, the industry’s productivity in the next
decade will depend on the Nation’s economic growth
and consequent energy needs. Changes in govern­
ment policies or in the international situation are not
dealt with in this chapter.
Capital investments have been increasing almost
steadily since the 1960’s. By January 1977, operable
capacity had risen 50 percent over the decade and 20
percent since January 1973, reversing concerns
about capacity shortages. Due to uncertainties of
supply and demand and rapidly rising costs, howev­
er, there is no general agreement on future capital
outlays for capacity expansion. But large invest­
ments are anticipated to accommodate changing
demand for the industry’s products and government
environmental and energy policies.
About 160,000 people were employed in the indus­
try in 1977, the largest number since 1962. Following
a sharp decline in the first half of the !960’s reflect­
ing very rapid productivity growth, employment was
Reprinted from BLS Bulletin 2005 (1979),
Technological Change and its L abor Im pact in Five Engergy Industries.




S97

relatively stable until 1973. Since 1973, however,
employment has been moving up as technology
changes require more unit labor and as the number
of very small refineries increases. The outlook to
1985 is for a resumption of the decline.

Technology in the 1970’s
Petroleum refining is a series of processes of phys­
ical separations and chemical reactions, it involves
three major groups of processes: Separation, conver­
sion, and treating. First the hydrocarbon compounds
in the crude oils are separated through heating and
distillation to recover the lighter products such as
gasoline, kerosene, and distillate fuels. Some com­
pounds heavier than gasoline may be “ cracked” or
chemically converted into higher quality products.
Desired products may also be built up by chemical
reactions such as alkylation. Others are chemically
rearranged, by catalytic reforming, for example. In
addition, at some stage of manufacture the products
may be treated to remove impurities such as sulfur
or metals.
In the past, the objective of U. S. refineries was
to maximize gasoline production rather than the out­
put of heavier fuels. Consistent with this objective,
they were geared, primarily, to producing high-oc­
tane gasoline from low sulfur (sweet) crude petro­
leum. Moreover, in general, there were no restric­
tions on levels of sulfur and other impurities in pe­
troleum products.
However, the picture is changing. First, there
appears to be a long-term shift in emphasis from
gasoline to heavy fuels based on a projected slow­
down in gasoline demand and an increase in the
market for heavy fuels as a result of the natural gas
shortage. Second, environmental protection regula­
tions encourage or require low-sulfur, low-lead prod­
ucts, as well as the reduction of noxious wastes
from the refining process itself. At the same time,
however, the availability of low-sulfur varieties is
declining. As a result of these conditions, refineries
must make adjustments to accommodate product
changes.

In addition, changes are taking place in the struc­
ture of the industry. Although a number of small
refineries are being built, in general, process units
are becoming larger, and functions are being consoli­
dated to increase productivity. Average capacity has
increased very sharply to over 60,000 barrels per day
(as of January 1977), and the labor implications (dis­
cussed in the employment section) are significant.
Capacity varies considerably among refineries, how-,
ever, ranging from 500 to 640,000 barrels per day. Sn
general, the smaller plants consist of a crude oil dis­
tillation unit plus the necessary auxiliary units, while
the larger refineries are considerably more complex.
They include, in addition to distillation facilities,
Tab!© 4.

cracking, reforming, coking, hydrogen-treating, alky­
lation, fuel desulfurization, and other processing un­
its. Key advances in the basic refining processes in
the last decade, their labor impact, and their rate of
difusion are presented in table 4 and discussed in
greater detail below.

Computer control

High-speed digital computers improve production
efficiency and raise quality through - more precise
control of the production process. Other benefits
also are often cited, such as better technical and
operating data and improved plant safety.

tectaology ebsnges in petroleym srafmiing

Technology

Description

Labor implications

Diffusion

C om puter control

H igh-speed digital com puters, in
association with highly com plex
instrum entation, m onitor and/or
control various refinery process­
es; th e y .a re used in testing and
research laboratories and for
m anagem ent inform ation.
Use
minimizes co sts and im proves
product quality.

A ffects o p e ra to r’s duties prim ari­
ly, assum ing earlier installation of
sophisticated
instrum entation;
requires com puter-related techni­
cians.

