View original document

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

Li.3 :1H-6~S



Bulletin No. 1458
W. Willard Wirtz, Secretary
Arthur M. Ross, Commissioner




Bulletin No. 1458
W. Willard Wirtz, Secretary
Arthur M. Ross, C om m issioner

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C., 20402 - Price $1.50


Sixteen years have passed since the first effort was made to put together into
one handbook statements of the methods used by the Bureau of Labor Sta­
tistics (B L S) in preparing its major statistical series. The policy of telling
users how B L S figures are obtained was in existence long before the first
handbook was published (Bulletin 933, 1950). The Bureau underlined this
policy again a few years later in an expanded and revised edition of the
1950 bulletin (Bulletin 1168, 1955). During the decade since the second
edition, new series have been added and surveys undertaken to widen and
sharpen our understanding of the economic and social forces molding our
national destiny. Over the same period, new and better ways of doing the job
of measurement and factfinding have been found. It is time to bring the
handbook up to date and to describe these new techniques and new surveys.
The present volume attempts to do this in a simple and direct manner, so as to
make more useful and understandable the factual picture presented by the
Bureau’s figures.
For each major program there is a brief account of how it came into being
and what it attempts to do. Where the original data come from is noted,
terms are defined, and the concepts adopted are outlined. The statements
which follow describe how the data are put into final form, tell how the results
may be used, and state the limitations to be kept in mind in using the informa­
tion. Occasionally, in the interest of being precise, there is added a form, a
table, or a mathematical formulation. The purpose is always to tell the user
of our statistics how and why the Bureau does the job. Since this volume
permits only brief treatment of each subject, lists of other publications which
have more complete information, or which can furnish additional insights
into the subject matter are provided at the end of most chapters.
B L S statistics are used for many purposes, and sometimes the data ideally
suited to one purpose may have limitations for another. The user, confronted
with a specific problem, must judge whether the accuracy and reliability of
the selected statistics are adequate to meet the needs of his problem. The
chapters on techniques used in preparing B L S series and conducting B L S
surveys contain the information necessary for evaluating the fitness of the
statistics for the various uses to which they may be put.
This handbook was prepared by the staff of the Bureau of Labor Statistics
under the general direction of Bobert J . Myers, Deputy Commissioner, Robert
B. Steffes coordinated the preparation of manuscripts and provided technical






Manpower and employment statistics
1. Labor force, employment, and unemployment____________
2. Employment, hours and earnings_______________________
3. Labor turnover______________________________________
4. Employment of scientific and technical personnel__________
5. Occupational outlook_________________________________
6. Projections of the labor force__________________________
7. Industry-occupational matrix__________________________


Prices and living conditions

Consumer expenditures and income_____________________
Standard budgets____________________________________
Consumer prices_____________________________________
Wholesale prices_____________________________________
Industry-sector indexes_______________________________
Daily spot market prices______________________________


Wages and industrial relations

Occupational pay and supplementary benefits_____________
Employee earnings and hours frequency distributions______
Union wage scales___________________________________
Current wage developments___________________________
Employer expenditures for supplementary compensation
19. Work stoppages_____________________________________
20. Collective bargaining agreements_______________________
21. Union membership___________________________________


Productivity and technological developments

Output per man-hour measures: Private economy_________
Output per man-hour measures: Industries_______________
Technological change_________________________________
Construction labor requirements________________________


Industrial safety
26. Frequency and severity of work injuries_________________
27. Work injuries in maritime activities_____________________
28. Injury rates and accident causes________________________




Other programs
29. Foreign labor conditions and international comparisons_____
30. Economic growth studies______________________________
Appendix A. The BLS seasonal factor method___________________
Appendix B. Industrial classification___________________________
Appendix C. Geographic classification__________________________



BLS Handbook of Methods for Surveys and Studies
The country is hungry for information; everything of a
statistical character, or even of a statistical appearance
is taken up with an eagerness that is almost pathetic;
the community have not yet learned to he half skeptical
and critical enough in respect to such statements.

With these words Gen. Francis A. Walker
greeted Carroll Wright in 1873, as Dr. Wright
assumed charge of the Massachusetts Bureau of
Statistics and Labor. And when as TJ.S. Com­
missioner of Labor, he issued his first annual re­
port in March 1886, Carroll Wright established the
policy of explaining his statistical methods to his
readers and of seeking to avoid misinterpretation
of the figures presented. For example, he said:
In stating the facts as they have been found by the
agents of the Bureau, many terms are used which are ca­
pable of varied application— some even are of doubtful
meaning when considered metaphysically, but all such
terms are used in this report in their common acceptation;
as, for instance, the term “overproduction” is used to indi­
cate that condition of a locality, state, or country when
more goods have been produced than are sufficient to
meet the ordinary dem and............1

In the same report there are statements on test­
ing the validity of figures (p. 141), problems of
nonresponse (p. 90), and restrictions on coverage
(headnotes to tables). Warnings as to inade­
quacies of available information occur frequently.
During the 8 decades which followed the initial
report, the definitions, methods, and limitations of
the data published by the Bureau of Labor and its
successor, the Bureau of Labor Statistics, have
been explained again and again. The reason for
this is not merely to make the readers “skeptical,”
“critical,” and aware of the known limitations of
the statistics, but also to instruct them in the
proper use of the information and to assure them
that proper standards have been observed.
Furthermore, whereas one might expect to breed
a certain amount of doubt about a statistical sur­
vey by revealing its lack of perfection, frankness
about unavoidable defects more often has the

opposite effect, and public confidence in the work
is reinforced in the process. The most grave
doubts arise when things crying for explanation
are not explained.
The Committee on Government Statistics and
Information Services emphasized 25 years ago
that the Central Statistical Board “should urge
on each collecting agency the importance of pub­
lication by agencies of frank appraisals of the
extent of noncomparability, incompleteness, and
inaccuracy which may be inherent in their reports
at any given time. This candid policy should
enhance and not diminish the scientific prestige
of the collecting agency.” 2
Full understanding of the statistical series and
studies of the Bureau of Labor Statistics is not to
be gained solely from detailed descriptions of
them, but also from appraisal of the philosophy
and approach of the Bureau and of the manner in
which it functions.

The history of the Federal Bureau of Labor ex­
tends back to 1884. Before the creation of the
cabinet post of Secretary of Labor, the Bureau
for a time was known as the Department of Labor.
From the Bureau’s beginnings in the administra­
tion of President Arthur until it became a part
of cabinet department under President Wilson,
it accumulated nearly 3 decades of experience in
collecting, interpreting, and presenting facts cru­
cial to the welfare of workers. Details of early
1 See Industrial Depressions, The F irst Annual Report of the
Commissioner of Labor3 March 1886, pp. 15-16.
2 Committee on Government Statistics and Information Services,
Government Statisticsj April 1937, p. 53. Recommendation # 1 6
of the Committee sta te s: “ Continued criticism and analysis
should be made of (a) statistical definitions, specifications, and
classification; (b) coverage of supposedly complete surveys and
of samples used for current reporting; (c) timing of periodic
surveys and current reports, component items and weighting
systems of index numbers; and ( f ) methods and practices in the
presentation of data. . . . Frank appraisals of comparability,
completeness, and accuracy should be published.” (pp. 48-49).




Bureau history and of developments of later years
are to be found elsewhere.3 However, in describ­
ing the various statistical programs 4 in this vol­
ume, some of the events which led to the develop­
ment of particular statistical measures are re­
counted. Against this historical background
emerges the philosophy and posture of the Bureau
of Labor Statistics as the impartial observer and
interpreter of trends important to the welfare of
workers. Voluntary reporting and the preserving
of the confidential nature of reported data are
important characteristics of B L S programs.

Voluntary Reporting and
In the 80-year history of the Bureau’s operation
it has asked hundreds of thousands of firms and
individuals to provide information closely related
to their daily affairs and their personal lives. To
some of them who have supplied the desired infor­
mation, the Bureau has gone back a second time,
a third time, and perhaps dozens of times, for later
information on the same subject or for new types
of information. The response has been remark­
able in its generosity, even when it is remembered
that a sustained effort has been made to keep the
requests reasonable. In no small measure, the
cooperation received is due to the great care taken
to avoid identifying the firm or the person sup­
plying the information. The fact that Bureau
employees pledge themselves to protect these data
is less important than that they have a deep under­
standing of the adverse longrun consequences of
even a single lapse. They are aware of the greater
worth, in terms of pure statistical validity, of the
information provided voluntarily as compared
with that supplied under legal sanctions. The
only inducement employed is to tell the respondent
that his contribution is important to the ultimate
success of the survey and that he may find the sur­
vey results useful in his own pursuits. The policy
of not identifying the respondent is implemented
by combining the data reported by the different
sources and issuing the findings in summary form.
Another assurance given the respondent is that
his report will be used for statistical purposes only.

Attempts to “break” this policy, by organizations
or individuals who wanted access to data in our
possession and were willing to go to the courts to
secure it, have been successfully resisted.5 An­
other form this problem takes is the case in which
an administrative agency of government seeks
court action to compel a company to release its file
copy of information provided in confidence to a
statistical agency.6
While it cannot be proved that these policies
result in more reliable statistics, Bureau Commis­
sioners and their staffs over the years have been
convinced from their experience that it is so. It
is notable that some other Federal agencies (es­
pecially the Bureau of the Census), well-equipped
with authority to compel the submittal of certain
reports, rarely if ever invoke this power. Bather,
they choose to rely upon forms of suasion similar
to our own. The Bureau of Labor Statistics,
while its functions as a statistical agency are pre­
scribed by law,7 has always relied upon voluntary
cooperation of respondents in collecting

BLS Role, Staff and Organization
Among Federal agencies collecting and issuing
statistics, the Bureau of Labor Statistics has been
termed a “general-purpose statistical agency.” 8
The Bureau’s figures are prepared to serve the
needs of business, labor, Congress, the general
3 See, for example, the Secretary of Labor’s F irst Annual Report,
1918, for history 1884-1913 ; and U.S. Department of Labor, The
Anvil and the Plow, 1964; pages 4-5, 19-20, 49-51 (1913-30) ;
pp. 63-64, 87-90, 117-119, 136-137, and 155 (1930-48) ; pp. 172173, 187, 206-207, and 230-234 (1949-63).
4 For a handy reference to B L S programs, showing their
principal characteristics, see U.S. Department of Labor, Bureau
of Labor Statistics, Major B L S Programs (issued annually).
*F o r example, see Norwegian Nitrogen Company v. United
States, 288 U.S. 294 ; United States v. Kohler, 13 Fed. Rules Serv.
33.333 (E.D., Pa., 1949) ; Hawes v. Walsh, 277 Fed. 569, the
Court of Appeals for the D istrict of Columbia. In all of these
cases, the courts sustained the policy of protecting the confiden­
tiality of information given voluntarily and in confidence to an
agency of the Federal Government.
6 See Supreme Court of the United States, St. Regis Paper
Company, Petitioner, v. United States, No. 47, October term,
7 Excerpts from 29 U.S.C. 1, acts of June 27, 1884, ch. 127,
23 Stat. 60; June 13, 1888, ch. 389, § 1, 25 Stat. 182; Feb. 14,
1903, ch. 552, § 4, 32 Stat. 826; Mar. 18, 1904, ch. 716, 33 Stat.
136; Mar. 4, 1913, ch. 141, § 3, 37 Stat. 737.
8 Bureau of the Budget, Statistical Services of the United States
Government (revised edition), 1963, p. 7.


public, and especially, the administrative and
executive agencies, for information on economic
and social trends and situations. While the data
serve some administrative purposes, they are free
from the constraints which sometimes result from
the close ties normally existing between operations
and operational statistics. B L S statistics are
often quite specialized, yet they meet general
economic and social data requirements. As the
needs of users are likely to differ from each other
and also over time, no statistic is ideal for all.
This makes it important that the characteristics of
the measures and their possible limitations be well
The Bureau plays a larger and more significant
role than merely publishing general purpose sta­
tistics. Its activities frequently influence, and
sometimes are crucial, to the determining and
shaping of public policy. The Bureau’s experts
have the keen understanding of economic and
social forces which results from intensive and con­
tinuous involvement in factfinding and in the pain­
staking anlysis of data.

The Bureau’s work extends beyond the initial
collection and processing of data. Over the years,
it has developed a staff of professional analysts,
trained in the disciplines of economics and other
social sciences, to search out the implications of
survey findings for the welfare of workers and to
present them as cogently and as promptly as pos­
sible in written and oral form. How successfully
this can be accomplished depends greatly upon the
competence of the analysts and of their supporting
In BLS, analytical and statistical work is per­
formed by e c o n o m i s t s , statisticians, and
mathematical-statisticians with the aid of an ex­
perienced corps of programers, systems analysts,
and other professionals, as well as statistical
clerks. For analytical work, economists at even
the lowest grade level must meet Civil Service
Commission requirements roughly equivalent to a
college major in economics. There are compar­
able requirements for other professionals. The

222—617— 6<



greatest effort is made to locate the best of gradu­
ating seniors, Masters, Ph.D.’s, and those with re­
search experience, in the colleges, State agencies,
business organizations, and labor unions. The
Bureau provides training needed for on-the-job
skills, as background to special assignments, to
keep professionals abreast of changes in their
fields, and to aid higher level and executive pro­
fessionals in obtaining the best results from their
In training staff, a special effort is made to im­
part detailed knowledge of the techniques used in
collecting and compiling the statistics, so that
maximum application of data results to current
problems can be made without a risk of exceeding
the limits of their significance.

The statistical programs of the Bureau were de­
veloped, for the most part, independently of each
other, taking on characteristics suited to the re­
quirements of the subject under observation. As
a result, the Bureau was organized according to
subject-matter areas, an arrangement which has
proved efficient and has been continued over the
years. Expertise in techniques, economic analysis,
and other staff activities across subject-matter lines
were added to provide better utilization of the
Bureau’s resources.
As the Bureau’s collection activities increased,
regional offices were established in 1943 to provide
administration of the field programs and staff.
Another function of the offices was to disseminate
data to local users and to furnish technical advice
and assistance to State agencies and other cooperat­
ing organizations. An important aspect of the
work of the regional staffs has been the function of
explaining the concepts and techniques which we
utilize in compiling our statistics.
Special recognition of the competence of the
Bureau in the field of statistics was given by the
Secretary of Labor in 1955 when he delegated to
the Commissioner of Labor Statistics the respon­
sibility for continuously reviewing all of the statis­
tical programs of the Department of Labor and of
making recommendations for their improvement.



Consultation and Advice on Statistical
A statistical program too much detached from
the uses of its data may fail in its principal mis­
sion. To avoid sterility and stagnation, the Bu­
reau continuously invites advice and ideas from
users and experts in business, labor, and academic
organizations and individual members of the pub­
lic. Over the years, the advice the Commissioner
of Labor Statistics has received on policy and
technical matters from responsible parties, relat­
ing to the collection and analysis of our statistics,
has usually been sound and therefore very helpful.
Of course, decisions on statistical policy have al­
ways been the final responsibility of the Commis­
In order to keep in touch with the current and
anticipated needs of business and labor groups
and to seek advice on technical problems, the Com­
missioner first established standing research ad­
visory committees in 1947. These groups, now
called the Business Research Advisory Council
and the Labor Research Advisory Council, serve
in an advisory capacity with respect to technical
problems, consult on Bureau programs, and pro­
vide perspectives on Bureau programs in relation
to needs of their members. The councils accom­
plish their work in general sessions and also
through committees designated to subject-matter
fields on a more specialized basis. Committee
memberships are augmented by other persons in
industry or labor who have special competence
although not council members. The councils may
take formal action through resolutions or recom­
mendations on matters regarded as appropriate
for such action, but such resolutions are merely
advisory. Members of the councils and the sub­
committees serve in their individual capacities,
not as representatives of their organizations.
The members of the Labor Research Advisory
Council are designated by the Commissioner of
Labor Statistics under authorization by the Secre­
tary of Labor, from nominations by the Director
of Research, AFL-CIO. All research directors
of international unions represented in the A F L CIO, the Railway Labor Executives’ Association,
and the railroad operating unions are invited to

attend the general meetings of the council. The
council provides general direction to the advisory
activities of trade union research directors in
relation to the Bureau.
The members of the Business Research Advisory
Council are designated by the Commissioner
under authorization of the Secretary of Labor,
after consultation with the National Association
of Manufacturers, the U.S. Chamber of Com­
merce, and other organizations broadly representa­
tive of American business. Members serve in their
individual capacities, not as representatives of
their companies.
The Bureau often seeks the advice of profes­
sional economists, statisticians, social scientists,
educators, and others, either in their individual
capacities or as members of professional organiza­
tions. This is most likely to occur when a concep­
tual or theoretical question arises which is con­
sidered fundamental to the work of the Bureau
in a specialized field, and where professional ac­
ceptance of the Bureau’s work in that field may be
reinforced by the findings of an independent
It is a fundamental objective of the Bureau that
its statistical practice be built soundly upon estab­
lished statistical theory. The objective can be
realized only if B L S practitioners are trained in
statistics and if they keep their knowledge up-todate. For this reason the Bureau encourages their
participation in activities of the professional
societies, their efforts to improve their education
in statistics, and continuing contacts with other
experts in their disciplines. The efforts of other
statistical organizations, public and private, are
studied unremittingly so that B L S may reap
benefits from the experiences of others.

Standard Definitions
Where related statistics cut across program lines
or across Government bureaus, the Bureau of
Labor Statistics cooperates to the maximum extent
possible in the Bureau of the Budget’s effort to
obtain adherence to standard definitions of terms
for maximum comparability. The use of the
definition of establishment is a case in point.
The business establishment has been found to be
the most satisfactory source of data for most in-


dustrial statistics because it is the first level of
business organization for which complete
records—i.e., production, employment, purchases,
sales, wages, inventories, etc.—are generally main­
tained. The establishment is the primary unit of
organization in the business economy and is the
first integrated level or combination of employees
devoted to the production of a related group of
products or services. In B L S programs in which
it is applicable, the standardized definition of the
establishment is utilized.9
The Bureau also follows the Bureau of the
Budget’s definition of “production and related
workers,” 10 and uses the “standard payroll pe­
riod.” 11 The reader is referred to appendixes B
and C for descriptions of the standards followed


with respect to industrial classification and geo­
graphic classification.
9 See appendix B.
10 Standard Definitions of Types of Workers, Bureau of the
Budget, November 7, 1944. “ Production and related workers are
defined to include working foremen and all nonsupervisory
workers (including leadmen and trainees) engaged in fabricating,
processing, assembling, inspection, receiving, storage, handling,
packing, warehousing shipping, maintenance, repair, janitorial,
watchman services, product development, auxiliary production
for plant’s own use (e.g., powerplant) and record-keeping and
other services closely associated with the above production opera­
tions. Excluded are supervisory employees (above the working
foremen level) and their clerical staffs.”
11 Standard Definition of Payroll Periods for Employment
Reports, Bureau of the Budget, March 28, 1952. “ In order to
maintain a coordinated system of employment reports and to re­
duce the reporting burden on respondents, requests made to em­
ploying establishments for statistical information from their
payroll records on the number of employees, payrolls, hours
worked, or related items, should refer to the payroll period con­
taining the 12th of the month.”

Manpower and Employment Statistics
Chapter 1. Labor Force, Employment, and Unemployment
Each month, the Bureau analyzes and publishes
statistics on the labor force, featuring information
on employment and unemployment classified by a
variety of demographic, social, and economic char­
acteristics. These statistics are derived from the
Current Population Survey, which is conducted by
the Bureau of the Census. The survey is based on
a probability sample of households, representing
the civilian noninstitutional population of the
United States.
Concepts of the labor force, employment, and
unemployment now in use were introduced in the
latter stages of the depression of the 1930’s, chiefly
in the interest of deriving more objective measure­
ments of unemployment and employment than
were previously available. Prior to the thirties,
aside from attempts in some of the decennial cen­
suses, there were no direct measurements of the
number of jobless persons. The development of
mass unemployment in the early thirties made
the need for statistics urgent, and widely conflict­
ing estimates based on a variety of indirect tech­
niques began to appear. Dissatisfied with these
methods, many research groups, as well as State
and municipal governments, began experimenting
with direct surveys of the population or samples
of the population. In these surveys, an attempt
was made to classify the population as employed,
unemployed, or out of the labor force, by means
of a series of questions addressed to each individ­
ual. In most of the surveys, the unemployed were
defined as those who were not working but were
“willing and able to work.” This concept, how­
ever, did not meet the standards of objectivity
that many technicians felt were necessary in order
to measure either the level of unemployment at a
point in time or changes over periods of time.
The criterion “willing and able to work,” when
applied in specific situations, appeared to be too
intangible and too dependent upon the interpreta­
tion and attitude of the persons being interviewed.

Out of this experimentation, a set of concepts
was developed in the late 1930’s which sought to
meet these various criticisms. According to the
new concepts, the classification of an individual
depends principally upon his actual activity
within a designated time period, i.e., was he work­
ing, looking for work, or engaged in other activi­
ties? These concepts were adopted for the na­
tional sample survey of households initiated by
the Works Progress Administration in 1940.
Originally termed the Monthly Report on Un­
employment when the WPA was responsible for
the collection of labor force statistics, the house­
hold survey was transferred to the Bureau of the
Census late in 1943. Its name was changed at
that time to the Monthly Report on the Labor
Force (M RLF), which was also the title of the
Census reports and subsequently the B L S reports
describing the sample results every month. The
survey title was changed once more—in 1948—
to the present “Current Population Survey”
(CPS) in order to reflect more accurately its
expanding role as a source for a wide variety of
demographic and economic characteristics of the
population. In 1959, responsibility for analyzing
and publishing the CPS labor force data was
transferred to the BLS, while the Census Bureau
has continued to collect and tabulate the statistics.