Installation in one-fourth of refin­
eries constituting m ore than twothirds of industry crude capacity.

Im proved cracking

Improved
riser-cracking
tech- •
niques use cataly sts with more
tolerance to fe ed sto cks of higher
metal content to provide greater
yields o f desired products and
higher octane ratings. Im proved
hydrocracking
provides
more
feedstock flexibility.

Increased labor productivity; di­
rect effects are minimal.

R iser m ethod constitutes approxi­
m ately 40 percent of U .S. c ra ck ­
ing capacity. H ydrocracking ca­
pacity is equal to 16 percent of
the total. Diffusion is expected to
be relatively slow.

D esulfurization ad v an ces

H igh-activity catalysts and other
advances efficiently reduce sulfur
content. H ydrogen-based process­
es enable refineries to process
so u r crude, to m ake low -sulfar
feed sto ck s for m odern </.<aiytic
reform ing units, to produce resid­
uals and distillates to environm en­
tal specifications, and to meet
pollutant em ission controls.

Additional processing, increases
unit labor requirem ents for tech­
nicians and m aintenance person­
nel.

Process units being built into new
refineries. Diffusion will depend
on environm ental protection re­
quirem ents and type of crude
available. H ydroprocessing capac­
ity increased 30 percent betw een
1975 and 1.978, and is expected to
increase another 5 percent by
1980.

O ctane-im proving processes

C atalytic reform ing, alkylation,
and isom erization increase gaso­
line octane ratings w ithout lead
additives. N ew bim etallic cata­
lysts are im proving all reform ing
m ethods. C ontinuous reform ing
elim inates periodic shutdow ns for
catalyst regeneration.

Direct labor effects depend on
refinery com plexity. Small plants
may need additional operators
and m aintenance w orkers. In all
cases, productivity would be ad­
versely affected.

By !978, reform ing accounted for
22 percent of crude capacity;
isom erization, 2 percent; alkyla­
tion, 5 percent. Low -lead require­
m ents suggest increased im por­
tance of octane-im proving proc­
esses.

E nergy conservation m ethods

Increased use of heat exchangers,
furnace air preh eaters, therm al
insulation, gas and hydraulic tur­
bines, w aste-heat steam genera­
tion, process im provem ents.
j

Increases m aintenance labor, par­
ticularly in older refineries; also
increases dem and for engineering
skills.

By early 1976, energy use was cut
!0 p ercent below 1972. E xpecta­
tions are fo r an additional 15-per­
cent cut by 1985.

Preventive m aintenance technologies

U se of ultrasonic testing. X-ray
testing, infrared cam eras, m agnet­
ic particle testing, and corrosion
probes to determ ine equipm ent
reliability.

N ew er sophisticated preventive
m aintenance equipm ent requires
highly trained personnel, but may
require few er unit em ployee
hours as dow ntim e is reduced.
M aintenance craft consolidation
also reduces unit labor require­
m ents.

N ew testing m ethods are widely
used; use depends on com plexity
and age of equipm ent.




198

In process control, digital computers are applied
to various refining processes ranging from crude dis­
tillation to on-line gasoline blending. Open-loop con­
trol is most common: data received from plant wide
on-stream sensors are monitored and the operator is
notified when machine changes are required. Howev­
er, closed-loop control is increasing in use in the
newest installations. The trend is toward use of minsor microcomputers which, while linked to a central
control center, control separate functions. Required
adjustments in the production process are made au­
tomatically, thus eliminating some of the operator’s
functions.
Digital computer use is generally more common in
large complex plants. Approximately one-fourth of
American refineries use digital computers in various
applications,1 but these plants constitute more than
two-thirds of total U.S. capacity.1* With current
2
trends towards the construction’ of larger, more
complex plants, it is expected that practically all
future refineries will incorporate one or more pro­
cess control digital computer systems.
The use of very sophisticated instrumentation
generally precedes or accompanies computer instal­
lation, and so the labor implications of the computer
cannot be easily sorted out. In a refinery visited by
the BLS staff, the computer monitored information
from more than I,(XX) electronic instruments, such as
chromatographs, mass spectrometers, and octane
analyzers, that are located at the process unit and
continuously measure product qualify. Their import­
ance lies in the speed with which problems can be
corrected, and also in1 their tie-in to computer con­
trol. All U.S. refineries use analyzers, but the num­
ber and sophistication of the instruments vary with
the size and complexity of the plant.
The effect on employment is associated largely
with the degree of sophistication of the refinery’s
instrumentation. For example, fewer analyzer repair­
ers, operators, and lab technicians may be required
where on-line monitoring is possible.3 In a modern
plant, one technician may take the place of three or
four technicians or operators in an older plant which
still maintains sample testing and manual recording.
On the otfier hand, jobs such as programmers and
systems analysts increase with the installation of
computers. A BLS study4 shows, however, that the
number of some computer-related jobs in a plant
may decrease after the initial phases of installation
and programming are completed.
1International Petroleum

Encyclopedia, 1977, pp. 443 - 44.

2 Ibid., op. 316 -2 0 .

1 “ Process Computers— They Do Pay Off in Refineries,” Oil
and Gas Journal, Dec. 3, 1973, p. 62.
4 Computer Manpower Outlook, Bulletin 1826 (Bureau of Labor
Statistics, 1974), pp. 36-37.



With the installation of computer process control,
changes are necessary in the operator’s duties. One
example is clearly shown in a BLS survey of em­
ployment implications of computer process control.-5
The duties of an operator of a fluid catalytic crack­
ing unit before computer control were to manually
adjust automatic analog controllers at the control
console and to monitor automatic data logging equip­
ment. After installation, the computer controls and
monitors a large part of the process and automatical­
ly logs the data, although the operator still performs
manual control. In case of emergency, the operator
can take control of any part or all of the process.
improved eemiking
Fluid catalytic cracking is a refining process that
converts heavier oils into lighter, more valuable
products such as gasoline, primarily by chemical
reaction in the presence of a catalyst. The.--technique
of riser cracking was developed concurrently with a
new generation of highly active catalysts in the early
1960’s. Considered more efficient than older tech­
niques, this method presently is in use in over 40
percent of U.S. cracking capacity.
Hydrocracking, an older, but greatly improved
method of cracking, has several advantages over the
conventional “ cat-cracker.” These include the capa­
bility to meet environmental specifications of low
sulfur and nitrogen more efficiently and flexibility to
handle variations in crude stock and in products de­
sired. Currently, however, hydrocracking accounts
for only about 16 percent of cracking capacity, and
the diffusion is expected to be slow, due primarily to
the very high investment and energy usage required
for this process.
For both types of cracking, continually improved
catalysts and regeneration methods enable more
efficient processing of oils with high contents of sul­
fur and metals. To meet new product specifications
and increased use of high-sulfur crudes, refineries
with older cat-crackers may need to change over to
the more efficient processing methods.
Sn general, the effect on labor utilization of im­
proved cracking procedures in place of older crack­
ing methods is minimal.
0®sylfyrizati©n advances
Hydroprocessing to reduce impurities, particularly
sulfur, will become increasingly important as de­
mand increases for low-sulfur residual and distillate
fuels. In addition, stricter environmental protection
regulations and greater utilization of high-sulfur
crudes due to dwindling supplies of sweet crudes
5 Outlook for Computer Process Control: Manpower Implica­
tions in Process Industries, Bulletin 1658 (Bureau of Labor Statis­
tics, 1970), p.29.

19®

tane ratings over those possible with conventional
catalysts. In addition, the process of continuous re­
forming eliminates periodic shutdowns because, un­
like other reforming methods, it continuously rege­
nerates the catalyst.
To meet the low-lead requirements, refiners in­
creased their reforming capacity by roughly 19 per­
cent between 1972 and 1978. In that period, the use
of bimetallic catalysts more than doubled, to over 60
percent< total reforming capacity. At the start of
of
1978, continuous reforming accounted for 4 percent
of total catalytic reforming capacity compared with
about 69 percent for semi-regenerative reforming,
and almost 27 percent for the older process of cyclic
reforming.7
In addition to reforming, the processes of alkyla­
tion and isomerization provide increased octane rat­
ings in gasoline. Developed in' the early 1940’s to
produce aviation fuels, these processes are not wide­
ly used now, representing only about 5 percent and 2
percent, respectively, of crude capacity (compared
to 22 percent for reforming). However, they will
become increasingly important as leaded gasoline is
phased out.
Although the phasedown of leaded gasoline may
not be as severe for the large refineries capable of
wide process adjustment, it will be particularly diffi­
cult for smaller, older refineries, geared primarily to
producing leaded gasoline: Some of these smaller
refineries may have problems associated with capital
acquisition or procurement and construction of the
needed equipment. In addition, more operators may
be needed in refineries lacking process control sys­
tems. More maintenance labor may also be required
by the small plant. In all cases, however, productivi­
ty would be adversely affected by the additional
processes required to increase octane ratings.