Description of Survey
The Current Population Survey provides sta­
tistics on the civilian noninstitutional population
14 years of age and over. Figures on the Armed
Forces (obtained monthly from the Department
of Defense) are added to the CPS estimates to
derive estimates of the total labor force and the
total noninstitutional population. Persons below
14 years of age are excluded from coverage in the
survey because child labor laws, compulsory school
attendance, and general social custom prevent most
of these children in the United States from work-



ing. The institutional population, which is ex­
cluded from coverage, consists of inmates of penal
and mental institutions, tuberculosis sanitariums,
and homes for the aged, infirm, and needy.1
The CPS is collected each month from a proba­
bility sample of approximately 35,000 households.
Participation in the survey is on a purely volun­
tary basis. Respondents are assured that all
information obtained is completely confidential
and is used only for the purpose of estimating
national totals. The B L S policy of confidenti­
ality is explained more fully in the Introduction.
The time period covered in the monthly survey
is a calendar week. A calendar week was selected
as the survey reference period because the period
used must be short enough so that the data ob­
tained is “current” and the time reference not so
short that the occurrence of holidays or other acci­
dental events causes extremely erratic fluctuations
in the information obtained. A calendar week
fulfills these conditions as well as being a conven­
ient and easily defined period of time. Since July
1955 the week containing the 12th day of the
month has been used.

or private employment agency, writing letters of
application, canvassing for work, etc.—and who,
during the survey week, were awaiting the results
of these efforts, are considered to be looking for
work. Also included as unemployed are those
who did not work at all during the survey week
and (a) were waiting to be called back to a job
from which they had been laid off, (b) were
waiting to report to a new wage or salary job
scheduled to start within the following 30 days
(and were not in school during the survey week),
or (c) would have been looking for work except
that they were temporarily ill or believed no work
was available in their line of work or in the


classified as employed, unemployed, or in the
Armed Forces are defined as “not in the labor
force.” These persons are further classified as
“engaged in own housework,” “in school,” “unable
to work” because of long-term physical or mental
illness, and “other.” The “other” group includes,
for the most part, retired persons, individuals re­
ported as too old to work, the voluntarily idle, and
seasonal workers for whom the survey week fell
in an “off” season and who were not reported as
looking for work. Persons doing only incidental
unpaid family work (less than 15 hours in the
specified week) are also classified as not in labor
The category “not in labor force—in school” in­
cludes persons attending school during the sur­
vey week who had new jobs to which they were
scheduled to report within 30 days. All persons—
whether or not attending school—who had new
jobs not scheduled to begin until after 30 days
(and not working or looking for work) are classi­
fied as not in labor force.

The criteria used in classifying persons on the
basis of their labor force activity are as follows:
Employment. Employed persons comprise (1) all

those who, during the survey week, did any work
at all as paid employees, or in their own business,
profession, or farm, or who worked at least 15
hours as unpaid workers in a family-operated en­
terprise and (2) all those who were not working
or looking for work but who had jobs or businesses
from which they were temporarily absent because
of illness, bad weather, vacation, labor-manage­
ment dispute, or various personal reasons. E x ­
cluded from the employed group are persons whose
only activity consisted of work around their own
home (such as housework, painting, repairing,
etc.) or volunteer work for religious, charitable,
and similar organizations.
Unemployment. Unemployed persons include
those who did not work at all during the survey
week and were looking for work. Those who had
made efforts to find work within the preceding
60-day period—such as by registering at a public

Labor Force . The civilian labor force comprises

the total of all civilians classified as employed or
unemployed in accordance with the criteria de­
scribed above.
The total labor force includes
members of the Armed Forces stationed either in
the United States or abroad.
Not in Labor Force . All civilians who are not

1 For a fuller explanation of tlie Current Population Survey,
see Concepts and Methods Used in Household Statistics on Em­
ployment and Unemployment from the Current Population Sur­
vey, United States Department of Commerce and United States
Department of Labor, 19G4.



Problems of Concept. The basis of the labor force
classification used in the Current Population Sur­
vey is the activity of an individual during a par­
ticular calendar week each month. Obviously, a
person could have engaged in more than one activ­
ity during the period. Thus, in classifying per­
sons, it is necessary to assign a priority to the
various activities for which information was ob­
tained. In this way, an individual is classified in
only one group and unduplicated totals of the
employed, the unemployed, and persons outside
the labor force can be obtained.
In this classification system, the highest priority
is assigned to “working.” Thus, if a person did
any work—as defined in the concepts—during the
survey week, he is classified as “at work” and is
included with the employed, even though he may
also have looked for work, gone to school, or done
something else.
The activity “looking for work” is given second
priority in the classification scheme. In defining
the unemployed, a slight departure is made from
a strict “activity” concept to include certain
groups who, although not actively looking for
work in the specified week, report that they would
have been doing so except for certain special cir­
cumstances, such as illness or waiting to be called
back to a job from which they had been laid off.
Some modification of the “activity” concept has
been made also in the case of the employed. I f
the only criterion considered were activity during
a calendar week, large numbers of persons who
have definite job attachments but were temporar­
ily absent from work in the survey week for rea­
sons such as illness, vacation, or bad weather,
would be excluded from the labor force count. Be­
cause their absence probably would not exceed a
week or two, their exclusion from the labor force
might result in an unrealistic count of the economi­
cally active population. Therefore, a third cate­
gory is set up within the priority system. This
category consists of persons who were neither
working nor looking for work but who had jobs
or businesses from which they were temporarily
absent because of a labor dispute, illness, vacation,
bad weather, or a related reason during the survey
week. This group, “persons with jobs but not at
work,” is measured separately but is added to the
“at work” group to derive estimates of the total
number of employed persons.

The Survey Design
The Current Population Survey sample is lo­
cated in 357 sample areas comprising 701 counties
and independent cities with coverage in every
State and the District of Columbia. In all, some
40.000 housing units and other living quarters are
designated for the sample at any time, of which
about 35,000, containing about 80,000 persons 14
years and over, are occupied by households eligible
for interview. The remainder are units found to
be vacant, converted to nonresidential use, contain­
ing persons with residence elsewhere, or ineligible
for other reasons. Of the occupied units eligible
for enumeration, about 4 to 6 percent are not inter­
viewed in a given month because the residents are
not found at home after repeated calls, are tempo­
rarily absent, or are unavailable for other reasons.2
Selection of Sam ple A reas . The entire area of the
United States consisting of 3,128 counties and in­
dependent cities is divided into 1,913 primary
sampling units. With some minor exceptions, a
primary sampling unit (PSU) consists of a county
or a number of continguous counties. Each of the
212 standard metropolitan statistical areas
(SM SA’s) 3 in existence at the time of the 1960
Census constitutes a separate PSU. By combin­
ing counties to form P SU ’s, greater heterogeneity
is accomplished. Moreover, another important
consideration is to have the P SU sufficiently com­
pact in area so that, with a small sample spread
throughout, it can be efficiently canvassed without
undue travel cost. A typical primary sampling
unit, for example, includes both urban and rural
residents of both high and low economic levels
and encompasses, to the extent feasible, diverse oc­
cupations and industries.
The P SU ’s are grouped into 357 strata. Among
these P SU ’s, 107 of the largest standard metro­
politan statistical areas (including all with over
250.000 inhabitants) and five other areas (not
SMSA’s) are separate strata representing them­
2 For a more detailed description of the sample design and
other technical phases of the CPS program, see U.S. Bureau of
the Census, Technical Paper No. 7, “ The Current Population Sur­
vey— A Report on Methodology,” (1963).
8 See appendix C.


selves. In general, however, a stratum consists of
a set of PSU ’s as much alike as possible in various
characteristics such as geographic region, popula­
tion density, rate of growth in the 1950-60 decade,
percentage nonwhite, principal industry, and type
of agriculture.
Except for the 112 areas mentioned above, each
of which is a complete stratum, the strata are es­
tablished so that their sizes in terms of 1960 popu­
lation are approximately equal. From each stra­
tum a single PSU is selected to represent the en­
tire stratum. In the 112 strata in which there is
only a single PSU (the 107 SMSA’s and 5 special
cases), the single P SU automatically falls in the
sample. Where the stratum has more than one
PSU, the sample PSU is selected in a random man­
ner in such a way that its probability of selection
is proportionate to its 1960 population. For ex­
ample, within a stratum the chance that a PSU
with a population of 50,000 would be selected is
twice that for a unit with a population of 25,000.
Selection of Sam ple Households. The sample de­

sign calls for a sampling ratio which depends on
the predetermined total sample size. For 1965, it
was 1 household for every 1,662 households in each
stratum. The sampling ratio used within each
sample PSU depends on the proportion that the
population of the sample area was of the stratum
population at the time of the 1960 Census. In a
sample area which was one-tenth of the stratum,
the within-PSU sampling ratio which results is
1 in 166.2, thereby achieving the desired ratio of 1
in 1,662 for the stratum. For each PSU which is
a stratum representing only itself, the sampling
ratio is 1 in 1,662 regardless of the size of the PSU.
Within each of the 357 sample P SU ’s, the num­
ber of households to be enumerated each month is
determined by the application of the within-PSU
sampling ratio rather than through the assign­
ment of a fixed quota. This procedure makes it
possible to reflect, on a current basis, population
changes within the sample area. Consequently,
the sample as a whole properly reflects the chang­
ing distribution of the population and avoids the
distortion which would result from the applica­
tion of fixed quotas of households, or persons,
based on the population at an earlier date.
Within each designated PSU, several stages of
sampling may be used in selecting the units to be


enumerated. The first step is the selection of a
sample of Census enumeration districts (ED ’s),
which are administrative units used in the 1960
Census and contain, on the average, about 250
households. These are selected systematically
from a geographically arranged listing, so that the
sample ED ’s are spread over the entire PSU. The
probability of selection of any one ED is propor­
tionate to its 1960 population.
The next step is to select a cluster of approxi­
mately 6 households to be enumerated within each
designated ED. This is done, wherever possible,
from the list of addresses for the ED compiled dur­
ing the 1960 Census or, if the addresses are incom­
plete or inadequate, by area sampling methods.
The address lists are used in about two-thirds of
the cases, primarily in urban areas, and area sam­
pling is applied in the remainder. In using the
Census lists, a systematic sample of clusters of 18
consecutive addresses is selected from the ED and
every third address within the cluster is designated
for the current sample. This provides a slightly
more reliable sample than would be the case for a
cluster of 6 consecutive units. The remaining 12
units in the larger cluster are available for future
The list sample is supplemented by a selection
of the appropriate proportion of units newly con­
structed in the P SU since the Census date. The
addresses of these units are obtained mainly from
records of building permits maintained by the
offices responsible for issuing permits in that area.
A special procedure for updating parts of the
census lists is also followed to reflect units missed
in the Census or new construction in areas where
there is no adequate system of building permits.
In those enumeration districts where area sam­
pling methods are used—mainly rural areas—the
ED ’s are subdivided into segments, that is, small
land areas having well-defined boundaries and in
general an expected “size” of about 6 housing units
or other living quarters. For each subdivided
enumeration district, one segment is designated for
the sample, with the probability of selection pro­
portionate to the estimated “size” of the segment.
Prior to 1962 when the 357-area sample was de­
signed, area sampling methods were used through­
out. The change to list sampling, where feasible,
resulted in reduced sampling error; it is also a
more economical procedure.



Part of the sample is
changed each month. A primary reason for ro­
tating the sample is to avoid the problems of un­
cooperativeness which arise when a constant
panel is interviewed indefinitely. Another rea­
son for replacing households is to reduce the cumu­
lative effect of biases in response which sometimes
occur when the same persons are interviewed in­
definitely. To accomplish this rotation of the
sample on a gradual basis, maps and other ma­
terials for several samples are prepared simul­
taneously. For each sample, eight systematic
subsamples (rotation groups) of segments are
identified. A given rotation group is interviewed
for a total of 8 months, divided into two equal
periods. It is in the sample for 4 consecutive
months one year, leaves the sample during the fol­
lowing 8 months, and then returns for the same 4
calendar months of the next year. In any 1 month,
one-eighth of the sample segments are in their first
month of enumeration, another eighth are in their
second month, and so on, with the last eighth in
for the eighth time (the fourth month of the sec­
ond period of enumeration). Under this system,
75 percent of the sample segments are common
from month to month and 50 percent from year to
year. This procedure provides a substantial
amount of month-to-month and year-to-year
overlap in the panel (thus reducing discontinuities
in the series of data) without burdening any
specific group of households with an unduly long
period of inquiry.
Rotation of Sample.

Collection Methods
Each month, during the calendar week contain­
ing the 19th day, interviewers contact some re­
sponsible person in each of the sample households
in the Current Population Survey. At the time of
the first enumeration of a household, the inter­
viewer prepares a roster of the household members,
including their personal characteristics (date of
birth, sex, race, marital status, educational attain­
ment, veteran status, etc.) and their relationship
to the household head. This roster is brought up
to date at each subsequent interview to take ac­
count of new or departed residents, changes in
marital status, and similar items. The informa­
tion on personal characteristics is thus available
each month for identification purposes and for

cross-classification with economic characteristics
of the sample population.
At each monthly visit, a questionnaire is com­
pleted or revised for each household member 14
years of age and over. The interviewer asks a
series of standard questions on economic activity
during the preceding week (the calendar week con­
taining the 12th day of the month, called the
“survey week”). The primary purpose of these
questions is to classify the sample population into
the three basic economic groups—the employed,
the unemployed, and those not in the labor force.
Additional questions are asked each month to
help clarify the information on employment status.
For the employed, information is obtained on
hours worked during the survey week, together
with a description of the current job. For those
temporarily away from their jobs, the enumerator
records their reason for not working during the
survey week and whether or not they were paid
for their time off. For the unemployed, he
records (1) the length of time they have been look­
ing for work, (2) whether they are seeking full- or
part-time work, and (3) a description of their last
full-time civilian job. For those outside the labor
force, their principal activity during the survey
week—whether keeping house, going to school, or
doing something else—is recorded.
The questionnaires containing the information
obtained for each person in the sample are sub­
jected to a field edit by clerks in each of the 12
regional offices of the Census Bureau. The field
edit serves to catch omissions, inconsistencies, il­
legible entries, and errors at the point where
correction is still possible. Many of the error cor­
rections made in the field edit prevent delays in
further processing of the questionnaires in
After the field edit, the questionnaires are for­
warded to the Washington office of the Census
Bureau. All of the questionnaires are received in
Washington by the end of the week after enumera­
tion. The raw data are transferred to computer
tape and checked for completeness and consistency.
The interviewers on the Current Population
Survey are chiefly part-time workers, although
most of the staff at any time consists of persons
who have had several years experience on the
survey. They are given intensive training when
first recruited and have either direct or home study


training each month prior to the survey. More­
over, through editing of their completed question­
naires, repeated observation during enumeration,
and a systematic reinterview of part of their
assignments by the field supervisory staff, the work
of the interviewers is kept under control and errors
or deficiencies are brought directly to their

Estimating Methods
To increase the reliability of the labor force
statistics derived from the sample, the estimation
procedure uses two stages of ratio estimates and
a “composite estimate5’. It is possible to achieve
this rather complicated procedure rapidly and
automatically because of the availability of high­
speed electronic digital computers. The principal
steps involved are as given below.
Adjustment for Households Not Interviewed .

The weights for all households interviewed are
adjusted to the extent needed to account for units
occupied by persons eligible for interview but for
which no interview was obtained because of
absence, impassable roads, refusals, or other rea­
sons. This adjustment is made separately by
groups of P SU ’s and, within these, for each color
(white, nonwhite) and residence group of house­
holds (urban, rural nonfarm, rural farm). The
adjustment is made separately within each pair of
rotation groups (the incoming pair, the two con­
tinuing pairs, and the outgoing pair).
Ratio Estim ates . The distribution of the popula­

tion selected for the sample may differ somewhat,
by chance, from that of the Nation as a whole in
such basic characteristics as age, color, sex, and
farm-nonfarm residence, among other things.
These particular population characteristics are
closely correlated with labor force participation
and other principal measurements made from the
sample. Therefore, some of the sample estimates
can be improved substantially when, by appropri­
ate weighting of the original returns, the sample
population is brought as closely into agreement as
possible with the known distribution of the entire
population with respect to these characteristics.
Such weighting is accomplished through two
stages of ratio estimates as follows:


1. F irst stage . The first stage of ratio estimates
takes into account differences in the distribution
by color and residence of the population estimates
from the sample PSTJ’s and that of the total popu­
lation in each of the four major regions of the
country. However, independent distributions of
the total population by residence, cross-classified
by color, are not available on a current basis.
Therefore, using 1960 Census data, estimated
population totals by color and residence for a
given region are computed from population
counts for PSIPs in the CPS sample. Eatios are
then computed between these estimates (based on
sample PSIPs) and the actual population totals
for the region as shown by the 1960 Census. In
deriving these ratios, PSIPs that comprised entire
strata and were selected with certainty (usually
referred to as “self-representing55 PSIPs) are ex­
cluded from the computations, since they represent
only themselves. In tabulations of the monthly
results from the Current Population Survey, the
weights for all sample households from non-selfrepresenting PSIPs in a given region are multi­
plied by the population ratio for that region for
the appropriate color-residence group.
2. Second stage . The second stage of ratio
estimates takes account of current differences be­
tween the population distributions of the sample
and that of the Nation as a whole by age, color,
and sex. Independent estimates of the entire
population, by these characteristics, are prepared
each month. They are calculated by carrying for­
ward the most recent census data (I960) to take
account of subsequent aging of the population,
mortality, and migration between the United
States and other countries. The CPS sample re­
turns (taking into account the weights determined
after the first stage of ratio estimates) are, in
effect, used to determine only the distribution
within a given age-color-sex group by employ­
ment status and various other characteristics. In
developing statistics, these sample distributions
are multiplied by the ratio of the independent
population estimate to the sample estimate for the
appropriate age-color-sex group.
The last step in the prep­
aration of estimates makes use of a composite
estimate. In this procedure, a weighted average
of two estimates is obtained for the current month
Composite Estim ate.



for any particular item. The first estimate is the
result of the two stages of ratio estimates described
above. The second estimate consists of the com­
posite estimate for the preceding month to which
has been added an estimate of the change in each
item between the preceding month and the present
month, based upon that part of the sample which
is common to both months (75 percent). While
the weights for the two components of such a
composite estimate do not necessarily have to be
equal, in this instance the weights used for com­
bining these two estimates are each one-half.
Equal weights in this case satisfy the condition
that for virtually all items there will be some gain
in reliability over the estimation procedure after
the first two stages of ratio estimates.
The composite estimate results in a reduction
in the sampling error beyond that which is
achieved after the two stages of ratio estimates
described above; for some items the reduction is
substantial. The resultant gains in reliability are
greatest in estimates of month-to-month change,
although gains are also obtained for estimates of
level in a given month, change from year-to-year,
or change over other intervals of time.

Presentation and Uses
The Current Population Survey (CPS) pro­
vides a large amount of detail on the economic
status and activities of the population of the
United States. It is the source of estimates of
total employment, both farm and nonfarm; of
nonfarm self-employed persons, domestics, and
unpaid helpers in nonfarm family enterprises; and
of total unemployment, whether or not covered by
unemployment insurance. It is a comprehensive
source of information on the personal character­
istics of the total labor force and of the employed
and unemployed, such as age, sex, race, and marital
It provides distributions of workers by the num­
bers of hours worked (as distinguished from
aggregate or average hours for an industry), per­
mitting separate analyses of part-time workers,
workers on overtime, etc. The survey is a com­
prehensive current source of information on the
occupation of workers (whether teachers, stenog­
raphers, carpenters, laborers, etc.). It also pro­

vides limited statistics on the industries in which
they work.
Information is available from the survey not
only for persons in the current labor force but
also for those who are outside the labor force,
some of whom may be considered to be a “labor
reserve.” The characteristics of such persons—
whether married women with or without young
children, disabled persons, students, older retired
workers, etc.—can be determined. Also, through
special inquiries, it is possible to obtain informa­
tion on their skills and past work experience, if
Each month, a significant amount of basic infor­
mation about the labor force is analyzed and pub­
lished in Employment and E arnings and Monthly
Report on the Labor Force . The tables in this
report provide information on the labor force, em­
ployment, and unemployment by a number of
characteristics, such as age, sex, color, marital
status, industry, and occupation. Approximately
40 of the most important estimates from the CPS
are presented each month on a seasonally ad­
justed basis. These estimates are adjusted by the
B L S Seasonal Factor Method, which is described
in the appendix on seasonal adjustment.
The CPS is also used for a program of special
inquiries to obtain detailed information from par­
ticular segments of the labor force or for
particular characteristics of the labor force. Ap­
proximately 6 to 8 such special surveys are made
each year. The inquiries are repeated annually
in the same month for some topics, including the
extent of work experience of the population during
the calendar year, marital and family characteris­
tics of workers, the employment of school-age
youth, the employment of recent high school grad­
uates and dropouts, the extent of overtime work at
premium pay, and the prevalence of multiple jobholding. Surveys have been made periodically
on such subjects as job mobility, length of time on
current job, and educational attainment of work­
ers. In addition there have been one-time surveys.
Generally, the persons who provide informa­
tion for the monthly CPS questions also answer
the supplemental questions. Occasionally, the
kind of information sought in the special survey
requires the respondent to be the person about
whom the questions are asked.
Information obtained through the supplemental
questions is combined with data in the regular


schedule to provide tabulations of all the desired
personal and economic characteristics of the per­
sons in the special survey. Reports on these spe­
cial surveys are first published in the Monthly
Labor Review. Reprints of the articles, together
with technical notes and additional tables, are
published as Special Labor Force Reports.