will lead to strong growth in desulfurization capaci­
ty.
Desulfurization is an important factor in better
enabling refineries to process sour crudes. For this
purpose, sulfur removal may follow initial crude dis­
tillation. Desulfurization is also performed down­
stream to meet the stringent requirements of cataly­
tic reformers and to control pollutant emissions from
the catalytic cracking process. Residual and heavy
gas-oil desulfurization, the smallest but most rapidly
growing segment of hydroprocessing capacity in the
United States, is performed as the last step in the
production of those fuels. New refineries are being
designed to produce low-sulfur products from highsulfur crudes, and existing plants are revamping
their process units when changes become necessary.
Various hydrogen-based processes (hydrodesulfurization, hydrorefining, hydrotreating) are used for
sulfur removal. All are based on chemical reactions
between oil and hydrogen in the presence of a cata­
lyst. Advances involving separate demetallization
processes and new high activity catalysts are reduc­
ing problems related to metals accumulation and
need for frequent catalyst regeneration. However,
the costs are still quite high. With some desulfuriza­
tion processes, off-site regeneration of catalysts by
specialized companies is increasing as an economical
solution to catalyst-related problems.
The trend is clearly toward an increasing capabili­
ty of refineries to process sour crudes. Total hydro­
processing capacity increased 30 percent between
1975 and the start of 1978 and is expected- to in­
crease an additional 5 percent by 1980. Residual and
heavy gas oil desulfurization capacity more than tri­
pled from 1975 to |978 6
Since additional processing is required for desul­
furization and demetallization, unit labor require­
ments for operations and maintenance personnel
may increase. These increases may be temporary,
however, as the processes become integrated into
the overall operation of the refinery,
Oetasie improvement

Catalytic reforming, a process which. improves the
octane rating of gasoline or fuels, is particularly
important today in 'view of: the Federal Govern­
ment’s requirements for lower lead and lead-free
gasoline. To increase yields of high-octane gasoline
without lead, low-sulfur feedstock is necessary. The
desulfurization of feedstock, discussed in the pre­
vious section, is therefore necessary.
New bimetallic catalysts are making all reforming
processes more efficient, increasing yields and oc-6

7lbid.

6 “ Federals Shape U.S. Refining Industry,” Oil and Gas Jour­
nal, Mar. 20, 1978, pp. 6 3 -6 6 .




Energy co n serv atio n
Because of the high costs of new refining technol­
ogies, particular emphasis is bei/ig placed on reduc­
ing costs through energy conservation—the more
efficient utilization and generation of fuel and pow­
er. 8 Current technologies such as heat exchangers,
furnace air preheaters, gas and hydraulic turbines,
waste-heat steam generation, and thermal insulation
are* now being improved. Minor process adjust­
ments, automatic instrumentation, increased mainte­
nance, and intensified surveillance of operations are
also important in reducing refinery energy consump­
tion.
The labor implications of energy conservation in
the refinery are considerable. Some companies have

8 Oil and Gas Journal, Mar. 29, 1976, p. 74.

zm

set up energy systems departments whose managers
closely control energy use. In addition to managerial
and engineering skills, more employee hours of
skilled craft and maintenance workers may he re­
quired for efficient energy utilization, particularly in
the older refineries.

There is no general agreement on the level of
domestic demand for refinery products in 1985JO
Many observers expect gasoline demand to peak out
in the next few years, while demand for distillate
and residual fuels is expected to increase as utility!
and industrial users substitute fuel oil for natural
gas.

Imports

Preventive msiomferaanc®
Special emphasis is being placed on preventive
maintenance, particularly the use of electronic in­
struments to locate defects in and measure the deter­
ioration of equipment before problems arise.
Through the use of ultrasonics, X-rays, and electri­
cal corrosion probes, wear and corrosion in pipes
and vessels can be measured on- or off-stream, sonic
testers can detect high-frequency sounds generated
by gas leaks from valves and fittings, and magnetic
particle tests and infrared cameras can also pinpoint
structural defects in some equipment.
Preventive maintenance reduces downtime and
maintenance costs, but the effect on labor is difficult
to assess. Maintenance labor requirements vary with
the complexity and age of the refinery, the sulfur in
the crude, and the extent to which maintenance is
subcontracted. Newer refineries may have less
maintenance because modem materials, e.g., corro­
sion resistant, are more fully utilized. In general,
however, important changes are occurring which are
reducing unit labor requirements for maintenance
personnel. These are discussed in the section on
employment and occupational trends.

Production add Pr®du<eiw% ©ytloolk
Gytjpyt
The steady growth in petroleum refining output
since World War II was interrupted only by small
declines in 1949 and 1958 and again in 1974 and
1975. Overall, from I960 to 1976, output rose at an
average rate of 2.9 percent annually.9 The growth rate,
however, was considerably more rapid in the strong
economy of 1960-66 (3.2 percent) than in the 1966-67
period (2.4 percent). The latter period included the em­
bargo and the 1974-75 production cutback associated
with the economic recession and energy conservation.
But in 1976 and 1977, output jumped to peak levels,
recording the most rapid annual rates of growth since
1955.

Until the early 1960’s, the United States was ...selfsufficient in refined petroleum. Even in the first half
of the 196Q’s, domestic refining capacity could sup­
ply more than nine-tenths of domestic demand;
product imports consisted almost entirely of residual
fuel. But in the 1965— period, demand for petro­
73
leum products expanded considerably more rapidly
than capacity, and by January 1973 operable capaci­
ty could supply less than 80 percent of the demand.
Thus, the gap between domestic demand and supply
had been growing for years when the crude oil em­
bargo and price increases intensified the problem.
Product imports rose to peak levels in 1973, averag­
ing 3 million barrels daily." Residual fuel was still
the major refined product imported but other imports
had also risen substantially.
However, the demand/supply situation reversed
itself following the crude oil embargo when concern
rose sharply about our self-sufficiency. Capacity in­
creased as plants were expanded and new plants
were built, while demand declined after three de­
cades of continuous growth. Consequently, the gap
between refining capacity and demand greatly nar­
rowed, and our dependence on imported irefined
products in 1975 dropped back to roughly that of the
mid-1960’s. Although demand has been rising, capac­
ity increases continue to hold down the gap filled by
product imports. In 1976, im ports of .refined prod­
ucts averaged 2 million barrels daily, It 1/2 percent
of domestic demand, the lowest proportion since
1967.
But crude distillation capacity alone is not a mea­
sure of tj\e industry’s capability to provide for
domestic demand, even assuming available crude
supplies. Residual fuel has been and continues to be
our major import because,. (as discussed' earlier,,
domestic refineries have not been interested in or
geared to processing residuals. The problem is now
complicated by government regulations which re­
quire Sow-suSfur residuals, sometimes necessitating
changes in technology. Nevertheless econom ic in­
centives and the outlook for rising demand have re10 See P ro jectio n s o f E n erg y S u p p ly a n d D em a n d an d 'Their Im ­
p a cts: A nn ual R e p o rt to C on gress. Vol. II, 1977 ( U.S. Department

9

P ro d u c tiv ity In d exes fo r S e le c te d In dustries,

1977 E dition

Bulletin 1983 (Bureau of Labor Statistics, 1977). Output measure
based on Bureau of Mines data.




201

of Energy, Energy Information Administration, 1977), ch. 6, pp.
127-53.
’
1 Bureau of Mines data.
1

Table 5.

suited in more domestic processing of residual fuel—
from 30 percent of domestic demand in 1973 to
about 56 percent in 1977. Imports of residua! fuel hit
a low of 1.2 million barrels a day in 1975 and have
not increased greatly since then, in spite of a sizable
increase in demand.
The outlook for imports of refined products is not
clear.
Opinions differ as to domestic capacity
growth and shifts in demand, aside from the avaiSa-,
biJity. of crude supplies or, in. the longer, run, the
possibility that oil-producing countries will move
into refining.