Geographic . The Current Population Survey is

designed to produce reliable National estimates.
It is not designed to produce estimates for States
and areas. A sample which could produce State
estimates as reliable as those now published for
the Nation would have to be approximately fifty
times as large as the present sample.
Sources of E rrors in the Survey Estimates. The

estimates from the survey are subject to sampling
errors, that is, errors arising from the fact that the
estimates each month are based on information
from a sample rather than the whole population.
In addition, as in any survey work, the results are
subject to errors made in the field and to errors
that occur in the process of compilation.
Classification errors in labor force surveys may
be particularly large in the case of persons with
marginal attachments to the labor force. These
errors may be caused by interviewers, respondents,
or both, or may arise from faulty questionnaire
design. In spite of a continuous quality control
program, interviewers may not always ask the
questions in the prescribed fashion. To the ex­
tent that varying the wording of the question
causes differences in response, errors or lack of
uniformity in the statistics may result. Similarly,
the data are limited by the adequacy of the infor­
mation possessed by the respondent and the will­
ingness to report accurately.
The estimates from the survey are subject to
various other types of errors beyond those already
mentioned. Some of these are:
Nonresponse. About 4 to 6 percent of occu­
pied units are not interviewed in a typical month
because of temporary absence of the occupants,
refusals to cooperate, or various other reasons.
Although an adjustment is made in weights for
interviewed households to account for noninter­


views, they still represent a possible source of bias.
Similarly, for a relatively few households, some of
the information is omitted because of lack of
knowledge on the part of the respondent or because
the interviewer forgot to ask certain questions or
record the answers. In processing the completed
questionnaires, entries are usually supplied for
omitted items on the basis of the distribution in
these items for persons of similar characteristics.
2. Independent Population Estim ates. The in­
dependent population estimates used in the estima­
tion procedure (see discussion under “Ratio
Estimates,” p. 11) may also provide a source of
error, although on balance their use substantially
improves the statistical reliability of many of the
important figures. Errors may arise in the inde­
pendent population estimates because of under­
enumeration of certain population groups or
errors in age reporting in the last census (which
serves as the base for the estimates) or similar
problems in the components of population change
(mortality, immigration, etc.) since that date.
3. Processing errors. Although there is a qual­
ity control program on coding and a close control
on all other phases of processing and tabulation of
the returns, some processing errors are almost
inevitable in a substantial statistical operation of
this type. It is likely, however, that the net error
arising from processing is fairly negligible.
Measuring the Accuracy of Results. Modern sam­
pling theory provides methods for estimating the
range of errors due to sampling where, as in the
case of the Current Population Survey sample, the
probability of selection of each member of the
population is known. Methods are also available
for determining the effect of response variability
in the Current Population Survey. A measure
of sampling variability indicates the range of dif­
ference that may be expected because only a sample
of the population is surveyed. A measure of
response variability indicates the range of differ­
ence that may be expected as a result of compen­
sating types of errors arising from practices of
different interviewers and the replies of respond­
ents ; these would tend to cancel out in an enumer­
ation of a large enough population. In practice,
these two sources of error—sampling and response
variability, as defined above—are estimated
jointly from the results of the survey. The com­



putations do not, however, incorporate the effect of
response bias, that is, any systematic errors of
response—for example, those that would occur if,
by and large, respondents tended to overstate
hours worked. Response biases occur in the same
way in a complete census as in a sample, and, in
fact, they may be smaller in a well-conducted
sample survey because for the relatively small
sample it is feasible to pay the price necessary to
collect the information more skillfully.
Estimates of sampling and response variability
combined are provided in Employment and E arn ­
ings and Monthly Report on the Labor Force and
in other reports based on the Current Population
Survey, thus permitting the user to take this factor
into account in interpreting the data. In general,
the smaller figures and small differences between
figures are subject to relatively large variation

and should be interpreted with caution. The
availability of the high-speed electronic computer
makes possible considerably more detailed esti­
mates that were possible earlier.
Estimation of response bias is one of the most
difficult aspects of survey and census work. Sys­
tematic studies on this subject are now an integral
part of the Current Population Survey,4 but in
many instances available techniques are not suffi­
ciently precise to provide satisfactory estimates of
response biases. Considerable experimentation is
in progress with the aim of developing more pre­
cise measurements and improving the overall
accuracy of the series.
4 For a summary of these studies, see Bureau of the Census,
Technical Paper No. 6, “The Current Population Survey Reinter­
view Program— Some Notes and Discussion” (March 1963).

Technical References

1. President’s Committee to Appraise Employment and Unemployment Statistics, M e a su rin g E m ­
ploym ent an d Unemployment (1962).
A review of all Federal statistical series on employment and unemployment and a comparison
of the sources and uses of each series. The discussion of labor force statistics includes a brief
history of their development, an evaluation of current concepts and techniques, and recommenda­
tions for further research and improvements.
2. U.S. Department of Commerce, Bureau of the Census, Technical P a p e r N o. 6 , “ The Current Popu­
lation Survey Reinterview Program— Some Notes and Discussion” (1963).
A summary of procedures and results of the Current Population Survey Reinterview Program
from 1955 through 1961 and some interpretations and comparative results from other studies.
3. U.S. Department of Commerce, Bureau of the Census, Technical P a p e r N o. 7, “ The Current Popu­
lation Survey— A Report on Methodology” (1963).
A brief history of the Current Population Survey (CPS) from its inception (1943) to the
present. A detailed description is given for both the sample design and survey procedures. Also
included is a detailed discussion of the various modifications in design and procedures and the
resultant gain in precision.
4. U.S. Department of Labor, Bureau of Labor Statistics and U.S. Department of Commerce, Bureau
of the Census, Concepts an d M ethods U sed in Household S ta tistic s on Em ploym ent an d Unem ploy­
ment fro m the Current P o p u lation Survey. BLS Report 279 and Current Population Reports,
Series P-23, No. 13 (1964).
A concise description of the methodology used in obtaining labor force information from sample
households. Labor force concepts and definitions are defined. The adequacy of labor force
data and quality controls are discussed, and major improvements in the Current Population Survey
are listed chronologically.
5. U.S. Department of Labor, Bureau of Labor Statistics, How the Government M e asu res Unem ploy­
ment (1964).
A short nontechnical discussion of the concepts and methods used in obtaining labor force
statistics from the Current Population Survey.

— C a r o l B . K a l is h


J ohn


B regger

Formerly of the Office of Manpower and Employment Statistics

Chapter 2. Employment, Hours, and Earnings
The first monthly studies of employment and
payrolls by the Bureau of Labor Statistics (B LS)
began in October 1915 and covered four manufac­
turing industries. Prior to that year, the princi­
pal sources of employment data in the United
States were the census surveys—the decennial Cen­
sus of Population, and beginning in 1899, the
quinquennial Census of Manufactures. There
had existed no regular compilation of employment
data between the Census dates.
By November 1916, the B L S program had been
expanded to cover 13 manufacturing industries,
and this number remained unchanged until 1922.
The depression of 1921 directed attention to the
importance of current employment statistics, and
in 1922 Congress granted additional funds to pro­
vide for program expansion. By June 1923, the
number of manufacturing industries covered by
the monthly employment survey had increased to
52. In 1928, concern over increasing unemploy­
ment induced Congress to provide additional ap­
propriations for the program. In the next 4 years,
38 manufacturing and 15 nonmanufacturing in­
dustries were added to the list of industries for
which the Bureau published monthly information
on employment and payrolls.
The onset of the Great Depression in 1930 and
the deepening economic crisis impelled President
Hoover to appoint an Advisory Committee on Em­
ployment Statistics to study the need for expanded
data in this field. The Committee made its re­
port in the spring of 1931 with a number of rec­
ommendations for extension of the program. The
most important of these called for the development
of series on hours and earnings. For the fiscal
year 1932, Congress granted the Bureau a substan­
tial increase in the appropriation for the program.
In January 1933, average hourly earnings and
average weekly hours for the first time were publisted for all manufacturing, for 90 manufac­
turing industries, and for 14 nonmanufacturing
During the period of the Great Depresssion
when mass unemployment threatened to become a
continuing aspect of American life, there
was much controversy among various authorities

concerning the actual number of the unem­
ployed. These discussions pointed up the fact
that no reliable measures of either unemployment
or employment existed. In the early years of the
Roosevelt administration, the Secretary of Labor
made frequent references to the value of the Bu­
reau’s employment estimates as an indirect meas­
ure of unemployment. This interest stimulated
efforts to develop comprehensive estimates of total
wage-and-salary emploment in nonagricultural in­
dustries, and in 1936, the Bureau first published
such a figure.
The preparation of these estimates of overall em­
ployment totals on a monthly basis was contingent
on the development of benchmark data. It was
recognized, even in the 1920’s, that month-tomonth employment trends derived from a sample
of establishments might be fairly accurate for
short periods, but over long intervals of time the
series would not represent the true movement of
employment, unless they were adjusted periodi­
cally to reasonably complete counts of employ­
ment, called benchmarks. The first such adjust­
ment was made in 1935, when the Bureau’s employ­
ment series in manufacturing were adjusted to
totals from the Census of Manufactures for 1923,
1925,1927,1929, and 1931. These series were sub­
sequently adjusted to the successive biennial Cen­
suses of Manufactures, through that of 1939. For
nonmanufacturing industries, benchmarks were
developed from various sources, including the
Censuses of Business taken at intervals from 1929
Throughout the period from 1915 to the begin­
ning of World War II, there was a constantly
growing interest in employment statistics for
States and areas. Even before the Bureau of L a­
bor Statistics entered the field in 1915, three States
(Massachusetts, New York, and New Jersey)
were preparing employment statistics. As early
as 1915, New York and Wisconsin had entered into
“co-operative” agreements with the Bureau,
whereby sample data collected by the State agency
were to be used jointly with the Bureau of Labor
Statistics for the preparation of State and national
series. By 1928, five other States had entered into
such compacts, and another five were added by
1936. Over the years, the amount of published



data on employment and payrolls for States and
areas underwent a constant expansion. In 1940,
estimates of total nonagricultural employment for
all 48 States and the District of Columbia were
published for the first time.
The onset of World War I I in 1939, followed by
the entry of the United States after the assault on
Pearl Harbor in December 1941, placed additional
demands upon the Bureau’s Employment Statis­
tics program. The added responsibilities pointed
up the need for greater uniformity among the
various programs of establishment statistics on
employment and related subjects which were being
prepared by the BLS, the Bureau of the Census,
and the agencies administering the emerging social
security programs. While most improvements
had to await the end of the war, several important
advances took place during those years.
The most far reaching decision was to use as
employment benchmarks the data on employment
collected primarily for administrative purposes
by the newly organized social insurance programs.
Tabulations of such materials became available
about 1940 from the unemployment insurance pro­
gram and they soon became the preferred sources
of benchmark data. They covered several indus­
trial categories not covered by the Census of
Manufacturers and Business, respectively, and
they were available annually. After 1939, these
were taken only at 5-year intervals.
As the unemployment insurance program de­
veloped, the feeling grew that the proper place
to estimate State and area employment was in the
State agencies rather than in Washington. By
1949, all States had joined the system, and since
that year the industry employment statistics pro­
gram has been a fully integrated Federal-State
project which provides employment, hours, and
earnings information on a national, State, and area
basis in considerable industrial detail. This coop­
erative program has as its formal base of author­
ity a Congressional act of July 7, 1930 (ch. 873,
46 Stat. 1019; 29 U.S.C. 2). In 1965, cooperative
arrangements were in effect with 45 State employ­
ment security agencies affiliated with the Bureau
of Employment Security and with 6 State labor
1 Standard Industrial Classification Manual (Bureau of the
Budget, Executive Office of the President, 1957), p. 2.

Description of the Survey
The Bureau of Labor Statistics cooperates in
collecting data each month on employment, hours,
and earnings from a sample of establishments in
all nonagricultural activities including govern­
ment. In 1965, this sample included over 135,000
reporting units. From these data a large number
of series on employment, hours, and earnings in
considerable industry detail are prepared and pub­
lished monthly for the United States as a whole,
for each of the 50 States and the District of Co­
lumbia, and for a majority of the metropolitan
areas. The data include series on total employ­
ment, production or nonsupervisory worker em­
ployment, women employed, average hourly earn­
ings, average weekly earnings, average weekly
hours, and average weekly overtime hours (in
manufacturing). For many series, seasonally
adjusted data also are published.

An establishment is defined as an economic unit
which produces goods or services, such as a fac­
tory, mine, or store. It is generally at a single
physical location and it is engaged predominantly
in one type of economic activity. Where a single
physical location encompasses two or more distinct
and separate activities these are treated as sep­
arate establishments, provided that separate pay­
roll records are available and certain other criteria
are met.1 In the collection of data on employ­
ment, payrolls, and man-hours, the B L S usually
requests separate reports by establishment. How­
ever, when a company has more than one establish­
ment engaged in the same activity in a geographic
area, these establishments may be covered by a
combined report.
Industry employment statistics published by
B L S and the cooperating State agencies represent
the total number of persons employed either full­
time or part-time in nonagricultural establish­
ments during a specified payroll period. In
general, data refer to persons who worked during,
or received pay for, any part of the pay period
that includes the 12th of the month. However, at
the national level, data for Federal Government
establishments generally refer to civilian person­
nel who worked on, or received pay for, the last


day of the month, plus intermittent employees who
worked any time during the month (e.g., Christ­
mas temporary employees of the Post Office
Employed persons include both permanent and
temporary employees and those who are working
either full- or part-time. Workers on an estab­
lishment payroll who are on paid sick leave (when
pay is received directly from the employer), on
paid holiday or paid vacation, or who work during
only a part of the specified pay period are counted
as employed. Persons on the payroll of more
than one establishment during the pay period are
counted in each establishment which reports them,
whether the duplication is due to turnover or dual
jobholding. Persons are considered employed if
they receive pay for any part of the specified pay
period, but are not considered employed if they
receive no pay at all for the pay period. Since
proprietors, the self-employed, and unpaid family
workers do not have the status of “paid em­
ployees,” they are not included. Domestic work­
ers in households are excluded from the data for
nonagricultural establishments. Government em­
ployment statistics refer to civilian employees
The figure which includes all persons who meet
these specifications is designated “all employees.”
Major categories of employment are differentiated
from this overall total, primarily to ensure the
expeditious collection of current statistics on hours
and earnings; these groups of employees are des­
ignated production workers, construction workers,
or nonsupervisory workers depending upon the
In manufacturing industries, data on employ­
ment, man-hours, and payrolls are collected on
production workers. This group, in general,
covers those employees, up through the level of
working foremen, who are engaged directly in the
manufacture of the product of the establishment.
Among the exclusions from this category are per­
sons in executive and managerial positions, and
persons engaged in such activities as accounting,
sales, advertising, routine office work, professional
and technical functions, and force account con­
struction.2 Production workers in mining are
defined in a similar manner. A more detailed de­
scription of the classes of employees included in
the production and nonproduction workers cate­
gories in manufacturing is shown on the facsimile


of the B L S 790 C schedule on pages 18 and 19 of
this bulletin.
In contract construction, the term construction
workers covers workers, up through the level of
working foremen, who are engaged directly on the
construction project either at the site or working
in shops or yards at jobs ordinarily performed by
members of construction trades. Exclusions from
this category include executive and managerial
personnel, professional and technical employees,
and routine office workers.
Data on the employment, man-hours, and pay­
rolls of nonsupervisory workers are collected from
establishments in the transportation, communica­
tion, and public utility industries, in retail and
wholesale trade, in finance, insurance, and real
estate, and in most of the service industries.
Nonsupervisory workers include most employees
except those in top executive and managerial posi­
tions. (See facsimile of B L S 790 E, the report­
ing form for wholesale and retail trade, pp. 20
and 21.)
For government and for nonprofit institutions
and organizations, including hospitals and educa­
tional institutions, data are collected only for “all
The series on hours and earnings is based on re­
ports of gross payroll and corresponding paid
man-hours for production workers.3 In order to
derive these series, B L S collects the following
data: (See facsimile of B L S 790 C on pp. 18
and 19.)
1. The number of full- and part-time produc­
tion workers who worked during, or received pay
for, any part of the pay period including the 12th
of the month.
2. Total gross payrolls for production workers
before deductions for old-age and unemployment
insurance, group insurance, withholding tax,
bonds, and union dues. The payroll figures also
include pay for overtime, shift premiums, holi­
days, vacations, and sick leave paid directly by
the firm to employees for the pay period reported.
They exclude bonuses (unless earned and paid
regularly each pay period) or other pay not earned
in the pay period concerned (e.g., retroactive pay).
2 Force-account construction is construction work performed
by an establishment, primarily engaged in some business other
than construction, for its own account and use and by its own
3 Unless otherwise indicated, the references in this chapter
to production workers apply likewise to the construction worker
and nonsupervisory categories.


B L S 790 E



Budget Bureau No. 44-R745.13.Approval expiree January 31, 1967.





Return promptly each month in the enclosed envelope which requires
no postage




Columns 2 and 3. PAY PE R IO 1) DATES.—The standard survey
reference week for each month is the calendar week (Sunday through
Saturday) which includes the 12tli of the month. Enter the first and
last dates of your pay period which most nearly coincide with this
reference period.
Columns 4 and 5. PAY PERIOD—NUM BER OF DAYS —
Enter in column 4 for the entire pay period reported the number of
days on which the majority of nonsupervisory employees performed
work plua the number of holidays and vacation days during the
period for which the majority were paid. When the period is longer
than a week, enter in column 5 the number of sucli reported days
worked or paid for during the 7 consecutive day period which includes
the 12th of the month and falls entirely within the period reported in
columns 2 and 3.
Column 7. ALL EM PLO YEES—BOTH SE X E S.—Enter the
total number of persons on the payroll(s) covered by this report
who worked full- or part-time or received pay for any part of the
period reported. Include salaried officers of corporations and execu­
tives and their staffs, but exclude proprietors, members of unincor­
porated firms, and unpaid family workers. Include persons on
vacations and sick leave for which they received pay directly from
your firm for the period reported but exclude persons on leave without
company pay the entire period and pensioners and members of the
Armed Forces carried on the rolls but not working during the period
number of women employees included in column 7.
E ES.—Enter the number of nonsupervisory employees, both fulland part-time, on your payroll(s), whether wage or salaried, who
worked during or received pay for any part of the pay period reported.
Include persons on vacations or on sick leave when paid directly by
your firm.
The term “ nonsupervisory employees” includes employees such
as salespersons, shipping and receiving clerks, stock clerks, laborers,
caretakers, general office clerks, office-machine operators, drivers,
installation and repairmen, demonstrators, alteration hands, elevator
operators, porters, janitors and watchmen, and other employees below
the supervisory level, whose services are closely associated with those
of employees listed above. Included in the nonsupervisory category

are employees who may be “ in charge” of a group of employees but
whose supervisory functions arc only incidental to their regular work.
The term “ nonsupervisory employees” excludes officers of cor­
porations, principal executives such as buyers, department heads,
superintendents, and chain-store managers and others who are pri­
marily engaged in planning and directing the work of subordinates.
(Employees listed above should be excluded from column 9 but
included in column 7, All Employees.)

Column 10. PAYROLL.—Enter amount of pay earned during the
pay period by the nonsupervisory employees reported in column 9.
Payrolls should be reported before deductions for old-age and unem­
ployment insurance, group insurance, withholding tax, bonds, and
union dues. Include pay for overtime and for holidays, vacations,
and sick leave paid directly by your firm to employees for the pay
p e r io d

re p o rte d .

Exclude bonuses (unless earned and paid regularly each pay period)
or other pay not earned in pay period reported (e.g., retroacthc pay).
Also, exclude tips contributed by the customer, value of free rent,
fuel, or other payment in kind, or traveling or other expenses of
Columns 10A, 10B,and 10C. COMMISSIONS OF NONSUPER­
VISORY EM PLO YEES.—Enter commissions (not drawing accounts
or basic guarantees) paid to nonsupervisory employees reported in
column 9. If commissions are paid monthly or oftener, enter in
column 10A, the amount of commissions earned during a period as
close to the pay period reported as possible, and in columns 10B and
10C, the first and last dates of the period during which the commis­
sions were earned. If commissions are not paid monthly or oftener,
enter the total commissions paid since the last report and the first and
last dates of the period during which they were earned.
Column 11. MAN-HOURS.—Enter the sum of (1) man-hours
worked (including overtime hours) during the pay period by the
nonsupervisory employees reported in column 9, (2) hours paid for
stand-by or reporting, time, and (3) equivalent man-hours for which
employees received pay directly from the employer for holidays,
vacations, sick leave, or other leave paid to these workers. Do not
convert overtime hours or other premium paid hours to straight-time
equivalent hours.

Form BLS 790 E
Before entering data see
explanations on other side

BLS Codes



1 Report No. |


(Number of establishments)

Information reported on this form is strictly confidential, and will be seen only by sworn
employees of the Bureau of Labor Statistics and the State agencies cooperating in its statistical
programs. It will not be revealed to any other person or agency nor published in such a
manner that data relating to an individual company can be identified.