Average annual percent change1
Indicator
1960-75

1960-66

1966-75

Payroll per unit of value added ..

-3.2

-6.3

-1.0

Capital expenditures per produc­
tion worker .............................

13.4

9.1

10.7

1 Linear least squares trends method.
SOURCE: Bureau of the Census.

Productivity doff@r©pc©s
Data on productivity differences among establish­
ments in an industry with a high degree of speciali­
zation may provide some insight into the faqtors
associated with high productivity performance within
the industry. In a study of 1967 Census d a ta ,14 pe­
troleum refineries w eres ranked by value added per
production worker hour to provide a rough indica­
tion of the range of productivity differences. In this
industry, average value added per production worker
hour in the highest quartile was almost 11 times
greater than in the lowest quartile.
Wide productivity differences in the refining indus­
try may reflect differences in size, management,
complexity (type of processing), labor, capital out­
lays,-etc., but the limited data preclude general con­
clusions. Nevertheless, the 1967 data suggest that
size may be important (table 6). Establishments in
the highest quartile had an average employment
almost four times greater than those in the lowest
quartile. This is verified by studies15 which show
that labor productivity increases with capacity and
with employment, up to a point. Small plants must
maintain a minimum staff of operators and mainte­
nance and technical personnel to run the refinery; as
capacity increases, the number of production work­
ers needed per thousand barrels of output declines
sharply. But at some point, the advantages of size
may be offset by duplication of process units.

Productivity
Productivity in refining rose sharply in the postWorld War SI period. From 1960 to 1977, output per
employee hour in the refining industry increased at
an average rate of 4.3 percent annually, compared
with 2.6 percent in all manufacturing. 15
Productivity growth was considerably more rapid
and’ steadier, however, from 1960 to 1967 (7.1 per­
cent) than from 1967 to 1977 (3.0 percent).
Sn the last few years of the 1960’s, productivity
growth leveled off at a relatively low rate. In 1972
and 1973, productivity rose very sharply; this was
followed by a sizable decline in 1974. While the re­
covery since then has been moderate, productivity in
1977 was back to the high level of 1973. These errat­
ic productivity movements were associated with the
embargo and the events that followed. There were
erratic changes in refining output, discussed above,
'such as the unusually steep increases in output in
4973, 1976 and 1977, and the decline in 1974, the
first since 1958. There were also unusual changes in
employee hours.
A roughly similar pattern of change in the industry
since 1960 is evident in data on payroll per unit of
value added, i.e., labor as a percent o f the value of
shipments less materials and other costs. As shown
in table 5, payroll per unit of value added fell at an
average annual .rate of 3.2 percent from 1960 to
1975, compared with 1.1 percent for all manufactur­
ing, indicating a relatively greater increase in effi­
ciency. The stronger industry position in the first
half of the period 1960 -6 6 is evident in the sharp
decline in payroll/value added of almost 6 1/2 per­
cent annually. In contrast, the ratio showed only a
minor change of about 1 percent'in the 1966-75 per­
iod, having registered sizable increases for several
years.1
3
2

CapoteS expenditures
Capital expenditures for refining plants increased
11.2 percent annually from 1960 to 1975 to a total of
$2.2 billion—more than four and one-half times the
outlay in 1960. The increase, however, was not even
over those years. In the first half of the 1960’s, an­
nual outlays declined or remained relatively con­
stant; from 1965 to 1971 they rose extremely rapidly;
14 Based on unpublished data prepared by the Bureau of the
Census for the National Center for Productivity and Quality of _
Working Life.
15 Studies by W .L. Nelson published in the Oil and Gas Jour­
nal. See “ Maintenance Material and Labor” , Jan. 13, 1975, pp.
5 7 -59.