One pay period
only ( preferably 1
ueek) which in­
cludes the 12th
of the month

Y ear


during the pay period
on which the majority
of nonsupervisory em­ DO
ployees worked plus holi­ N O T
days and vacation days USE
paid for. ( Nearest


Include all persons who worked
during or received pay for any
part of period regardless of type
of work performed

Enter in these columns the number of nonsupervisory em­
ployees who worked during or received pay for any part
of the period reported, the pay earned (excluding com­
missions), and all hours worked or paid for. Include pay
and man-hours for overtime, sick leave, holidays, and


C o m m is s io n s o f N o n s u p e r v is o r y
E m plo yees

M o n th

the entire


(Bothdatetr inclusive)




which in­
cludes the
12th of the



Both sexes

Women only



Jan ...

N u m b er of
N o n su per­
v is o r y

N o n s u p e r v is o r y E m plo yee
P ayro ll

E m ployees

(omit cents)


________ (10)________

N o n s u p e k v is o k y E m ployee
M a n -H o u r s

Amount of

(omit fractions)

(omit cents)



$ ............................

Period in which



( Bothdatett inclusive)
_(I°C ).



____ (I D )___


$ ......................

Ju n e.



Enter in column 13 the main factors responsible for significant month-to-month changes in the report above. Examples are: Wage rate increase, more

overtime, strike, weather..

business, fire, temporary summer help

If any GENERAL WAGE-RATE CHANGES (not individual changes for length of service, merit, or promotion) have occurred since last month’s report, note the amount of increase or






Mar...................................................... ....................... ..................... July.................................................................................................
A p r.......................................................... .............................. .....................................................

Aug.............. —- ...................................................................... --

(Person to be addressed if questions arise regarding this report)

N o v ... .


U.S. GOVERNMENT PRINTING OFFICE: 1 9 6 5 -0 -7 8 0 -7 8 0




B L S 790 C



Budget Bureau No. 44-R745.13.
Approval expires January 81,1987.




W A S H IN G T O N , D .C .



Return promptly each month in the enclosed envelope which requires
no postage



Columns 2 and 3. PAY PERIOD DATES.—The standard survey
reference week for each month is the calendar week (Sunday through
Saturday) which includes the 12th of the month. Enter the first
and last dates of your pay period which most nearly coincide with
this reference period.
Columns 4 and 5. PAY PERIOD—N U M BER OF DAYS.—
Enter in column 4 for the entire pay period reported the number of
days on which the majority of production and related workers per­
formed work plus the number of holidays and vacation days during
the period for which the majority were paid. When the period is
longer than a week, enter in column 5 the number of such reported
days worked or paid for during the 7 consecutive day period which
includes the 12th of the month and falls entirely within the period
reported in columns 2 and 3.
Column 7. ALL EM PLO YEES—BOTH S E X E S —Enter the
total number of persons on the payroll(s) covered by this report
who worked full- or part-time or received pay for any part of the
period reported. Include salaried officers of corporations and execu­
tives and their staffs, but exclude proprietors, members of unincor­
porated firms, and unpaid family workers. Include persons on
vacations and sick leave for which they received pay directly from
your firm for the period reported but exclude persons on leave without
company pay the entire period and pensioners and members of the
Armed Forces carried on the rolls but not working during the period
number of women employees included in column 7.
WORKERS.—Enter the number of production and related workers,
both full- and part-time, on your payroll (s). whether wage or salaried,
who worked during or received pay for any part of the pay period
reported. Include persons on vacations or on sick leave when paid
directly by your firm.
The term “ production and related workers*' includes working
foremen and all nonsupervisory workers (including leadmen and
trainees) engaged in fabricating, processing, assembling, inspection,
receiving, storage, handling, packing, warehousing, shipping, truck­
ing, hauling, maintenance, repair, janitorial, watchman services,
product development, auxiliary production for plant’s own use
(e.g., power plant), and recordkeeping, and other services closely
associated with the above production operations.
The term “ production and related workers’* excludes employees

engaged in the following activities: Executive, purchasing, finance,

accounting, legal, personnel, cafeterias, medical, professional, and
technical activities, sales, sales-delivery (e.g., routemen), advertising,
credit, collection, and in installation and servicing of own products,
routine office function, factory supervision (above the working fore­
man level); and force account construction employees on your payroll
engaged in construction of major additions or alterations to the plant
who are utilized as a separate work force. (Employees in the above
activities should be excluded from column 9 but included in column 7,
All Employees.)
Column 10. PAYROLL.—Enter amount of pay earned during
the pay period by the production and related workers reported in
column 9. Payrolls should be reported before deductions for old-age
and unemployment insurance, group insurance, withholding tax,
bonds, .and union dues. Include pay for overtime and for holidays,
vacations, and sick leave paid directly by your firm to employees for
the pay period reported.
Exclude bonuses (unless earned and paid regularly each pay
period), or other pay not earned in pay period reported (e.g., retro­
active pay), and value of free rent, fuel, meals, or other payment in
Column 11. TOTAL MAN-HOURS.—Enter the sum of (1)
man-hours worked (including overtime hours) during the pay period
by the production and related workers reported in column 9, (2) hours
paid for stand-by or reporting time, and (3) equivalent man-hours
for which employees received pay directly from the employer for
holidays, vacations, sick leave, or other leave paid to these workers.
Do not convert overtime or other premium paid hours to straighttime equivalent hours.
Column 1()Y. OVERTIM E PA Y RO LL—Enter that part of the
payroll included in column 10 which represents the straight-time
pay plus the premiums paid for hours worked in excess of either the
straight-workday or workweek. Payments for Saturday and Sunday
hours (or 6th and 7th day hours) should be included only if overtime,
premiums were paid. Payments for holiday hours worked should not
be included unless the payments are at more than the straight-time
Do not include pay related to hours for which only shift differential,
hazard, incentive, or other similar types of premiums were paid.
If there were no overtime payments, enter “ 0” in column I0Y.
Column 11Y. OVERTIM E MAN-HOURS.—Enter the number
of man-hours included in column 11 for which overtime was paid (as
described above). If none, enter “ 0” in column 11Y,


Report No.

BLS Codes

Form BLS 790 C


Before entering data see explanations on other side

(Number of establishments)


Information reported on this form it strictly confidential, and w ill be teen onlj/ by sworn
employees of the Bureau of Labor Statistics and the State agencies cooperating in Us statistical
programs. It w ill not be revealed to any other person or agency nor published in such a manner
that data relating to an individual com pany can be identified.



Y ear
a n d


One pay period
only ( preferably 1
week) which in­
cludes the 12th
of the month

during the pay period
on which the majority of
production and related
workers worked plus
holidays and vacation
days paid for. ( Nearest

M day)

Include all persons who worked
during or received pay for any
I>art of period regardless of
type of work performed


Enter in columns 9,10, and 11 the number of production and related workers who worked during
or received pay for any part of the period reported, the pay earned, and all hours worked or
paid for. Include pay and man-hours for overtime, sick leave, holidays, and vacations.

Enter in column 13 tho main factors responsible for
significant month-to-month changes in employment,
average hours worked (col. 11-5-col. 9), average hourly
earnings (col. 10-j-ool. 11), etc., as indicated by this
report. Examples are:

Enter in column 11Y the total number of man-hours for which overtime premiums were paid
and in column 10Y the total pay for these hours (straight-time plu s overtime premiums)

M ore business
Tem porary sum m er help
W a n e r a te in c r e a s e
Layoff for retooling

o n th



( Both data i inclusive)





period NDO
includes USE
the 12th
of the


Both sexes


Women only



N t'M R E R O f
P r o d u c t io n
W o rk ers


T o t a l P r o d u c t io n
W o r k er P a y ro ll
I n c l u d in g O v e r ­
t im e P a y

(orm'l cents)
(1 0 )

$ ............................


o ta l



r o d u c t io n

o r k e r

o u r s



a n

O v e r t im e
M a n -H o u r


(omU fractions)

r o d u c t io n


o r k e r


v e r t im e



n c l u d in g


a y r o ll

Included in
column 1 0


a n


o u r s

Included in
column 11

(omU cents)

(omU fractions)

(1 0 Y )

(1 1 Y )

(1 2 )

$ ......................



F ir e

It any general wage-rate changes (not indi­
vidual changes for length of service, merit, or
promotion) have occurred since last month’s
report, note the amount of increase or decrease
(as + 2 % , —b t). the effective date of the change,
and the approximate number of production
workers affected.
(U )





(Person to be addressed if questions arise regarding this report)





BLS 790 Industry Class Supplement

Budget Bureau N o . 44-R745.13.
A pp roval expires January 31, 1967.



Return this form as soon as possi­
ble in the enclosed envelope which
requires no postage.



Information reported on this form is strictly confidential, a n d will be seen only by sworn,
employees of the Bureau of L abor Statistics a n d the State agencies cooperating in its statis­
tical programs.
It will not be revealed to a n y other person or agency nor published in
such a m a n n e r that dat a relating to a n individual comp a n y can be identified.

(C ity)



S ta te

R eport N o.

E m p l.

In d .

M o.

Y r.


_ u i_

P ro p o se d In d .

M o.

In d .

Y r.

This report will be used to insure the proper industrial classification of your regular MONTHLY REPORT ON EMPLOY­
MENT, PAYROLL, AND HOURS and should cover the entire activity of the same establishment.
Classification will be by industry on the basis of the principal product or activity of your establishment during the calendar year
1965. Describe your processes or goods produced m your own words, making the distinctions requested on the list of special character­
istics provided on the enclosed sheet. This list is not complete but represents the kind of information which should be reported.

r in c ip a l


r o d u c ts



c t iv it ie s


u r in g

(List items separately)

P ercent op
T otal S a les
V a lu e or
R e c e ip t s
D u r in g


r in c ip a l


a t e r ia l s



(F o r each product listed in column ( a ) )





M ater
U sed
P rod uced
I n T h is

a s

s t a b l is h m e n t


(b )

ia l


1A. Manufacturing (Specify below) „


IB. Nonmanufacturing (Specify below)

2. Is this establishment part of a multiunit company?
□ Yes □ No
If “ Yes,” enter name and location of controlling company.


Combined Total


3. Space for Your Comments. (Use this space for further explanation of information about the products or activities of your estab­
lishment, additional information concerning materials used, etc.)

(Person to be addressed if questions arise regarding this report)




Tips and the value of free rent, fuel, meals, or
other payment in kind are not included.
3. Total man-hours worked (including over­
time hours) during the pay period, hours paid for
standby or reporting time, and equivalent man­
hours for which employees received pay directly
from the employer for sick leave, for holidays,
vacations, and other leave paid to these employees.
Overtime or other premium paid hours are not
converted to straight-time equivalent hours.
4. Overtime man-hours for which premiums
were paid because the hours worked were in excess
of the number of hours of either the straight-time
workday or workweek. Saturday and Sunday
hours (or 6th and 7th day hours) are included as
overtime only if overtime premiums were paid.
Holiday hours worked as overtime are not in­
cluded unless they are paid for at more than the
straight-time rate. Hours for which only shift
differential, hazard, incentive, or similar types of
premiums were paid are excluded from overtime
5. Overtime payrolls, which are the total
amounts of pay for the overtime man-hours as
described in (4).
Overtime hours and payroll data are collected
only from establishments engaged in manufactur­
ing industries.

Industrial Classification
All national, State, and area employment, hours,
and earnings series data are classified in accord­
ance with the Standard Industrial Classification
Manual , U.S. Bureau of the Budget, 1957, as
amended by the 1963 supplement. (See appendix
B of this bulletin for a detailed description of this
Exporting establishments are classified into sig­
nificant economic groups on the basis of major
product or activity as determined by the establish­
ments5 percent of total sales or receipts for the
previous calendar year. This information is
collected once each year on an “Industry Class
Supplement55 to the monthly report form. (See
p. 22 for a facsimile of this form.) All data for
an establishment making more than one product
or engaging in more than one activity are classified
4 Young, Dudley E . and Goldstein, Sidney, “ The B L S Employ­
ment Series and Manufacturing Reporting Practices,” Monthly
Labor Review, November 1957, pp. 1367-1371.


under the industry of the most important product
or activity, based on the percentages reported.

Time Period
Employment, hours, and earnings are measured
for the pay period including the 12th of the month,
which is standard for all Federal agencies collect­
ing employment data on an establishment basis.

Sample Data

Data Sources

Each month the State agencies cooperating with
the Bureau collect data on employment, payrolls,
and man-hours from a sample of establishments.
The respondent extracts these figures from his pay­
roll records. These data are readily available as
the employers must maintain such records for a
variety of tax and accounting purposes. A re­
sponse analysis survey of the reporting practices
of a scientifically selected sample of reporting es­
tablishments in manufacturing industries showed
that the reports were made out almost exclusively
from their payroll records. The survey also
showed that, while a number of employers did not
report precisely the data requested on the schedule
for all items, these deviations were not all in the
same direction. On balance, they tended to offset
each other, and the net effects of incorrect report­
ing were quite insignificant.4
Participation in the industry employment sta­
tistics program is entirely voluntary on the part
of the reporters. However, in many industries,
particularly in manufacturing, employers who
have a high percentage of total employment in the
industry supply reports regularly, and many have
done so over a long period of years.

Benchmark Data
An employment benchmark is defined as a rea­
sonably complete count of employment. The esti­
mates are adjusted periodically, annually if possi­
ble, to new benchmark levels. Since 1939, the
basic sources of benchmark information for “all
employees55 have been periodic tabulations of em­
ployment data by industry and, beginning with
1959, by size of establishment. These are com­
piled by State employment security agencies, ac­
cording to uniform procedures specified by the
Bureau of Employment Security of the U.S. De­



partment of Labor, from reports of establishments
covered under State unemployment insurance
laws. The State employment security agencies re­
ceive quarterly reports, from each employer sub­
ject to the laws, showing total employment in each
month of the quarter, and total quarterly wages
for all employees.5 I f the employer has more
than 50 employees and operates more than one
establishment in a State, he is required to make
separate reports for each area (e.g., county) in
which he operates and for each establishment in
different industries. Employment is reported for
the pay period of the month including the 12th,
and reports are classified industrially according to
the Standard Industrial Classification. The State
employment security agencies cooperate closely
with the Bureau of Labor Statistics in the assign­
ments of industry classifications, so there is a high
degree of uniformity in this respect between the
benchmark and sample data.
In 1964, unemployment insurance data ac­
counted for three-fourths of the total benchmark.
For the group of establishments exempt from
State unemployment insurance laws because of
their small size6 and for certain classes of non­
profit institutions, the data used are those pro­
vided by the national old-age insurance program
administered by the Social Security Administra­
tion of the U.S. Department of Health, Education,
and Welfare.
For industries not covered by either of the two
programs, benchmarks are compiled from a num­
ber of special sources. The most important of
these are the Interstate Commerce Commission
(interstate railroads), the American Hospital As­
sociation (private nonprofit hospitals), the Office
of Education in the U.S. Department of Health,
Education, and Welfare, and the National Catholic
8 The State employment security agencies are required to sub­
mit tabulations of these reports to the Bureau of Employment
Security each quarter. These tabulations are due in the Wash­
ington office of B E S by the middle of the fifth month after the
end of the quarter of reference. For example, the first quarter
tabulation, which provides the basis of the B L S benchmarks, is
due on August 15. Review and editing of these tabulations and
preparation of national summaries from them requires several
months additional work on the part of both B L S and B E S before
the benchmark is completed.
6 In 1965, the unemployment insurance laws of 19 States and
the D istrict of Columbia covered all employers of 1 or more
workers, 4 States covered employers of 3 workers or more, and
the remaining 27 States covered employers of 4 workers or more.
7 For a more detailed description of the benchmark, see
Schechter, Samuel, “ The 1959 Benchmarks for the B L S Payroll
Employment Statistics,” Monthly Labor Review, December 1962,
pp. 1385-1392.

Welfare Conference (private schools, colleges, and
universities), the U.S. Civil Service Commission
(Federal Government), and the Governments Di­
vision of the Bureau of the Census (State and
local government) .7
Special efforts are made to classify establish­
ments into the same industrial groupings for
benchmark purposes as they are for monthly re­
porting. Wherever possible, employment for the
standard midmonth pay period for March is used
as the benchmark.

Collection Methods
The primary collection of the current sample
data is conducted by State agencies which have
cooperative agreements with the BLS. In most
States, this is the employment security agency,
affiliated with the Bureau of Employment Se­
curity, the organization which administers the
State unemployment insurance program. In a
few cases the State department of labor acts as the
agency. The agencies mail schedules to a sample
of establishments in the States each month. A
“shuttle” schedule is used (B L S form 790 series);
that is, one which is submitted each month in the
calendar year by the respondent, edited by the
State agency, and returned to the respondent for
use again the following month. The State agency
uses the information provided on the forms to
develop State and area estimates of employment,
hours, and earnings, and then forwards the data,
either on the schedules themselves or on punchcards or machine listings, to the Washington, D.C.,
office of the Bureau of Labor Statistics, where
they are used to prepare estimates at the national
The shuttle schedule has been used in this pro­
gram since 1930, but there have been substantial
changes in its design and in the data collected over
the period. All aspects of the schedule, its for­
mat, the wording of the requested items and defini­
tions, and the concepts embodied therein have been
subjected to a continuing and intensive review over
the entire period, not only by the staff of B L S and
of the State agencies, but also by other government
agencies and by numerous persons in private busi­
ness and labor organizations. The report forms
are not exactly alike for every industry, but most


of them request data on total employment, number
of women employees, number of production work­
ers (in manufacturing and mining), construction
workers (in contract construction industries), or
nonsupervisory workers (in other nonmanufac­
turing industries), and, for these workers, data
on payroll, paid man-hours, and for manufactur­
ing, overtime man-hours and payroll. The sched­
ule contains detailed instructions and definitions
for the reports. There are several variants de­
signed to meet the specific problems of different
industries. (See facsimiles of B L S 790 C and
B L S 790 E, pp. 18-21.)
The technical characteristics of the shuttle
schedule are particularly important in maintain­
ing continuity and consistency in reporting from
month to month. The design exhibits automati­
cally the trend of the reported data during the
year covered by the schedule, and therefore, the
relationship of the current figure to the data for
the previous month. The schedule also has opera­
tional advantages; for example, accuracy and
economy are obtained by entering identifying
codes and the address of the respondent only once
a year.
All schedules are carefully edited by the State
agencies each month to make sure that the data
are correctly reported and that they are con­
sistent within themselves, with the data reported
by the establishment in earlier months, and
with those reported by other establishments in
their industry. This editing process is carried
out in accordance with a detailed manual of in­
structions prepared by the Bureau of Labor Sta­
tistics. When the reports are sent to Washington,
they are screened by use of an electronic computer
to detect processing errors and reporting errors
which may have escaped the first editing. Ques­
tionable cases discovered at any stage of the editing
process are returned, if necessary, to the respond­
ent, for review and correction.8

Sampling is used by B L S in its industry employ­
ment statistics program for collecting data in most
industries, since full coverage would be prohibi­
tively costly and time consuming. The sampling
plan for the program must: (a) provide for the


preparation of reliable monthly estimates of em­
ployment, hours of work, and weekly and hourly
earnings which can be published promptly and
regularly; (b) through a single general system,
yield considerable industry detail for metropolitan
areas, States, and the Nation; and (c) be appro­
priate for the existing framework of operating
procedures, administrative practices, resource
availability, and other institutional characteristics
of the program.
In developing the sample design, the universe of
establishments was stratified first by industry and
within each industry by size of establishment in
terms of employment, using six standard size
classes. Within each industry, an optimum allo­
cation design was obtained by sampling with
probability proportionate to average size of estab­
lishment within each of the strata. A total size of
sample necessary to produce satisfactory estimates
of employment had to be distributed among the
size-class cells on the basis of average employment
per establishment in each cell. In practice, this
is equivalent to distributing the predetermined
total number of establishments required in the
sample among the cells on the basis of the ratio
of employment in each cell to total employment in
the industry. Within each stratum, the sample
members are selected at random.
Under this type of design, large establishments
fall into the sample with certainty. In nearly all
industries, establishments with 250 or more em­
ployees are included in the sample with certainty
and in many industries the cutoff is lower. The
sizes of the samples for the various industries were
determined empirically on the basis of experience
and of cost considerations. In a manufacturing
industry in which a high proportion of total em­
ployment was concentrated in a relatively few
large establishments, a high percentage of total
employment had to be included in the sample.
Consequently, the sample design for such indus­
tries provides for a complete census of the larger
establishments with only a few chosen from among
the smaller establishments, or none at all if the
concentration of employment in the larger estab­
lishments is great enough. On the other hand, in
an industry where a large proportion of total em­
ployment is in small establishments, the sample
8 Mendelssohn, Rudolph C. “Machine Methods in Employment
Statistics,” M onthly Labor Review, May 1955, pp. 567-569.



design calls for inclusion of all large establish­
ments, and also for a substantial number of the
smaller establishments. Many industries in the
trade and service divisions fall into this category.
In order to keep the sample to a size which
can be handled with available resources, it is
necessary to accept samples in these divisions
with a smaller proportion of universe employment
than is the case for most manufacturing indus­
tries. Since individual establishments in these
nonmanufacturing industries generally show less
fluctuation from regular cyclical or seasonal pat­
terns than establishments in manufacturing indus­
tries, these smaller samples (in terms of employ­
ment) generally produce reliable estimates.
This sample design, although aimed primarily
at meeting the needs of the national program, pro­
vides a technical framework within which State
and area sample designs can be determined. Since
the estimates for States and areas generally are
not prepared at the same degree of industry detail
as the national estimates, the national design
usually provides sufficient reports for the prepara­
tion of State and area estimates.9

Estimating Procedures
The “all-employee” estimates by industry are
based on reasonably complete employment counts
or “benchmarks.” To obtain employment esti­
mates for the individual estimating cells, the fol­
lowing three steps are necessary:
1. A total employment figure (benchmark) for
the estimating cell, as of a specified month, is
obtained from sources which provide a reasonably
complete count of employment for the cell.
2. For each cell, the ratio of employment in one
month to that in the preceding month (i.e., the
link relative) is computed for sample establish­
ments which reported in both months.
9For the national sample, additional reports needed for State
and area samples are added to those required by the national
10If permanent changes in the composition of the sample take
place, the “production-worker, all-employee” ratios and the
“women” ratios calculated from the sample are modified by a
wedging technique described in this chapter under “Hours and
Earnings.” (See p. 27.)

Beginning with the benchmark month, the
estimate for each month is obtained by multiplying
the estimate for the previous month by the link
relative for the current month.
Application of the estimating procedure in pre­
paring a series is illustrated by the following
example: Assume that total employment for a
given series was 50,000 in July. The reporting
sample, composed of 60 establishments, had 25,000
employees in July and 26,000 in August, a 4-per­
cent increase. To derive the August estimate, the
change for identical establishments reported in the
July-August sample is applied to the July
5°,0°0 x § ^

(or 1.04) =52,000

This procedure for estimating current employ­
ment is known as the benchmark and link-relative
technique. It is an efficient technique, taking
advantage of a reliable complete count of employ­
ment and of the high correlation between levels of
employment in successive months in identical
In addition to estimates of total employment by
industry, the Bureau publishes data on production,
construction, or nonsupervisory worker employ­
ment. For this purpose, the sample ratio for the
current month of production workers to total em­
ployment is used. For example, the 60 sample
firms which had 26,000 employees in August, re­
ported an August production-worker figure of
19,500 resulting in a ratio of
or .750.
Using this ratio, production-worker employment
in August is estimated to be 39,000 (52,000 multipled by .750=39,000). A similar ratio method
is used to estimate the number of women
The estimates for each type of series (all em­
ployees, production workers, and women em­
ployees) for individual estimating cells are
summed to obtain the corresponding totals for
broader industry groupings and divisions.
Appropriate revisions, based on new bench­
marks, are introduced into the employment series
as required to correct for classification changes
and for deviations resulting from the use of sample
trends. In general, the benchmark month is
March. The employment estimates which had
been published previously for that month are com­
pared with the new benchmark data. The amount



of adjustment in the published employment in­
formation is indicated by this comparison. The
all-employee series, for months between the cur­
rent and the last preceding benchmark, are ad­
justed by wedging or tapering out the difference
between the current benchmark and the estimate
for the benchmark month back from the current
benchmark to the last previous benchmark. This
difference is assumed to have accumulated at a
regular rate. The series for months subsequent
to the benchmark month are revised by projecting
the level of the new benchmark by the trend of
the unadjusted series.
A comparison of the amounts of the revisions
made since 1957 is presented in table 1.
The B L S employment statistics sample is not,
strictly speaking, a true probability sample and
hence no formal statement on sampling error can
be made. Nevertheless the B L S estimates of em­
ployment, hours, and earnings are judged to have
a high degree of reliability for most industries.
This reliability is due to the relatively large per­
centage of the employment universe covered by the
sample, the frequent adjustments of employment
estimates to benchmark levels, the use of special
techniques (such as stratification by size and/or
region), and the use of “bias adjustment factors”
which are anticipatory corrections applied to the
employment estimates for each month, to reduce
the amount of adjustment required at the next
benchmark. These “bias adjustment factors” are
based on the percent of adjustment made at the
most recent benchmark revision, modified where
necessary by the experience of earlier benchmark
Differences between the benchmarks and the
estimates result less from sampling and response
errors than from changes in the industrial classi­
fication of individual establishments (owing to
changes in product or activity) which are not re­
flected in the levels of estimates until the data are
adjusted to new benchmarks. At more detailed
industry levels, particularly within manufactur­
ing, changes in classification account for a major
part of the adjustment. However, this factor be­
comes of less importance at broader aggregations
of industries where differences due to classification
are generally small. Another cause of differences,
generally minor, between the estimates and the
benchmark arises from improvements in the
quality of benchmark data.
222-617— 66----- 3

T able

1. N o n a g r ic u l t u r a l E m p l o y m e n t E s t im a t e s ,
b y I n d u s t r y D iv is io n , a s a P e r c e n t a g e o f T h e
B e n c h m a r k fo r R e c e n t Y e a r s
Industry division
Total........ ...................................

Mining.................. ...............................
Contract construction.......... ................
Manufacturing____ ______ ____
Transportation and public utilities___
Wholesale and retail trade_____ ____
Finance, insurance, and real estate___
Service and miscellaneous............ .......
Government________ ___________
Federal........... ............................
State and local................................






101.0 99.3

99.5 100.0 100.3
100.9 101.5 101.5
99.8 100.2 100.1
100.1 100.4 100.0
99.4 100.4 100.6
100.7 99.4
97.9 99.7 100.8
99.8 99.0 103.8
100.0 100.0 100.0
99.7 98.7 »105.1

1962 1961

99.2 99.4
93.9 99.9
99.4 99.7
100.4 100.7
100.1 100.5
99.9 101.0
98.0 99.4
100.0 100.0
100.0 100.0
100.0 100.0

i State and local benchmark derived from October 1962 Census of Govern­
ments. Last previous revision of estimate was made to benchmarks based
on April 1957 Census of Governments.

Hours and Earnings
Independent benchmarks are not available for
the hours and earnings series. Consequently, the
levels shown are derived from the B L S reporting
Since 1959, when benchmark data stratified by
employment size became available through the
employment security system, estimates of employ­
ment, hours, and earnings have been prepared by
a cell structure which makes use of size and in some
cases regional stratification. Experience in the
preparation of current estimates shows that the six
size classes as described under the sampling
design can be combined into a maximum of
three size classes for the purposes of preparing
current estimates of hours and earnings, when
stratification by size is needed.11 Whenever a new
national benchmark becomes available, national
estimates of average weekly hours and average
hourly earnings using eight size strata and four
regional strata (Northeast, North Central, South,
and West) are prepared. These estimates are used
as a standard against which the published averages
are compared. In some cases, this comparison
indicates that some modification of the pattern of
stratification is needed. I f this is the case, a
change is introduced into the estimating structure
at the time of the next benchmark revision.
11 These combinations of size classes have been made because
of operational economies. In particular, the preliminary esti­
mates are based on less than full samples, and if the estimation
of preliminary estimates was attempted using the full stratifica­
tion pattern, there would be a number of cells for which there
were no samples. Experiments and tests over several years
indicate that estimates of hours and earnings prepared from the
BLS sample using a maximum of three size strata generally do
not differ significantly from those computed with four size strata
or more.



a. Average weekly hours and gross earnings .
To obtain average weekly hours for an individual
estimating cell, the sum of the man-hour totals
reported by the plants classified in that cell is
divided by the total number of production work­
ers reported for the same establishments. Simi­
larly, in computing average hourly earnings, the
reported payroll total is divided by the reported
man-hour total.
The first ratio estimates of average hourly earn­
ings and average weekly hours are modified at the
estimating cell level by a wedging technique de­
signed to compensate for changes in the sample
arising mainly from the voluntary character of
the reporting.
For example, a first estimate of average hourly
earnings for the current month, Z7i, is obtained
from aggregates from a matched sample of estab­
lishments reporting in the current and previous
month. Similarly an estimate of average hourly
earnings, Z70, for the previous month is calculated
from the same matched sample. Hence, Ui —U0
is a measure of absolute change between the 2
Note is then taken of the published estimate of
average hourly earnings for the previous month,
say V0. Because the panel of establishments re­
porting in the sample is not absolutely fixed from
month to month, there may be differences between
V0 and U0.12 A final figure for the current month
is obtained by making use of both pieces of infor­
mation ; the estimate is
F i = (0.9 Vo + 0.1 U0) + {U i- U o )
The procedure, reflected in this last equation, ac­
cepts the advantage of continuity from the use of
the matched sample, and at the same time tapers
or wedges the published estimate toward the level
of the latest sample average. The same procedure
is used to adjust the production-worker all-em­
ployee-ratio and the ratios for women with re­
gard to changes in the composition of the sample
between successive months.
Weekly hours and hourly earnings for industries
and groups above the basic estimating cell level
12If the difference between the estimate and the average com­
puted from the sample ( V o — U « ) , is too great, the sample average
is accepted once it has been established that the difference is
due to a permanent change in the composition of the sample,
and the series is regarded as discontinuous at that point. In
general, a difference greater than 3 percent is considered as de­
fining a discontinuity or “break.”

T a b l e 2. N u m b e r of I n d u st r ie s for W hich “ P r im a r y ”
S e r ie s A r e P u b l ish e d U n d er t h e BLS I n d u str y
E m ploym en t
S t a t ist ic s
P rogram — E m plo ym en t ,
H o u rs , and E a r n in g s , J a n u a ry 1966
Industry division
Total..................... .
Contract construction__
Manufacturing....... .........
Transportation and
public utilities..............


Finance, insurance, and
real estate__________
Service andmiscellaneous_

Hours Average
duction Women
earnings 2 hours
ployees workers 1













* Production workers in manufacturing and mining; construction workers
in contract construction; nonsupervisory workers, all other divisions.
2 Average hourly earnings, average weekly hours, and average weekly

are weighted averages of the figures for component
cells and industries. The average weekly hours
for each estimating cell are multiplied by the cor­
responding estimate of production-worker employ­
ment, to derive aggregate man-hours. Payroll
aggregates are the product of aggregate man-hours
and average hourly earnings. Payroll and man­
hour aggregates for industry groups and divisions
represent the sum of aggregates for component
Average weekly hours for industry groups are
obtained by dividing the man-hour aggregates by
the corresponding production-worker employment
estimates. Average hourly earnings for groups
are computed by dividing the payroll aggregates
by the man-hour aggregates. This method is
equivalent to weighting weekly hours by estimated
universe production-worker employment and
hourly earnings by estimated universe man-hours.
For all levels, from individual estimating cells
to major industry divisions, average weekly earn­
ings are computed by multiplying average hourly
earnings by average weekly hours. Man-hour data
were not collected for many industries in the fi­
nance and service divisions prior to 1964. In these
industries when reporting of man-hour data is
incomplete, average weekly earnings are obtained
by dividing the sum of the reported payroll totals
by the total number of nonsupervisory workers
for the same establishments.
b. Overtime man-hours. To obtain average
weekly overtime hours in manufacturing indus­
tries, the sum of the overtime man-hours reported


is divided by the number of production workers
in the same establishments.
c. Spendable average weekly earnings . Before
the majority of workers in lower income brackets
were subject to Federal income and social security
taxes, gross average weekly earnings were a satis­
factory measure of trends in weekly earnings avail­
able for spending. After Federal income taxes
began to affect the spendable earnings of an ap­
preciable number of workers, a method was de­
veloped for approximating spendable earnings by
deducting these taxes from gross earnings.13
The amount of individual income tax liability
depends on the number of dependents supported
by a worker as well as on the level of his gross
income. Spendable earnings for workers in the
manufacturing, mining, contract construction, and
wholesale and retail trade divisions are computed
and published for a worker with no dependents
and a worker with three dependents.
Gross and spendable weekly earnings also are
computed and published in terms of 1957-59 dol­
lars, to give an approximate measure of changes
in “real” gross and spendable weekly earnings,
that is, in purchasing power since that base period.
This series is computed by dividing the weekly
earnings average (in current dollars) by the B L S
Consumer Price Index for the same month.
d. Average hourly earnings, excluding overtime,
in manufacturing are computed by dividing the
total production-worker payroll for the industry
group by the sum of total production-worker man­
hours and one-half of total overtime man-hours,
which is equivalent to payrolls divided by straighttime man-hours. This method excludes overtime
earnings at 1% times the straight-time rates; no
further adjustment is made for other premium
payment provisions.
e. Indexes of aggregate weekly man-hours and
payrolls are prepared by dividing the current
month’s aggregates by the averages for the 1957-59

Seasonally Adjusted Series
Many economic statistics, including employment
and average weekly hours, reflect a regularly recur­
ring seasonal movement which can be measured on
is “The Calculation and Uses of the Spendable Earnings Series,”
April 1965, pp. 405-409.

M onthly Labor Review ,


the basis of past experience. By eliminating that
part of the change which can be ascribed to usual
seasonal variation, it is possible to observe the cycli­
cal and other nonseasonal movements in these
series. Seasonally adjusted series are published
regularly for selected employment and hours series.
The seasonal adjustment method used for these
series is an adaptation of the standard ratio-tomoving average method, with a provision for
“moving” adjustment factors to take account of
changing seasonal patterns. A detailed descrip­
tion of the method is given in appendix A of this
The seasonally adjusted series on gross average
weekly hours and average overtime hours are com­
puted by applying factors directly to the corre­
sponding unadjusted series, but seasonally adjusted
employment totals for all employees and produc­
tion workers by industry divisions are obtained by
summing the seasonally adjusted data for compo­
nent industries. Seasonally adjusted series also
are prepared for aggregate weekly man-hours for
mining, contract construction, manufacturing, and
the major industries in manuf acturing.

Presentation and Uses
At the national level, the program produces each
month a total of over 2,000 separate published
series. Tables 2, 3, and 4 provide a summary of
the detail which is published currently. Table 2
describes the “primary” series produced by the
program, that is, those computed directly from the
sample and benchmark data.
Table 3 indicates the “special” series which are
obtained from the primary series by application of
special adjustments, while table 4 lists the season­
ally adjusted series by type and industry division.
In addition to the series published on a current
monthly basis, a single annual figure for employ­
ment in March of each year (based on benchmarks)
is published for a number of industries for which
monthly estimates do not currently meet estab­
lished standards for publication. In late 1964,
following revision to the 1963 benchmark, data for
210 such industries were published.
In December 1964, employment and hours and
earnings statistics were available for 50 States, the
District of Columbia, and 154 areas. Approxi­



mately 6,600 employment series and hours and
earnings series for about 3,100 industries were
published for these States and areas by the State
agencies. The employment series usually covered
total nonagricultural employment, major industry
divisions (e.g., contract construction, manufactur­
ing), and major industry groups (e.g., textile mill
products, transportation equipment, retail trade)
for each State and area. Additional industry de­
tail frequently is provided for the larger States
and areas, particularly for industries which are
locally important in the various jurisdictions.
The series on employment, and hours and earn­
ings appear in several B L S publications. The
summary data are first published each month in a
press release which contains preliminary national
estimates of nonagricultural employment, weekly
hours, and gross average weekly and hourly earn­
ings in the preceding month, for major industry
categories. The release also includes seasonally
adjusted data on employment, average weekly
hours, and average overtime hours. The prelimi­
nary estimates are based on tabulations of data for
less than the full sample to permit early release of
figures. This release normally is issued 4 weeks
after the week of reference for the data. The press
release also includes a brief analysis of current
trends in employment, hours, and earnings, point­
ing up current developments as compared with
those for the previous month and the same month
in the preceding year.
National estimates in the detail described in
tables 2, 3, and 4 are published in the monthly
report Employment and Earnings and Monthly
Report of the Labor Force . This publication is
T a b l e 3. N u m b e r of I n d u st r ie s fo r W hich S p e c ia l
S e r i e s a r e P u b l ish e d U n d er T h e B L S I n d u st r y
E m ploym en t
S t a t ist ic s
P rogram — E m ploym en t ,
H o u rs , and E a r n in g s , J a n u a ry 1966

Industry division

Index of

Contract construc­
Trade__ _______
Mining, construc­
tion and manu­

Index of Spendable Gross
aggregate average
weekly earnings earnings
payrolls earnings 1 (1957-59 (excluding
dollars) overtime)











1In current and 1957-59 dollarss


T a b l e 4. N u m b e r of S e a so n a lly A d ju s t e d S e r i e s
P u b l ish e d U n d e r T h e B L S I n d u str y E m plo y m en t
S t a t ist ic s P rogram — E m ploym en t , H o u r s , and
E a r n in g s , J a n u a r y 1966
Seasonally adjusted series
Industry division

Mining, construction,
and manufacturing. _
Contract construc­
and public
Finance, insur­
ance, and real
estate___ _____
Service and

Pro­ Indexes Aver­
duc­ of em­
ploy­ tion
ploy­ week­
ees work­ ment





















issued about 5 weeks after the week of reference.
Employment data for total nonagricultural em­
ployment and for the major industry divisions,
as well as hours and earnings for all manufactur­
ing, are published for States and areas in Em ploy­
ment and Earnings and Monthly Report of the
Labor Force , 1 month later than those for the

Nation. Special articles analyze long-term eco­
nomic movements or describe technical develop­
ments in the program. Many of the national
series are republished in the Monthly Labor R e­
view with data shown for each series for the most
recent 13 months.
Following each benchmark revision, an histori­
cal volume called Employment and Earnings S ta ­
tistics fo r the United States is published. This
provides historical data, monthly and annual aver­
ages, from the beginning date of each series, in a
few instances as far back as 1909. A companion
volume, Employment and Earnings Statistics fo r
States and Areas , provides historical data ( annual
averages) on all employees and on productionworker hours and earnings series published by
State agencies for States and areas back to the
beginning of these series, in some instances to 1939.
This volume is published annually. Detailed in­
dustry rates are available monthly in releases pub­
lished by the cooperating State agencies.
The data are disseminated also through the
publications of many other Federal agencies; e.g.,



the Department of Commerce, the Board of Gov­
ernors of the Federal Reserve System and the
Council of Economic Advisors republish all or
part of the data. They are also regularly repub­
lished in summary form or for specific industries
in many trade association journals, the labor
press, and in general reference works.
These series are used by labor unions, business
firms, universities, trade associations, private re­
search organizations, and many government agen­
cies. Research workers in labor unions and indus­
try, as well as others responsible for analyzing
business conditions, use the trends reflected in these
particular statistics as economic indicators. The
average weekly hours series are utilized as lead
indicators of swings in the business cycle. Labor
economists and other social scientists find these
series to be an important indicator of the Nation’s
economic activity, as well as a measure of the well­
being of the millions of Americans who depend
on salaries and wages. Industrial growth and
progress may be assessed by using the employment
and hours series in conjunction with other eco­
nomic data to yield measures of productivity.
Analysts study employment trends to detect
changes in industrial structure, and to observe
growth and decline proclivities of individual in­
dustries. They also are used in the Bureau’s Occu­
pational Outlook program as a basis for projection
of future trends.
Executives use the employment, earnings, and
hours data for guidance in plant location, sales,
and purchases. Also, firms negotiating long-term
supply or construction contracts often utilize series
on average hourly earnings as an aid in reaching
an equitable agreement; “escalation clauses” may
be included in the contracts, which permit an in­
crease or a lowering of the settlement price
depending on the movement of average hourly
earnings in a selected industry.14 Wide need has
been demonstrated by both labor and business for
industry series on hourly earnings and weekly
hours, to provide a basis for labor-management
negotiations. They not only furnish current and
historical information on a given industry but
provide comparative data on related industries.
14Weinberg, Edgar, “BLS Earnings Series as Applied to Price
Escalation,” M onthly Labor Review , July 1952, pp. 57-59.
15 See chapter 1 for a description of this survey.

Total employment in nonagricultural establish­
ments from the “payroll” survey is not directly
comparable with the Bureau’s estimates of the
number of persons employed in nonagricultural in­
dustries, obtained from the monthly “household”
survey.15 The “payroll” series excludes unpaid
family workers, domestic servants in private
homes, proprietors and other self-employed per­
sons, all of whom are covered by the household
survey. Moreover, the “payroll” series counts a
person employed by two or more establishments at
each place of employment, while the “household”
survey counts him only once, and classifies him
according to his single major activity. Certain
persons on unpaid leave are counted as employed
under the “household” survey, but are not included
in the employment count derived from the “pay­
roll” series. In addition to these differences in
concept and scope, the surveys employ different
collection and estimating techniques. Therefore,
although each survey measures changes in employ­
ment, direct comparability should not be expected.
However, over time, the trends are similar. The
household survey places its primary emphasis on
the employment status of individuals and also pro­
vides a great deal of information on the demo­
graphic characteristics (e.g., sex, age, race) of the
labor force. The survey is not well suited to pro­
viding detailed information on the industrial and
the geographic distribution of employment. The
establishment survey, while providing limited in­
formation on personal characteristics of workers,
is an excellent vehicle for obtaining these detailed
industrial and geographic data, and in addition, it
provides hours and earnings information which is
directly related to the employment figures. The
payroll and household surveys therefore may be
regarded as complementary.
Employment estimates derived by the Bureau
of the Census from its quinquennial censuses and
from the annual sample surveys of manufacturing
establishments may differ from B L S employment
statistics. The most important reason for differ­
ence stems from the degree to which multiproduct
establishments file separate or combined reports



in one survey but not the other, which may result
in different industrial classification of employ­
ment. There is also a significant difference at
the more detailed industry levels, since Census
classifies auxiliary units and central and district
administrative units on the basis of the most ap­
propriate 2-digit major group, while B L S codes
these units to the most appropriate 4-digit indus­
try. For broad categories, however, the two sur­
veys do show similar levels and trends.