12P rq jp ctio n s o f E n e rg y S u p p ly ar]d f b r r j ^ d , p. ,137^

13 Productivity Indexes, 1977 Editioni plpl 7 5 -1 6 .




Iradieatoirs of change in peSiroSeum refining, 1960-75

202

small refineries, 19 of these new plants had less than
a 10,000-barrel daily capacity. Only one had a capac­
ity of more than 40,000 barrels.
There is no general agreement on the outlook for
capital expenditures for expansion. In addition to
judgments on the need for additional capacity, capi­
tal outlays for expansion will be influenced by the
increasingly heavy costs of new plant and equip­
ment. In general, however, there is agreement on the
necessity to modify existing facilities to cope with
changing demand and supply conditions. Even in
this, there is a wide range of views relating to the
future course of gasoline demand and likely develop­
ments in, coal gasification and liquefaction, In addi­
tion, future government environmental and energy
policies will affect capital outlay decisions. Of great
concern is the increase in environmental protection
costs, which averaged 12 percent of the total petro­
leum industry’s outlay in 1975;18 data for the refining
sector alone are not available.

Table 6. Value added and employment in petroleum refin­
ing: Ratios of “highest quartile” to “lowest quartile” plants
and to average plant, 1967

Measure

Ratio of highest
Ratio of highest
quartile to lowest quartile to aver­
age
quartile

Value added per production
worker hour..........................

10.7

1.8

Average employment per establishment.................................

3.7

1.5

NOTE: Establishments Were ranked by the ratio of value added per
production worker hour.
SOURCE: Based on unpublished Census Bureau data prepared for
the National Center for Productivity and Quality of Working Life.

in 1972 and 1973 they declined; and in 1974 and
1975 they jumped very sharply. The average
outlay in the I960— period was $917 million.
75
These data, however, reflect costs unadjusted for
changes in prices. Adjusting the dollar figures by the
Nelson index of refinery construction costs16 reveals
that real investment barely doubled from 1960 to
1975. From 1966 to 1975, real investment rose one
and one-half times compared with three and one-half
times for dollar outlays. However, increases in re­
finery costs were offset by the greater efficiency of
plant and equipment. When adjusted for productivity
changes by Nelson’s “ true cost” index, adjusted
real capital outlays rose two and one-half times in
those 9 years, and almost three and one-half times
from 1960.
Petroleum refining is highly and increasingly capi­
tal intensive. Labor costs were less than )3 percent
of the value of the product in 1975 (compared with
48 percent in all manufacturing), having dropped
sharply and steadily from 34 percent in, i960. As
capital expenditures rose sharply and the number of
production workers declined from 1960 to 1975, capi­
tal outlays per production worker rose almost seven­
fold1 After adjustment for price and productivity in­
.
crease, real outlays per production worker rose al­
most fivefold.
These large capital expenditures resulted in addi­
tions to daily capacity of 3.2 million barrels, an in­
crease of 24 percent in the 5 years from January
1972 to January 1977. From January 1974 to January
1977, 28 “ grass roots’’ plants were built, accounting
for slightly over 20 percent of the total.increase in
operating capacity in those years.17 However, almost
all of these were very small, of very simple design
and limited flexibility. With incentives available to

Employment and Occupational Trends
Employment
A bout 160,300 people were employed in the refining
industry in 1977, the largest number since 1962. A
decline starting in the late 1940’s continued unabated
through the m id-1960’s, reflecting the very sharp in­
crease in productivity through m ost o f the period. After
1973, however, employment turned up again.

In the first half of the 1960’s, the sizable employ­
ment decline was associated with a sharp reduction
in the number of refineries and a productivity growth
rate which was more than double the rate of growth
of output, forom the mid-1960’s to 1973, employment
was relatively stable, although it dipped to a low
point of 145,(X ) in 1969. After 1973, however, sever­
X
al years of rising employment brought the level up to
that of the early 1960’s. This change in the direction
of employment reflected, in addition to technology
changes which required more unit labor, an increase
in the number of very small refineries. Overall, from
1960 to 1977, a relatively moderate annual average
employment decline of 0.3 percent was registered.
Refinery employment to 1985 /is projected to re­
sume its decline. Based on the economic assump­
tions stated in the introduction, the BLS projects a
decline to 137,000 employees in 1985, or a drop of
1.9 percent annually from 1977 to 1985.
These data reflect the technological, structural,
and skill changes which have affected employment in

16
Nelson Index published in the Oil and Gas Journal. See issue
of Jan. 26, 1976.
* Trends in Refinery Capacity and Utilization (Federal Energy
7
Administration),June 1976, pp. 4 and 7, and June 1977, p.i 14.