Hours and Earnings
The workweek information relates to average
hours paid for, which differ from scheduled hours
or hours worked. Average weekly hours reflect
the effects of such factors as absenteeism, labor
turnover, part-time work, and strikes.
The gross average hourly earnings series reflect
actual earnings of workers, including premium
pay. They differ from wage rates, which are the
amounts stipulated for a given unit of work or
time. Gross average hourly earnings do not
represent total labor costs per man-hour for the
employer, for they exclude retroactive payments
and irregular bonuses, various welfare benefits,
and the employer’s share of payroll taxes. Earn­
ings for those employees not covered under the
production worker and nonsupervisory-employee
categories are, of course, not reflected in the
The series on spendable weekly earnings measure
the disposable earnings of workers who earn the
average gross weekly earnings, have the specified
number of dependents, and take the standard de­
ductions for Federal income tax purposes. Spend­
able earnings reflect deductions only for Federal
income and social security taxes (calculated on the
basis of total annual liabilities), and thus represent
only a rough approximation of disposable earn­
ings. They do not take into account payroll de­

ductions for such purposes as State income taxes,
union dues, or group insurance, and they do not
reflect such factors as total family income or tax
deductions above the standard amount.
The “real” earnings data (those expressed in
1957-59 dollars), resulting from the adjustment
of gross and spendable average weekly earnings by
means of the Bureau’s Consumer Price Index,
indicate the changes in the purchasing power of
money earnings as a result of changes in prices for
consumer goods and services. These data cannot
be used to measure changes in living standards as
a whole, which are affected by other factors such
as total family income, the extension and incidence
of various social services and benefits, and the
duration and extent of employment and
To approximate straight-time average hourly
earnings, gross average hourly earnings are
adjusted by eliminating only premium pay for
overtime at the rate of time and one-half. Thus,
no adjustment is made for other premium pay­
ment provisions such as holiday work, late-shift
work, and premium overtime rates other than at
time and one-half.
The ultimate goal of the program is to provide
current estimates of employment, hours, and
earnings for all nonagricultural industries in the
Nation as a whole, and also for all significant
industries in all States and all Standard Metro­
politan Statistical Areas, as defined by the U.S.
Bureau of the Budget. While very substantial
progress toward this objective has been made over
the years, and particularly since the end of World
War II, there remain some important areas where
the goal is yet to be realized. Efforts constantly
are being directed toward strengthening the
sample so that series for employment, hours, and
earnings for additional industries may be pub­
lished, and also toward developing series for addi­
tional standard metropolitan areas.


Technical References

1. “ The Calculation and Uses of the Spendable Earnings Series,” M onthly L ab or Review , April
1966, pp. 405-409.

2. Mendelssohn, Rudolph C., “ Machine Methods in Employment Statistics,” M onthly L ab or Review ,
May 1955, pp. 567-569.
A description of the use of electronic data-processing equipment in the preparation of em­
ployment statistics, with particular reference to screening employers' reports for errors.
3. President's Committee to Appraise Employment and Unemployment Statistics, M e asu rin g E m ­
inent an d U nem ploym ent 1962.
A comprehensive review and critique of the methods and concepts used by various Federal
Government programs providing statistics on employment, unemployment, and the labor force
in the United States.
4. Schechter, Samuel, “ The 1959 Benchmarks for the BLS Payroll Employment Statistics,” M onthly
L ab or Review , December 1962, pp. 1385-1392.
A detailed description of the sources and construction of BLS employment benchmarks.
5. Weinberg, Edgar, “ BLS Earnings Series as Applied to Price Escalation,” M onthly L ab or Review ,
July 1952, pp. 57-59.
A discussion of the use of BLS average hourly earnings series in escalation clauses in contracts.
6. Wymer, John P., “ The Revised and Expanded Program of Current Payroll Employment Statistics,”
Em ploym ent an d E a r n in g s , November 1961, pp. iv-vii.
A description of the impact of a major benchmark adjustment and of important technical
innovations on the industry employment statistics series.
7. Young, Dudley E. and Sidney Goldstein, “ The BLS Employment Series and Manufacturing Re­
porting Practices,” M onthly L ab or Review , November 1957, pp. 1367-71.
A discussion of the findings in a survey analyzing the response patterns of manufacturing
establishments cooperating in the industry employment statistics program.

— J ohn P . W ym er
Office of Manpower and Employment Statistics

Chapter 3. Labor Turnover

Description of the Survey

Labor turnover, as used in the Bureau of Labor
Statistics program, refers to the gross movement
of wage and salary workers into and out of em­
ployment status with respect to individual
In January 1926, the Metropolitan Life Insur­
ance Co. began the collection of labor turnover
data from a small sample of manufacturing estab­
lishments. By February 1927, the sample in­
cluded 175 establishments with 800,000 employees,
which was estimated to be about 8 to 10 percent
of total manufacturing employment at the time.
The original purpose of this series was to provide
personnel managers with national figures on labor
turnover rates for manufacturing industries
against which they could measure the experience
of their own plants. Between November 1927
and July 1929, the Metropolitan Life Insurance
Co. published labor turnover rates for total manu­
facturing. By the latter date, the company felt
the project was sufficiently successful and well
established to warrant turning it over to the
Bureau of Labor Statistics for further develop­
ment. A decade later, in December 1939, series on
labor turnover rates were being published for 30
manufacturing industries, and the sample upon
which the rates for all manufacturing were based
contained 5,500 establishments and nearly
2,600,000 employees.
For a number of years, State employment secu­
rity agencies affiliated with the Bureau of Employ­
ment Security had collected labor turnover infor­
mation for use in job market analysis and as a
guide for the operations of the State employment
services. Cooperative arrangements between
these agencies and the Bureau of Labor Statistics
for the joint collection of labor turnover data
began with an agreement with Connecticut in 1954.
By 1964, the cooperative program had been ex­
tended to cover all 50 States and the District of
Columbia. In December 1964, these agencies pub­
lished nearly 10,000 labor turnover series in manu­
facturing and mining industries for States and
areas. These rates were based on a sample of
approximately 44,000 reports which also were
used to prepare the national series.

Labor turnover actions are divided into two
broad groups: accessions or additions to employ­
ment, and separations or terminations of em­
ployment. These two broad groups are further
divided; accessions into new hires and other acces­
sions, and separations into quits, discharges, lay­
offs, and other separations. Labor turnover is
expressed in the B L S series as a monthly rate per
100 employees. Separate rates are computed for
each of the component items.
The primary difference between types of sep­
arations is whether action is initiated by the em­
ployee or employer, i.e., whether it is voluntary
on the employee’s part or involuntary. Voluntary
actions—quits—are initiated by the employee for
an almost unlimited variety of reasons, financial,
personal, or social, (e.g., lack of housing and
transportation, poor community facilities, etc.).
Involuntary actions either may be initiated by the
employer or be beyond the control of both employer
and employee; these actions may arise from eco­
nomic causes such as business conditions, physio­
logical reasons such as aging, or performance rea­
sons such as incompetence.


Separations are terminations of employment of
persons who have quit or been taken off the rolls
for reasons such as layoff, discharge, retirement,
death, military service expected to last more than
30 consecutive calendar days, physical disability,
etc. Since January 1959, transfers of employees
to other establishments of the same company also
have been classified as separations.
Quits are terminations of employment initiated
by employees for any reason except retirement,
transfer to another establishment of the same firm,
or service in the Armed Forces. Included as quits
are persons who failed to report after being hired
(if previously counted as accessions), and unau­
thorized absences which, on the last day of the
month, have lasted more than 7 consecutive calen­
dar days.
Layoffs are suspensions from pay status (lasting
or expected to last more than 7 consecutive calen­
dar days), initiated by the employer without prej­


udice to the worker, for reasons such as lack of
orders, model changeover, termination of seasonal
or temporary employment, inventory-taking, intro­
duction of labor saving devices, plant breakdown,
or shortage of materials.
Discharges are terminations of employment
initiated by the employer for such reasons as in­
competence, violation of rules, dishonesty, laziness,
absenteeism, insubordination, failure to pass
probationary period, etc.
Other separations include terminations of em­
ployment for military duty lasting or expected to
last more than BO days, retirement, death, perma­
nent disability, failure to meet the physical stand­
ards required, and transfers of employees to an­
other establishment of the company.
Accessions are all permanent and temporary
additions to the employment roll, whether of new
or rehired employees. Transfers from another
establishment of the Same company also are
counted as accessions (beginning with January
New hires are permanent and temporary addi­
tions to the employment roll of persons who have
never before been employed by the establishment,
and former employees rehired although not specifi­
cally recalled by the employer. This category ex­
cludes transfers from other establishments of the
same company and employees returning from
military service or unpaid leaves of absence.
Other accessions include all additions to the em­
ployment roll other than new hires.

Industry Classification
The classification system used for compiling and
publishing rates is that described in the 1957
Standard Industrial Classification Manual as
amended by the 1963 supplement, issued by the
Bureau of the Budget. (See appendix B of this
bulletin for a detailed description of this system.)
Reporting establishments are classified on the
basis of major product or activity as determined
by annual sales data for the previous calendar year.
Most establishments in the labor turnover sample
also report employment, hours, and earnings under
the Bureau’s industry employment statistics pro­
gram, and are assigned the same industry classifi­
cation in both programs. Further discussion of
industry classification in the two programs is given
2 2 2 - 6 1 7 — 66-



under the heading, Industrial Classification in
chapter 2 of this bulletin.

Data Sources
The labor turnover program covers all industries
in the manufacturing and mining divisions, and
for the Nation as a whole, the telephone com­
munications and the telegraph communications in­
dustries. Each month the Bureau collects data
on labor turnover actions from a sample of estab­
lishments drawn from a list of those subject to
State unemployment insurance programs. (See
chapter 2, p. 15 of this bulletin.) The respondent
extracts the figures largely from his personnel
records, though some smaller establishments which
do not maintain special personnel records utilize
their payroll records in making out the reports.
A response analysis survey, which analyzed the
reporting practices of a scientifically selected
sample of the establishments in the labor turnover
panel, showed that a number of employers did not
report the figures for all items precisely as re­
quested on the schedule. Nevertheless, the effect
of these deviations on the published data appeared
to be quite insignificant, particularly for the
broader classes of turnover actions, namely total
accessions and total separations.

Collection Methods
Labor turnover data are collected primarily at
the State level by employment security agencies
from cooperating employers via the medium of a
mailed “shuttle” schedule, U.S. Department of
Labor form 1219. (See pp. 36 and 37 for a fac­
simile of this schedule.) The same form is re­
turned to the respondent each month of the year for
the entry of current data. The respondent reports
the total number of actions for each turnover item
during the calendar month, total employment, and
the number of women employed. These employ­
ment figures, which are the bases used to compute
the rates, represent the number of persons who
worked or received pay for any part of the pay
period (usually 1 week) which includes the 12th
of the month. Turnover data reported for women
are limited to total accessions, total separations,
and quits.


DL 1219

Budget Bureau No. 44-Rl 064.6.
Approval expires January 31. 1967.


W A S H IN G T O N . D .C .


Report No.

Enter the data requested and return in
the enclosed envelope as soon as the
information is available each month.







Before entering data see explanations on other side

In fo rm a tio n rep o rted on th is fo rm is s tr ic tly con fiden tial, a n d w ill be seen
o n ly b y sw o rn em ployees o f th e B u reau o f L a b o r S ta tistic s, the B ureau o f E m p lo ym en t S e cu rity, a n d th e S ta te a gen cies co o p era tin g in th eir sta tistic a l p ro ­
g ra m s. I t unit n o t be revea led to a n y o th er p erson o r a g en cy n or published in
sucA a m a n n er th a t d a ta rela tin g to a n in d ivid u a l com p an y can be iden tified.

L O C A T IO N _______________________________________________________
(C ity)


e r io d



ea r

a bo r

C o v ered
T urnov

b y



(Cols. 4 through 14)
Preferably one calendar


Separations (during calendar month)

c c e s s io n s

(during calendar month)



c c e s s io n s


o n th



(Both date:s inclusive)



(Sum o f cols.
5 through 8)






(7 )

separations (Sum of cols.
10 and 11)

New hires



(I D

in col. 4

in col. 5

in col. 9


Y ear

N umber o f E m p l o y e e s who worked during
or received pay for any part of the pay pe­
riod which includes the 12th of the month


M o n th


Both sexes


E x p l.


Enter M ain F a ctors
Examples are:

R esp o n sible fo r any significant
M ore b u siness
S ea so n al in crea se


month-to-month changes in Sections I and II.

T e m p o ra ry su m m er h elp
F ire

S trik e
W eath er



(Person to be addressed if questions arise regardin g this report)




I n f o r m a t i o n o n la b o r t u r n o v e r , c o ls . 4 t h r o u g h 1 4 , i s r e q u e s t e d f o r t h e entire calendar month
o r , i f t h i s i s n o t p o s s i b l e , f o r a p e r io d w h ic h m o s t c l o s e l y c o v e r s t h e c a l e n d a r m o n th . I n e i t h e r
c a s e e n t e r in c o ls . 2 a n d 3 th e b e g in n i n g a n d e n d in g d a t e s f o r th e m o n t h ly p e r io d f o r w h ic h
tu r n o v e r d a t a a r e re p o rte d .
I n f o r m a t i o n o n e m p lo y m e n t, c o ls . 16 a n d 17 , i s r e q u e s t e d f o r one pay period only ( p r e f e r a b l y
o n e w e e k ) w h ic h in c lu d e s t h e 1 2 th o f th e m o n th .

SEPARATIONS (All Employees)

ACCESSIONS (All Employees)



(Column It)


(Column 5)

A q u it is a te r m in a tio n o f e m p lo y m e n t initiated by the
employee f o r a n y r e a s o n e x c e p t to r e tir e , to t r a n s f e r to
a n o th e r e s ta b lis h m e n t o f th e s a m e firm , o r f o r s e r v ic e in
th e A r m e d F o r c e s . In c lu d e a p e r so n w h o f a i l s to r e p o r t
a f t e r b e in g h ire d ( i f p r e v io u sly co u n te d a s a n a c c e s s io n )
a n d a n u n a u th o riz e d a b s e n c e i f o n th e l a s t d a y o f th e m o n th
th e p e rso n h a s b een a b s e n t m o re th a n 7 c o n s e c u tiv e c a l­
e n d a r d a y s.


(Column 9)

A n a c c e ssio n i s a n y p e rm a n e n t o r te m p o r a r y a d d itio n to th e em plo y ­
m e n t ro ll w h e th e r o f n ew o r fo r m e r em plo y ees, o r t r a n s f e r s fr o m
a n o th e r e s ta b lish m e n t o f th e co m p an y . E n t e r in col. 9 th e su m o f
co ls. 10 a n d 11.

E n t e r in co lu m n 4 th e su m o f c o lu m n s 5 th ro u g h 8.


(Column 10)

N ew h ir e s a r e te m p o r a r y o r p e rm a n e n t a d d itio n s to th e e m ­
p lo y m e n t roll o f (1 ) an y o n e w ho h a s n e v ei b e fo re b een em p lo y ed
in th is e s ta b lis h m e n t, o r ( 2 ) fo r m e r e m p lo y e e s you d id not c a ll
b ac k . F o r m e r e m p lo y e e s yo u did c a ll b ac k sh o u ld be in clu d ed
in " t o t a l a c c e s s io n s ” a n d “ o th e r a c c e s s io n s ," b u t not in "n e w
h ir e s .” P e r s o n s tr a n s f e r r e d fro m o th e r e s ta b lis h m e n ts o f th is
c o m p an y sh o u ld b e r e p o rte d in “ o th e r a c c e s s io n s .”

(Column 6)

A d is c h a r g e is a te r m in a tio n o f em p lo y ­
m ent

initiated by the employer f o r su c h

r e a s o n s a s in co m p e te n ce, v io latio n o f
in su b o rd in a tio n , f a il u r e to p a s s p r o b a ­
tio n a r y p e rio d , etc.

I n a b ility to m e e t o r ­

g a n iz a tio n ’s p h y s ic a l s t a n d a r d s sh o u ld
b e r e p o rte d in o th e r s e p a r a tio n s , col. 8.


(Column 11)

In c lu d e a ll a d d itio n s to th e e m plo y m en t
ro ll o th e r th a n n ew h ire s. T h is in c lu d e s
a ll em p lo y ees c a lle d b a c k to w o rk b y th e
e m p lo y e r fr o m a la y o ff a s defined f o r
col. 7, t r a n s f e r s fro m o th e r e s ta b lish ­
m e n ts o f th e co m p an y , a n d fo r m e r em ­
p lo y ee s r e tu r n in g fr o m m ilit a r y le a v e o r
o th e r a b s e n c e s w ith o u t p a y who have

ru le s, d is h o n e s ty , la z in e ss , a b s e n te e is m ,

been counted as separations.

(Column 7)

A la y o ff is a s u sp e n sio n fr o m p a y s t a t u s
( la s t in g o r e x p e c te d to l a s t m o re th a n 7
c o n se c u tiv e c a le n d a r d a y s w ith o u t p a y )
in itia te d b y th e e m p lo y e r without prej­
udice to the worker f o r su c h r e a s o n s a s :
lack o f o r d e r s , m o d el c h a n g e o v e r, t e r ­
m in a tio n o f s e a s o n a l o r te m p o r a r y
em p lo y m e n t, in v e n to ry - ta k in g , in tr o ­
d u c tio n o f la b o r s a v in g d e v ic e s , p la n t
b re a k d o w n , s h o r t a g e o f m a te r ia ls , e t c . ;

A C C E S S I O N S (Columns 12 through H )
E n t e r th e n u m b e r o f w om en in clu d ed
in th e fig u r e s re p o rte d f o r co ls. 4, 5, a n d
9. I f n o s e p a r a tio n s o f w om en o c c u rre d
d u r in g th e m o n th , e n te r “ 0 ” in co ls. 12
a n d 1 3 ; i f n o a c c e s sio n s, e n te r “ 0 ” in
col. 14.

include temporarily furloughed em­
ployees a n d employees placed on unpaid


(Column 8)
T h is g r o u p sh o u ld in clu d e o n ly
te rm in a tio n s o f em p lo y m e n t
f o r m ilit a r y d u ty la s t in g o r
e x p e c te d to l a s t m o re th a n 30
c a le n d a r d a y s , r e t i r e m e n t ,
d e a th , p e rm a n e n t d is a b ility ,
fa ilu r e to m e e t th e p h y s ic a l
s t a n d a r d s re q u ire d , a n d t r a n s ­
f e r s o f e m p lo y e e s to a n o th e r
e s ta b lis h m e n t o f th e c o m p an y .
N O T E : I f y o u in clu d e a n y
o th e r ty p e s o f s e p a r a tio n s in
th is co lu m n , m en tio n th e n u m ­
b e r an d ty p e u n d e r C o m m e n ts.


(Column 19)


(Column 16)

E n t e r th e t o t a l n u m b e r o f p e r s o n s (b o th s e x e s ) o n t h e p a y r o l l s o f th e
e s t a b l i s h m e n t ( s ) c o v e r e d in t h i s r e p o r t w h o w o r k e d f u l l- o r p a r t - t i m e o r
r e c e iv e d p a y f o r a n y p a r t o f th e p a y p e r io d ( p r e f e r a b l y o n e w e e k )
w h ic h in c lu d e s t h e 1 2 th o f th e m o n th .
I n c l u d e s a l a r i e d o ffic e r s o f
c o r p o r a t i o n s , e x e c u t i v e s a n d t h e ir s t a f f s , a n d e m p lo y e e s e n g a g e d in
f o r c e - a c c o u n t c o n s t r u c t io n b u t e x c lu d e p r o p r i e t o r s , m e m b e r s o f u n in c o r ­
p o r a t e d f ir m s , a n d u n p a id f a m i l y w o r k e r s . I n c l u d e p e r s o n s o n v a c a t i o n s
a n d s i c k le a v e f o r w h ic h t h e y r e c e iv e d p a y d ir e c t l y f r o m y o u r f ir m f o r
t h e p a y p e r io d c o v e r e d b u t exclude p e r s o n s o n l e a v e w ith o u t c o m p a n y
p a y t h e e n t i r e p e r io d a n d p e n s io n e r s a n d m e m b e r s o f th e A r m e d F o r c e s
c a r r i e d o n th e r o l l s b u t n o t w o r k in g d u r i n g t h e p a y p e r io d c o v e r e d .
N O T E : I f t h e n u m b e r d i f f e r s f r o m th e “ A ll E m p l o y e e s ” t o t a l r e p o r t e d
o n t h e E m p lo y m e n t , P a y r o ll , a n d H o u r s f o r m , e x p la in u n d e r C o m m e n t s .

(Column 17)

E n t e r th e n u m b e r o f w om en in clu d ed in th e fig u re re p o rte d f o r col. 16.
I f yo u h a v e no w om en em p lo y e e s, e n te r “ 0 ” .

E n t e r th e m a in f a c t o r s
re sp o n sib le f o r s ig n ifi­
c a n t m o n th - to - m on th
c h a n g e s in L a b o r T u r n ­
o v e r ite m s (c o ls . 4
th ro u g h 14) a n d E m p lo y ­
m e n t (c o ls. 16 a n d 1 7 ).
S o m e e x a m p le s a r e liste d
in th e h e a d in g o f coi. 19.