203

18 Data from the Bureau of the Census.

might have had a ratio of 2 maintenance workers to
1 operator. Now, (excluding contract workers, the
ratio of maintenance workers to operators in that
refinery may be 1 to 2.

the industry. A modern refinery today with an input
of 100,O X barrels per day employs about 300-350
C)
workers on three shifts. An older refinery with that
capacity which has been modernized employs about
700 workers; that same plant would have employed
almost 1,000 persons in the 1950’s.
Occupations
As discussed earlier, technological and structural
changes are altering traditional concepts of job con­
tent and duties. More importantly, duties are being
consolidated, as in the ca'se of maintenance crafts, or
partially removed from the refinery, as in the case of
contract maintenance.
Maintenance craft consolidation is an important
labor development of the last decade which increas­
es the flexibility of the work force while it reduces
the number of workers required per processing unit.
Under most maintenance consolidation plans, skilled
workers who have attained journeyworker status in
one craft are trained to handle other crafts (for ex­
ample, a boilermaker who learns pipefitting), thus
eliminating the need for several workers, each with a
specific craft duty. Such consolidation is becoming
more widespread. Of 104 refineries studied by BLS
in 1976, about one-fourth reported craft consolida­
tion plans, double the number reported in 1965.19 ln
most plants, consolidation was limited to two desig­
nated crafts but in many plants consolidation incor­
porated all maintenance crafts. These skill combina­
tions fall into a single job classification, “ general
mechanic.’’ A further development of this practice is
the combination of operative and maintenance skills
by one worker, who may be known as a “ running
operator.” Two running operators can handle a pro­
cessing unit of 1(H),OCX)-barrel capacity, compared to
three operators and a maintenance worker required
in the average refinery of similar capacity.
The trend to maintenance craft consolidation may
in time contribute most importantly to revising job
content and standard occupational patterns. By elim­
inating the lines of craft duties, craft consolidation
practices generally establish new single job classifi­
cations with new duties and training.
Contract maintenance is performed by workers
supplied by outside firms on a contract basis, and
permits a refinery to have a relatively small yearround maintenance staff. Although contract workers
are generally used for special peak work periods
such as during shutdown, they may also be em­
ployed year round on regular maintenance. Prior to
the practice of contract maintenance, a refinery

Hie practice of contracting out is reducing employee
hour requirements in the refinery ^buf data are not
available on its extent. The importance of this practice
is evident in the fact that in this industry 9 of 13 labormanagement contracts (covering 1,000 workers or
more) studied by the BLS in 1975 contained provisions
limiting subcontracting.20
Job and skill changes. lob content and skill re­
quirements are being substantially changed by the
sophisticated instrumentation, particularly for
maintenance and lab technicians. In many cases,
decisionmaking is being transferred from the em­
ployee to the machine. In the older plants, techni­
cians may take readings of various instruments ev­
ery 4 hours and may record the information manual­
ly. In modern refineries, computer consoles in each
process unit record the data from on-line instru­
ments and feed the data into a central computer sys­
tem. As many as a thousand signals can be received
by the computer and checked against limits for prob­
lem areas. In blending, for example, several prod­
ucts must go into the finished gasoline in proper
proportions. To do this, the operator merely sets the
controls for the specified percentages, etc. Samples
from the stream are then automatically analyzed;
changes can be made by the operator. The computer
will continuously monitor the process and report
information on demand.
Manual skills, already at minimum levels in the
refinery, continue to decline. Even the truckdriver
who loads the gasoline for delivery uses an automat­
ed system. The driver removes the cover of the
tank, puts a punchcard into a slot, and pushes a but­
ton. The required quantity of the correct product
fills the truck, after which the flow shuts off auto­
matically. When this job is not automated, loaders
and pumpers are employed to do the work.
Due to the many changes occurring in the indus­
try, the occupational distribution in 1985 is expected
to be significantly different from the 1970 pattern. In
the BLS projections of occupational employment,
professional and technical workers, more than onefifth of all refinery workers, are expected to decline
moderately from 1970 to 1985 but to retain about the
same share of total em