The State agency uses the information provided
on the schedule to develop labor turnover rates for
the States and for metropolitan areas, and for­
wards the data to Washington, where they are used
by the Bureau of Labor Statistics to prepare rates
at the national level.

Sampling is used by B L S for collecting data in
its labor turnover statistics program, since full
coverage would be prohibitively costly and time
consuming. The sampling plan for the program
must: (a) provide the preparation of reliable
monthly estimates of labor turnover rates which
can be published promptly and regularly; (b)
through a single general system, yield consider­
able industry detail for metropolitan areas, States,
and the Nation; and (c) be appropriate for the
existing framework of operating procedures, ad­
ministrative practices, resource availability, and
other institutional characteristics of the program.
In developing the sample design, the universe of
establishments was stratified first by industry and
within each industry by size of establishment in
terms of employment. Within each industry, an
optimum allocation design was obtained by
sampling with probability proportionate to aver­
age size of employment within each of the strata.
The total size of sample regarded as necessary to
produce satisfactory estimates of employment was
distributed among the size cells on the basis of
average employment per establishment in each cell.
In practice, this is equivalent to distributing the
predetermined total number of establishments re­
quired in the sample among the cells on the basis
of the ratio of employment in each cell to total
employment in the industry. Within each stratum,
the sample members are selected at random.
Under this type of design, large establishments
fall into the sample with certainty. In all manu­
facturing industries, establishments with 250 or
more employees are included in the sample with
certainty, and in some industries the cutoff is lower.
The sizes of the samples for various industries were
determined empirically on the basis of experience.
1For the national sample, additional reports needed for State
and area samples are added to those required by the national

The sample design, although aimed primarily at
meeting the needs of the national program, pro­
vides a technical framework within which State
and area sample designs can be determined.
Since, however, the rates for States and areas are
not generally prepared at the same degree of in­
dustry detail as the national rates, the national
design usually provides sufficient reports for the
preparation of State and area rates.1

Estimating Procedures
Turnover rates are estimates of ratios. For in­
dividual industries, the rates are computed by di­
viding the number of turnover actions of each type,
as reported by the sample establishments, by the
total number of employees reported by those estab­
lishments. The result is multiplied by 100. In an
industry sample, for example, 623 employees quit
between January 1 and 31, while 30,062 employees
worked or received pay during the week of Jan ­
uary 11-17. The January quit rate for the indus­
try is:

Turnover rates for industry groups are com­
puted by weighting the rates for the component
industries by the estimates of total employment,
prepared by the B L S industry employment sta­
tistics program. These estimates, which cover the
pay period including the 12th of the month, are
described in chapter 2 of this bulletin. Bates for
“all manufacturing” and for the durable and non­
durable goods subdivisions of manufacturing
are weighted by employment in the major industry
Women turnover items are subtracted from total
turnover items to obtain turnover items for men
for each sample report. Turnover rates for men
and women for individual industries are computed
by dividing the total number of turnover actions
by the corresponding employment totals for men
and women. Separate men and women turnover
rates for industry groups are obtained by weight­
ing the rates in individual industries by the em­
ployment estimates for men and women.
As of 1965, size stratification was not used in the
preparation of labor turnover rates. Until 1963
and 1964, a number of State agencies, including



those for several large States, were not fully inte­
grated into the program. Until this integration
was completed, the use of size stratification did not
appear to be desirable because of imbalances in the
sample both by State and industry. In 1965, it
was planned to introduce size stratification into the
preparation of labor turnover rates as soon as

Seasonally Adjusted Series
Many economic statistics, including labor turn­
over rates, reflect a regularly recurring seasonal
movement which can be measured on the basis of
past experience. By eliminating that part of the
change which can be ascribed to usual seasonal var­
iation, it is possible to observe the cyclical and
other nonseasonal movements in these series. Sea­
sonally adjusted labor turnover rates are published
at the all manufacturing industry level.
The seasonal adjustment method used for these
series is an adaptation of the standard ratio-tomoving average method, with a provision for
“moving” adjustment factors to take account of
changing seasonal patterns. A detailed descrip­
tion of the basic method is given in appendix A
of this bulletin.

The B L S publishes, on a national basis, monthly
series of labor turnover rates for selected indus­
tries. These series are currently published for the
manufacturing division, the durable and nondur­
able goods subdivisions, 21 major industry groups
in manufacturing, 194 individual manufacturing
industries, and 7 categories in mining and commu­
nications. Bates are available for all manufac­
turing from January 1930 and for telephone and
telegraph from 1943. For industry groups and
individual industries in the manufacturing and
mining divisions, all series begin with January
1958. Bates for certain highly seasonal industries,
for example canning and preserving, are not now
published separately but are included in the com­
putation of rates for the major manufacturing
groups. Before 1958, these industries and the
printing, publishing and allied industries major
group were not included in the rates for all manu­

facturing. The rates for all manufacturing for
years prior to 1958 were revised, however, to reflect
the influence of these industries.
Industry turnover rates on a national basis are
prepared by the B L S for all employees and for
men and women separately. For all employees,
monthly rates for total accessions, new hires, total
separations, quits, and layoffs are shown for manu­
facturing and mining industries. Except for the
new hire rates, the same items are published for
the telephone and telegraph industries. For men
and women, rates for total accessions, total separa­
tions, and quits are published for manufacturing
industry groups and subdivisions.
Preliminary turnover rates for the 21 major
industry groups in manufacturing are published
monthly in a B L S press release about a month
after the reference month, and in the Monthly
Labor Review 3 months after the reference month.
Preliminary rates for both detailed industries and
broad categories are published in Employment and
Earnings and Monthly Report of the Labor Force

about 2 months after the month of reference.
Separate rates for men and women are published
in Employment and E arnings and Monthly R e­
port of the Labor Force on a quarterly basis, for
the first month of each calendar quarter.
The labor turnover rates for all manufacturing
for selected States and metropolitan areas are
published each month in Employment and E arn ­
ings and Monthly Report of the Labor Force .
Detailed industry rates are available in releases
issued by the cooperating State agencies.
National labor turnover rates (monthly data
and annual averages) back to the beginning of
each series are published in the volume called E m ­
ployment and Earnings Statistics fo r the United
States . New editions of this volume are published

annually, following each adjustment of the Bu­
reau’s industry employment statistics series to new
benchmark levels.

Uses and Limitations
The two major causes of change in labor turn­
over rates are industrial expansion and contrac­
tion. In prosperous times, quit rates and accession
rates are high because of job availability; in
periods of economic recession, high layoff rates are




coupled with low quit and accession rates. Turn­
over rates are, therefore, regarded as good eco­
nomic indicators and are widely used by economic
analysts in both government and private industry.
Labor turnover rates by industry are also valu­
able for personnel planning and analysis. Em­
ployers use these rates as a yardstick against which
to measure the performance of their plants. For
example, they consider low quit rates to be an
indication of efficient operations and good labormanagement relations. A consideration of turn­
over is essential for scheduling production and
for planning the orderly recruitment and mainte­
a Because they are liable to misinterpretation, the Bureau does
not prepare cumulative annual rates of labor turnover. For
example, an annual quit rate could be obtained by dividing the
total number of quits during the year by average employment
during the year. An approximation of this figure can be ob­
tained by cumulating the 12 monthly rates. Suppose the annual
rate thus obtained amounted to 50 per 100 employees. This
might seem to imply that 50 percent of all employees in January
voluntarily left their jobs by the end of December. However,
many jobs in a given establishment are vacated and refilled more
than once during the year. The Bureau does not have infor­
mation on the number of employees who remained with the
establishment during the entire year. Over short periods of time,
labor turnover rates probably include relatively little repetitive
counting of employees who have held the same jobs, while over a
period of as long as a year there is considerable duplication.


m eth o d s

nance of an adequate manpower supply. Labor
turnover rates are also widely used by State em­
ployment services to plan and appraise their
The use of turnover rates to interpret changes
in the B L S monthly employment series is limited
for the following reasons: (1) The labor turnover
series measures changes during the calendar month,
while the employment series measures changes
from midmonth to midmonth; and (2) employees
on strike are not counted as turnover actions, al­
though such employees are excluded from the
employment estimates if the work stoppage lasts
throughout the report period including the 12th
of the month.
The Bureau publishes annual averages of labor
turnover rates, which are computed as the arith­
metic means of the 12 monthly rates. These can
provide a useful measure if a 1-month rate is not
suitable for some purposes, as for example when
the rate for a specific month is considered to be un­
usual or affected strongly by seasonal influences.2
— J




ym er

Office of Manpower and Employment Statistics

Chapter 4. Employment of Scientific and Technical Personnel
The growth in industrial employment of scien­
tists and engineers has been at once a symptom and
a driving force in the rapid technological progress
of American industry. For a number of years, a
major element in this growth has been the needs of
the Federal Government for increasingly complex
and technologically advanced National Defense
materiel. With expanding activities in space re­
search and technology, the Federal Government’s
needs, fulfilled largely through private industry,
are increasing rapidly. The surveys of scientific
and technical personnel, though still in their de­
velopment stage to an important degree, provide
a basis for evaluating future requirements for
such personnel.
The surveys of scientific and technical personnel
were initiated by the Bureau of Labor Statistics
in 1959, under the sponsorship of the National
Science Foundation. They are part of a compre­
hensive statistical program, coordinated by the
Foundation, which is designed to yield estimates
of the employment of scientists and engineers in
all sectors of the U.S. economy. The Bureau of
Labor Statistics conducts a survey of scientific and
technical personnel in industry on an annual basis,
and related surveys of State government agencies
and of local governments at less frequent intervals.
Since 1962, the Bureau of Labor Statistics has had
full responsibility for the conduct and publication
of these surveys, which continue to be planned in
consultation with the National Science Founda­

Description of Surveys
The surveys of scientific and technical person­
nel in industry gather data on the employment of
engineers as a group, and on scientists and techni­
cians by major occupational specialty. Engineers
and scientists are further distributed according to
whether they are primarily engaged in research
and development activities, in management and
administration, in technical sales and service, in
production and operations, or in “all other” func­

tions. In addition to the numbers of scientific and
technical personnel, data are also collected on the
total numbers of economists, statisticians, and psy­
chologists. Data are published separately for all
major industry groups, and in finer industry de­
tail for a selected number of major industry
groups. Beginning in 1966, data will be presented
also by geographic areas.
State government agencies and local govern­
ment units are surveyed to obtain information on
employment of engineers, scientists, technicians,
economists, statisticians, and psychologists. These
surveys also obtain employment data for several
health-service occupations. Employment is tabu­
lated by the various governmental functions in
which the workers are engaged. Data are also
published by State where appropriate.
For each professional occupation, respondents
are asked to report the number of persons whose
current positions require knowledge equivalent at
least to that acquired through completion of a 4year college course with an appropriate academic
major, regardless of whether they hold a college
degree. These surveys, thus, cover all persons
actually working in one of the designated occupa­
tions, specifically including those who do not hold
an appropriate degree or any degree at all, but
specifically excluding persons trained in the occu­
pation but currently employed in positions not
requiring the use of such training.
Technicians are defined as persons actually en­
gaged in technical work at a level which requires
knowledge of engineering, mathematical, physical
and life sciences, comparable to that acquired
through technical institutes, junior colleges, or
other formal post-high school training less exten­
sive than 4-year college training, or through
equivalent on-the-job training or experience.

Data Sources
The sources of the occupational data reported
by respondents are personnel records, and, espe­
cially for the smaller reporting units, personal
knowledge on the part of the persons completing
the reports. Discussions with a number of large



respondents have taught that their records typi­
cally contain much of the data in the desired form,
but that some adjustments by the respondent are
often necessary because the occupational classifica­
tions used in his records differ somewhat from
those specified for the surveys.
The employment benchmarks for the survey of
scientific and technical personnel in industry are
derived from employment data tabulated from the
first quarter reports of the unemployment insur­
ance program. The benchmarks used for the local
government surveys are derived from the periodic
censuses of local government units, by the Bureau
of the Census. The survey of State government
agencies does not use benchmarks, since the survey
is based on a compilation of data from all State
government agencies employing personnel in occu­
pations covered by the survey.

Collection Methods
Data are collected from respondents primarily
by mail, but personal visits are made to many large
employers, and to other respondents who indicate
particular difficulty in completing the question­
naires. These visits, which limited resources have
prevented from being either numerous or frequent,
are carried out by senior staff members assigned to
the surveys. There are normally two follow-up
mailings, and a sub-sample of residual nonre­
spondents are contacted further by telephone.
The response experience in these surveys has
been very encouraging. Kespondents supplying
usable information have constituted approxi­
mately 90 percent of the reports solicited in virtu­
ally every year, and have never been below 85

The basic sample for the industry survey was
drawn from lists of establishments reporting to
each of the 51 (State and D.C.) employment secu­
rity agencies for unemployment insurance (UI)
purposes, supplemented by a list of interstate rail­
roads and related companies supplied by the
Interstate Commerce Commission. Industry
classification of establishments is based on infor­
mation available to the State agencies.

Certain categories of establishments are elimi­
nated from the master list before the sample is
selected, either because a separate survey of the
given category is being made or because the num­
ber of scientific and technical personnel employed
are believed to be negligible. The categories of
organizations omitted are those classified accord­
ing to the Standard Industrial Classification sys­
tem in the following major industry groups: 01
and 02—farms; 80—medical and other health serv­
ices (except 807, medical and dental laboratories,
which is included); 82—educational services; 84—
museums, art galleries, and botanical and zoologi­
cal gardens; 86—nonprofit membership organiza­
tions; 88—private households; 89—miscellaneous
services (except 8911, engineering and architec­
tural service, which is included); 91 through 94—
government; and 99—nonclassifiable establish­
Establishments below a specified minimum size,
determined separately for each industry group,
are also excluded from the listing prior to
sampling. Because of the large number of estab­
lishments in the small size groups, minimum size
cutoffs are essential to the efficiency of the survey.
Since the excluded establishments employ very
few persons working as scientists, engineers, or
technicians, survey results are affected little by
these omissions.
Sample numbers are allocated among the various
industry-size strata in accordance with the prin­
ciple of optimum allocation, taking into account
the expected response rates by industry and by
size, so as to obtain maximum reliability within
available resources. The overall sample size is
determined so that the variance (two relative
standard errors) for the estimate of the total num­
ber of scientists and engineers for all industries
combined is about 3 percent. In every covered
industry, all establishments with 1,000 employees
or more are included in the sample. In other in­
dustry-size cells, the sampling ratios range from
1 in 1 to 1 in 100. In general, the larger the estab­
lishment and the greater the number of technical
personnel used by the industry, the higher is the
sampling ratio. All selections are made randomly
within the designated strata.
Since scientific and engineering employment is
concentrated to a significant degree in research
and development laboratories, not separately


identifiable in the U I universe, the probability
sample was supplemented. The essential rule for
unbiased supplementation is that the supplemen­
tary units must be drawn independently of the
probability sample; that is, the chance that a unit
is drawn in the probability sample must be inde­
pendent of the chance that the unit is selected as
a supplement. The initial supplementation was
drawn from a list of industrial research labora­
tories compiled by the National Academy of
Science—National Eesearch Council and from a
list of small business concerns interested in per­
forming research and development compiled by the
Small Business Administration. Beginning in
1966, when a new sample was drawn, supplemen­
tation was achieved by retaining in the mailing
list all establishments which reported employment
of 20 or more scientists and engineers in surveys
based on the previous sample; this procedure is to
be repeated whenever a new sample is drawn
(every 2 to 3 years). Establishments selected
as supplements are tabulated as a separate cell
within their industry and size class with a weight
of 1, regardless of whether they are also mem­
bers of the probability sample. (See Estimating
The mailing list for the State government sur­
veys is not a sample, but includes all agencies of
State governments which could conceivably
employ personnel in any of the designated occu­
pations. The agencies are identified from infor­
mation in directories and other documents
furnished by the States.
The survey of employment of scientific and
technical personnel in local government units is
based on a stratified random sample of the universe
of government units, other than school districts,
obtained from the Bureau of the Census. Stratifi­
cation is by State; by type of governmental unit—
county, municipality, township, and special dis­
trict; and by the 1960 population size of county,
municipality, and township, or, in the case of spe­
cial districts, by the governmental function of the
district as determined by the Bureau of the Census.
1 Symbolically,

M, where M Is the cell universe total employment,

2Pi is the sum of the primary item employment of the cell respondents, 2e<
js the sum of total employment of the cell respondents, and P ' is the primary
estimate. M is adjusted to prevent duplicate estimation for supplemental
sample reporters.


Sample numbers are allocated among the various
strata in accordance with the principle of optimum
allocation, taking into account the expected re­
sponse rates by type of government unit, and by
size or function. The sample size for each State
is determined so as to produce estimates of the
number of scientists, engineers, technicians, and
professional health workers with reasonable
sampling tolerances. The sampling ratios used
vary from 1 in 1 for all counties, municipalities
and townships with 1960 populations of 50,000
persons or more, and for special districts reported,
in the 1957 Census of Governments, to have 20
full-time employees or more, to 1 in 35 for small
units expected to employ relatively few scientific,
engineering, technical, or health personnel. All
selections are made randomly within the desig­
nated strata.

Estimating Procedures
For the survey of scientific and technical per­
sonnel in industry and for the survey of scientific
and technical personnel employed by local govern­
ment units, two different types of statistical find­
ings are derived from sample data. One type,
designated as primary estimates, includes the total
items for each question, such as total scientists and
engineers, total technicians, or total professional
health service personnel. The second type, called
secondary estimates, includes the components of
the primary estimates.
Primary estimates are obtained for probability
cells as the ratio of primary item employment to
total employment (of the reporting units in the
cell), multiplied by a total employment figure in
that industry and size class that is adjusted for
any supplemental units in that industry and size
class to prevent duplicate estimation for supple­
mental units.1 Primary estimates for supple­
mental cells are obtained by summing the primary
item employment for the supplemental reports
plus an estimate for nonrespondent supplemen­
tary units.
Secondary estimates are obtained by computing,
for each cell, a ratio of the given component item
to the corresponding primary item, using informa­
tion only for those establishments supplying data
on both items. The secondary estimates are then



derived by multiplying that ratio by the cell esti­
mate of the corresponding primary item.2
Estimates for the survey of scientific and techni­
cal personnel employed by State government
agencies are obtained by summing the reported
data. The response rate in this survey is extraor­
dinarily high—of the order of 96 to 98 percent—
and examination of the nonrespondents shows that
the number of scientists, engineers, and other per­
sonnel employed by them is negligible.

Analysis, Interpretation, and
A report on the findings of each survey is pub­
lished, usually within two years of the reference
date of the survey. Each report consists of an
analytical interpretation of the findings, sup­
ported by a statistical appendix containing in tab­
ular form all of the data that can be meaningfully
derived from the survey.

Uses and Limitations
The data derived from these surveys form the
essential statistical base (1) for evaluating the
adequacy of the scientific and technical manpower
resources of the United States in the light of the
demand placed or to be placed upon them, and (2)
for determining the rate of growth of these re­

sources. They have been used in evaluating the
impact of new or materially enlarged Federal pro­
grams, calling for substantial scientific and techni­
cal manpower input, upon the continuing needs of
the economy in general and of national defense in
particular. For example, there have been special
studies of the impact of the Apollo Program and
of the Supersonic Aircraft Program. One of the
earliest uses of these studies was in connection with
the impact upon the economy of the increased mili­
tary demands of the Korean War. Another use of
these data is as the basis for long-range projections
of future manpower requirements in science and
engineering. (See The Long-Range Demand for
Scientific and Technical Personnel, N SF 61-65.)
Still another use is to provide data for occupa­
tional guidance counselors and others who are con­
cerned with providing young people with informa­
tion on which to base a career choice.
The estimates derived from these surveys must
be interpreted as approximations. All of the sur­
veys are subject to possible response and process­
ing errors, although these are reduced insofar as
possible, through checking procedures and through
correspondence with reporters whose data are in­
ternally inconsistent or appear to involve misinter­
pretations of definitions or other instructions. In
addition, estimates derived from sample surveys
are limited by sampling error.
P * where * * ' 18 8,5 defined 111 footnote 1 ,2c» is the sum of a given
component item employment of the cell respondents, 2p»* is the sum of the
primary item employment of the same respondents, and C' is the secondary

Technical References

1. U.S. Department of Labor, Bureau of Labor Statistics. Em ploym ent o f Scientific an d Technical
Personn el in In d u stry , 1962 (BLS Bulletin 1418, 1964).
2. ----- . Em ploym ent o f Scientific an d Technical Personnel in State Government A gencies , 1962 (BLS
Bulletin 1412, 1964).
3. ----- . National Science Foundation, Em ploym ent o f Scien tific a n d Technical Personn el in State
Government A gen cies , Report on a 1969 Survey. NSF 61-17 (1961).
4. Reports prepared for the National Science Foundation by the U.S. Department of Labor, Bureau
% of Labor Statistics: National Science Foundation, Scien tific an d T echnical Personn el in A m erican
I In d u stry , Report on a 1959 Survey . NSF 60-62 (1960).
5. ----- . National Science Foundation, Scientific an d Technical Personn el in In d u stry , 1960. NSF
61-75 (1961).
6. ----- . National Science Foundation, Scientific an d Technical Personnel in In d u stry , 1961. NSF
63-32 (1963).

— William L. C opeland
Office of Manpower and Employment Statistics

Chapter 5. Occupational Outlook
The occupational outlook program originally
stemmed from a report of the Advisory Commit­
tee on Education appointed by President Roose­
velt. This Advisory Committee recommended, in
1938, that an occupational outlook service be set up
in the Bureau of Labor Statistics to make studies
and provide information for use of individuals in
choosing a career, and for the use of those responsi­
ble for the planning of education and training
programs. In 1941, the Occupational Outlook
Service was organized under a specific authoriza­
tion by the Congress. Although the first, prelimi­
nary studies were begun in 1941, it was not until
after World War II that the occupational outlook
staff was able to devote its efforts to the prepara­
tion of occupational reports for use in guidance.
In mid-1946, a manual of occupational outlook in­
formation was prepared for use in the Veterans
Adminstration (VA) counseling and rehabilita­
tion program.
In response to a resolution by the National
Vocational Guidance Association, calling upon the
Congress to authorize this type of information for
sale, and to requests by other private individuals
and groups, the first edition of the Occupational
Outlook Handbook was published in 1949. The
favorable public response to the Handbook was a
major factor in the Bureau’s decision to issue, with
the backing of the VA, a revised and enlarged
edition, which was released in 1951.
Following the conclusion of the Korean hostil­
ities, there was a sharp increase in public recogni­
tion of the key role of vocational guidance in staff­
ing essential occupations and effectively utilizing
the Nation’s manpower resources. This resulted
in the Congress in 1955 providing for the mainte­
nance of the Occupational Outlook Handbook and
its related publications on a regular, continuing,
up-to-date basis. In 1957, the third edition of the
Occupational Outlook Handbook was published;
also in that year, the Occupational Outlook Quar­
terly was originated as a companion piece to the
1 ‘‘America’s Industrial and Occupational Manpower Require­
ments, 1964-75,” prepared for the National Commission on
Technology, Automation, and Economic Progress, published in
Technology and the American Economy, Appendix Vol. I, Feb­
ruary 1966.

Handbook . The 1957 Handbook was followed in

due course by the 1959, 1961, 1963-64, and 196667 editions of the Handbook .

Description of Program
Under the occupational outlook program, the
Bureau of Labor Statistics conducts research in,
and provides information on, future occupational
and industry manpower requirements and re­
sources. It provides vocational guidance informa­
tion on expected employment opportunities for the
use of counselors, educators, and others helping
young people in choosing a field of work. It also
provides manpower information for local and na­
tional training authorities and policy makers for
use in developing programs of education and train­
ing. The results of the research are published in
the Occupational Outlook Handbook , the Occupa­
tional Outlook Quarterly , and special bulletins,
reports, and pamphlets.
In its 2y2 decades of industry and occupational
research, the occupational outlook program has
systematically accumulated and analyzed consider­
able manpower information on such topics as
employment trends for major industries of the
economy and for most major occupations; employ­
ment effects of a great many long-term programs
of government agencies, including those for de­
fense, highways, scientific research, space technol­
ogy, medical care, and education; and changes in
industry and occupational requirements.
Toward providing an overall framework of
future manpower requirements for the economy
as a whole, projections are developed for the broad
industry and occupational groups, and have been
published regularly.1 Every other year, hundreds
of different, detailed occupational and industry
statements are published in the Occupational Out­
look Handbook . In most of these occupational
outlook statements, information is provided on:
nature of work; places of employment, education
and training requirements; employment outlook
for about 10 years ahead, including, in most cases,
estimates of annual requirements for growth and
replacement needs; and earnings and working con­
ditions. In presenting outlook statements for in45



dustries, information is included on nature and
location of each industry and other industry char­
acteristics, as well as information on the industry’s
major occupations.
In presenting the employment outlook for an
occupation, information is provided not only on
the demand for workers but also on the potential
supply of workers from many sources—schools
and other training institutions, transfers from
other occupations, and reentries to the labor force.
The balance between supply and demand, in those
occupations for which an assessment is possible,
gives some indication of the nature of job competi­
tion in a specific field facing young people in the
years ahead.
In addition to the overall and detailed industry
and occupational projections developed for the
Handbook and described earlier, special manpower
studies are prepared, as part of the occupational
outlook program, that provide information, nar­
rower in scope and greater in depth, on the chang­
ing industrial structure and occupational compo­
sition of American industries, such as railroads,
civil aviation, and electronics.2 Other special
studies provide more technical information and
quantitative projections of manpower require­
ments and resources in specific occupations, indus­
tries, or for specific groups of workers—such as
teachers, technicians and nonwhite workers—
which include consideration of the current and
future demand-supply relationships and their im­
plications.3 Still others discuss only manpower
requirements trends and projections, especially in
2 Employment and Changing Occupational Patterns in the Rail­
road Industry, 1947-60 (B L S Bulletin 1344, 1963) ; Employment
Requirements and Changing Occupational Structure in Civil Avia­
tion (B L S Bulletin 1367, 1964) ; and Employment Outlook and
Changing Occupational Structure in Electronics Manufacturing
(B L S Bulletin 1363,1963).
* Maxine G. Stewart, “A New Look at Manpower Needs in
Teaching,” Monthly Labor Review, June 1964, pp. 639-644;
Technician Manpower: Requirements, Resources, and Training
Needs (B L S Bulletin 1512:, 1966) ; Jo e L. Russell, “ Changing
Patterns in Employment of Nonwhite Workers,” Monthly Labor
Review, May 1966, pp. 503-509; and “ Job Hopes High for
Record College Class,” U.S. Department of Labor, Press Release,
June 5, 1966 (USDL 7246).
4 Allan F. Salt. “ Estim ated Need for Skilled Workers, 1965-75.”
Monthly Labor Review, April 1966, pp. 365-371; and Joseph F.
Fulton, “ Employment Impact of Changing Defense Programs,”
Monthly Labor Review, May 1964, pp. 508-516; and Bernard
Yabroff, “ Trends and Outlook for Government Employment,”
Monthly Labor Review, March 1965, pp. 285-291.

those occupational groups where the supply of
workers is difficult to estimate, such as skilled
workers and workers in defense-related employ­

Sources of Data
The projections and other manpower informa­
tion developed in the occupational outlook pro­
gram utilize a wide variety of data sources, which
vary mainly with the particular occupation or
industry under examination. The following sec­
tions indicate some of the major sources of statisti­
cal and other information utilized in the program.
The basic statistics on current and past employ­
ment in occupations and industries have been based
mainly on Bureau of Labor Statistics household
data from the Monthly Report on the Labor Force
(M BLF) and establishment data from Em ploy­
ment and E arnings . (A single publication start­
ing with the February 1966 issue). Use is made
also of the scientific and technical personnel sur­
veys conducted by the Bureau, which contain data
on scientists, engineers, and technicians. The
decennial Census of Population is utilized for
data on most occupations not covered by the lim­
ited detail published in the M KLF and Censuses of
Business and Manufacturing are used to fill in
industry detail. Information from the Civil Serv­
ice Commission is used for data on Federal
Government workers. These basic sources of
occupational and industry employment statistics
are augmented by data from Federal regulatory
agencies, such as the Federal Aviation Agency
and Interstate Commerce Commission, which col­
lect industry and occupational statistics. In some
cases, employment data are obtained from unions,
industries, trade associations, and professional
societies. In most cases, however, these general
statistics serve only as a starting point for the
development of the specific, current estimates
needed for a particular report.
In developing analyses of past and projected
changes in employment requirements—which will
be described later—the outlook program utilizes
statistics of output, hours of work, and output per
man-hour. The maj or sources of the statistics used
are Bureau of Labor Statistics studies of produc­
tivity and technological development, the Federal


Reserve Board production indexes, and the U.S.
Department of Commerce output data from the
Annual Survey of Manufactures and the Census
of M anufactures . Industry associations and
unions also often provide similar types of data.
Estimates of the past and probable future sup­
ply of workers utilize entirely different sources of
information. U.S. Office of Education data on
enrollments and degrees in high school, post-sec­
ondary schools, and colleges and universities form
a major component of the supply estimates.
Bureau of Apprenticeship and Training statistics
on apprenticeship, as well as information on com­
pany training programs, provided through com­
pany reports and personal interviews, provide
other inputs into the supply estimates.
Special studies of various aspects of the supply
of workers provide essential information for the
development of the estimates and projections of
supply. A few examples of these types of sta­
tistical source materials are occupational mobility
studies (Bureau of Labor Statistics); Tables of
Working Life (Bureau of Labor Statistics); fol­
lowup studies of college graduates (National Sci­
ence Foundation); and many other specific types
of studies, often of a one-time nature. Earnings
information, which appears in many of the out­
look publications, is drawn primarily from wage
and earnings surveys conducted by the Bureau of
Labor Statistics, supplemented with additional in­
formation on many occupations from Federal reg­
ulatory agencies. Studies of union wage scales are
also used. Information is also drawn from reports
by the National Science Foundation, professional
societies and other groups.
Filling in gaps in the various types of statistics
used is information obtained from (1) personal
interviews with employers or others closely asso­
ciated with an industry or occupation; (2) reports
and interviews with professional or trade associa­
tions and licensing agencies; (3) union publica­
tions and officials; and (4) periodicals, trade jour­
nals, annual reports, and so on.

Methods of Analysis
The projections of requirements and resources
developed for the occupational outlook program
require varying methods of analysis, usually be­
cause of differences in the factors affecting a par­


ticular occupation or industry, but also because of
differences in the amount of data available for
analysis. The broad pattern of research, however,
is generally the same in all of the detailed, compre­
hensive occupational and industry studies.
The starting point in most studies is an analysis
of the factors affecting the demand for workers in
the occupation, and an assessment of how these fac­
tors may operate in the future. Occupational em­
ployment is affected by a host of factors. Techno­
logical change is the most often discussed factor
affecting occupational employment, but occupa­
tional changes are also influenced by other factors,
such as growth in population and its changing age
distribution, as in the case of teachers. Govern­
ment policy—relating, for example, to the magni­
tude of the defense and space programs and to
expenditures for research and development—also
plays a major role. Occupational employment is
also influenced by institutional factors, such as
union-management relationships and practices, as
in the case of railroad workers, or by the relative
supply of workers in other occupations, as for ex­
ample, the substitution effect resulting from short­
ages of engineers and their replacement by techni­
cians. Also influencing occupational employment
are changes in the total demand for the employing
firm’s product; changes in the level of income
and distribution of income among consumers, in­
dustry and government; and changing patterns of
It is apparent, in view of this multitude of fac­
tors, that no one technique can be used successfully
to project manpower requirements in all occupa­
tion and industries. The growth and decline of
each occupation is affected by its own complex of
factors. The number of teachers required, for
example, is affected by the number of pupils (re­
lated to birth rates and trends in the proportion of
children at each age who attend school) and by
trends in the ratio of teachers to pupils, which de­
pend upon education practices and available
Projections of requirements for scientists, engi­
neers, and technicians require consideration of dif­
ferent factors. These must take into account such
factors as the growing utilization of technical per­
sonnel, the increasing technological complexity of
industrial products and processes, changes in levels
of expenditures for defense, and growing research
and development activities. Requirements for



automobile mechanics are related to the number
of new automobiles and accessories and the age of
automobiles; for radio and TV repairmen, to the
number of radios and TV’s sold, and their age and
complexity; for policemen, to population and
urbanization; for truckdrivers, to improved equip­
ment and highways, and for competing methods of
For many occupations, the significant factors
influencing employment are the prospective levels
of demand for the products of the various indus­
tries in which the occupation is found, and the
effect of these changes in demand on employment
in the industries. Among the general factors
which must be considered in an analysis for an
industry study are expected changes in the total
domestic production of the industry’s product or
service, competition with other products or serv­
ices, expected technological changes, output per
man-hour, and changes in hours of work. More
specifically, in projecting the activity or produc­
tion level of an individual industry, it is necessary
first to establish the nature of the demand for an
industry’s products or services and the relation­
ship of this industry to the growth of the whole
economy. Obviously, an industry producing prod­
ucts directly for consumers will have a different
type of demand function than an industry which
is making raw materials to be used as a component
for further manufacturing.
In projecting the production of steel, for ex­
ample, consideration must be given to the expected
increase in population and the trend in steel out­
put per capita. The total requirements for steel
depend on the requirements projected for each of
the principal steel-using industries, such as the
automobile, construction, electrical appliances,
machinery, and containers industries; competition
to steel from other materials, such as aluminum
and plastics; and the import-export balance for
steel. In effect, it is necessary to project the out­
put of both domestic and foreign users of steel in
order to estimate total steel requirements. Future
industry production or activity levels can then be
translated into overall manpower requirements by
estimating changes in man hours per unit of out­
put for each industry, and by making assumptions
as to changes in hours of work.
Because of the tremendous amount of resources
necessary to make an extensive study of each in­
dustry in the economy, a more global type of anal­

ysis has been used to fill the gaps and to provide
an overall framework for the occupational and
industry projections. In making the analysis for
the products of each industry, the usual starting
point is the total demand of the economy for goods
and services; this can be apportioned among the
requirements for each major product or service
(classified by industry). The estimates of produc­
tion can then be translated into requirements for
workers in total and by occupation.
The general approach in the development of this
industry framework is to begin with the popula­
tion and labor force projections developed by the
Bureau of the Census and the Bureau of Labor
Statistics. Assumptions are made as to the size
of the Armed Forces, the level of unemployment,
annual hours of work, and output per man-hour.
Multiple correlations are made which take into
account past employment trends and relationships,
and variables such as unemployment, size of the
Armed Forces, gross national product, and popu­
lation. By this technique, preliminary projections
of manpower requirements are developed for each
industry for which adequate data are available.
The results of the multiple regression analysis
are then used as the basis for further judgment
decisions as to the level of manpower requirements
in the projected period, particularly for those in­
dustries for which detailed industry studies have
been made. The considerable amount of informa­
tion on individual industries developed in the
occupational outlook program over many years
and through discussion with representatives of in­
dustry and unions provides essential background
in making these judgments. Analysis of trends
and projections for the economy as a whole, or
for individual industries made by other groups,
such as the National Planning Association, Stan­
ford Kesearch Institute, State and local govern­
ments, and universities, also contribute to these
judgments. The adjusted overall industry pro­
jection framework is then utilized as a basis
for occupational requirements. These rough oc­
cupational projections are then analyzed and ad­
justed on the basis of the individual occupational
studies previously described, both the detail and
the control totals. (See description of IndustryOccupational Matrix, ch. 7.) In general, it may
be said that the projections are based heavily on
judgment as to the effect of the demand factors on
specific occupations.


Projections of changes in manpower require­
ments by occupation and industry provide only one
part of the inf ormation on the total number of
job openings which will need to be filled in the
years ahead. In most occupations, more workers
are needed yearly to fill positions left vacant by
those who leave the occupation to enter other occu­
pations or because of retirement or death, than are
needed to staff new positions created by growth of
the field. In estimating the total number of open­
ings likely to arise in an occupation, the occupa­
tional outlook program analyzes studies of occupa­
tional mobility among selected groups of workers,
and tables of working life.
These tables of working life, which are similar
to the actuarial tables of life expectancy used by
insurance companies, provide a basis for assessing
future rates of replacement resulting from deaths
and retirements, which are in turn affected by dif­
ferences in sex and average age of the workers in
particular occupations. Where men comprise the
great majority of workers, estimated replacement
rates for deaths and retirement usually average
between 1 and 4 percent a year. In occupations
in which women predominate, the rate is usually
much higher, and allowance must be made for the
large numbers of women who leave paid employ­
ment to get married and assume family responsi­
bilities but who return to paid employment after
marriage or raising a family. The replacement
rate among elementary school teachers, for exam­
ple, is estimated at 8 percent a year; many of these
teachers return to employment at a later date.
In appraising the overall employment oppor­
tunities in an occupation, estimates are also made
of the future supply of personnel, at least in those
fields in which the supply is identifiable. Statis­
tics on high school and college enrollments and
graduations are the chief sources of information
on the potential supply of personnel in the pro­
fessions and in occupations requiring extensive
formal education. Data on numbers of appren­
tices and graduates of vocational and technical
training programs provide some limited informa­
tion on new entrants into skilled trades. How­
ever, in many occupations most new entrants are
trained informally, through on-the-job training or
company training programs.
It is not enough to know, of course, how many
persons are being formally trained for an occupa­


tion, since not all those completing formal train­
ing or education in a particular field enter that
field upon completion of their courses. As a result,
special surveys are utilized to provide additional
information on the actual net supply of workers
from a training program or a field of study. These
include studies of employment plans of college
seniors, job placements of college graduates, and
jobs entered after completion of MDTA and other
types of training. Limited data on transfers out
of an occupation or re-entries into an occupation
are also utilized, although, in general, data on oc­
cupational mobility are available for only a few
The estimates of the future demand in an occu­
pation is then related to estimates of the future
supply to develop the employment outlook in that
field and to provide inf ormation to policy makers,
educators, and others on the implications of these

The Occupational Outlook Handbook is the ma­
jor publication of the occupational outlook pro­
gram. Oriented toward vocational guidance, the
Handbook is a basic reference source, published
every other year, which includes comprehensive
and non-technical job information on approxi­
mately 700 occupations and 30 major industries,
covering the entire spectrum of white-collar, bluecollar, and service occupations. An occupational
outlook report series provides reprints of individ­
ual statements from the Handbook .
The Occupational Outlook Quarterly provides a
continuous flow of current occupational and job
information between editions of the Handbook ,
together with the most recent information avail­
able on earnings, training requirements, and other
related topics. In addition to these two publica­
tions, developed mainly for use in vocational guid­
ance, the occupational outlook program conducts
technical and detailed studies on specific occupa­
tions and industries in order to furnish informa­
tion to manpower experts, personnel departments,
and others interested in the more technical aspects
of the Nation’s future manpower needs.
— H ow ard S t a m b l e r

Office of Manpower and Employment Statistics

Chapter 6. Projections of the Labor Force
Background and Uses
Projections of the future size of the labor force
are needed for a variety of planning purposes.
They provide a basis for establishing the amount
of employment growth the economy must generate
to maintain high levels of employment. They
serve as the basis for one approach in setting goals
for a general economic growth rate consistent with
full utilization of human resources. Projections
help to gain insight into the characteristics and
numbers of workers who will be available for
industry, and to see what this implies for educa­
tion, training, and personnel policies. In addi­
tion, labor force projections, together with
population projections, are used to estimate de­
mand for products, develop marketing plans, and
evaluate expansion programs. The U.S. Depart­
ment of Labor is particularly concerned with the
relationship between the expected labor supply
and the need for various skills and training created
by our changing technology.

Projections of the labor force as a whole and of
the separate age-sex groups are made for quin­
quennial dates usually for about 15 years ahead.
The schedule for preparing the projections has
been irregular owing, in part, to the timing of new
projections of the population of working age.
Labor force projections for 1970, 1975, and 1980
were published in February 1965, approximately
6 months after the latest population projections
became available.
Because social and economic factors affect the
supply of labor, certain assumptions need to be
made about conditions surrounding any set of
labor force projections. Generally, projections
have been made on the basic assumptions that past
trends in labor force participation would continue

into the future, and that the economy would con­
tinue to expand and maintain high levels of em­
ployment opportunity consistent with the situation
in 1955-57, when the unemployment rate averaged
about 4 percent. Another usual assumption is
that there would be no major war or significant
change in the size of the Armed Forces which
might substantially alter the previous work pat­
terns of the population. It also presupposes that
the trend toward increased school enrollment
beyond the high school level, which has a direct
bearing on labor force activity of young persons,
would be supported by adequate school facilities,
staff, and aid to students.
The general approach used in preparing the
Bureau’s labor force projections is to project the
proportion of the population in each age-sex group
or subgroup that is expected to be in the labor
force, i.e., the labor force participation rate at the
specified future date, and to apply these rates to
the expected population in each group.
In making projections for a given age-sex group
or its subgroup, the standard procedure is to fit
a line or curve to a series of points representing
the labor force participation rates for that group
for the years since 1947, and to extrapolate the
line or curve into the period covered by the projec­
tion. The procedure is modified, as appropriate,
to discount the temporary effect of factors judged
to be operative for only short periods.
The population projections used in projecting
the labor force are prepared by the Bureau of the
Census on the basis of analyzing past trends in
birth rates, death rates, and net immigration and
proj ecting these trends. Since the birth rates pose
the most uncertainty in projecting the population,
the Bureau of the Census prepares several series of
population projections on the basis of varying as­
sumptions with respect to birth rates. The uncer­
tainty of projecting birth rates does not directly
affect the level of the labor force projections 15
years ahead, since almost everyone of working age
(14 years and over) at that future date has already


been bom when the projections are made. How­
ever, the birth rates do have a bearing on projec­
tions of the labor force participation rates of
younger married women, because mothers of young
children are less likely to work. Because of this
indirect effect, it was necessary to select the one
series of population projections which seemed most
reasonable on the basis of an independent evalua­
tion of past trends in birth rates. For recent labor
force projections, series “B ” of the population
projections published in the Census Bureau’s Cur­
rent Population Reports , P-25, No. 286 was chosen.
The overall size of the labor force is built up by
age and sex, not only because the composition is
needed for many of the purposes noted earlier,
but also because the degree of labor force partici­
pation varies among the different age-sex groups,
and the historical trends in these rates also vary.
Some of the factors which help to explain the be­
havior of the labor force participation rates and
which affect particular groups include school at­
tendance, marital status, birth rates, and the avail­
ability of social security benefits and the expansion
of private pension plans. The method of project­
ing the labor force participation rates for the
various age-sex groups takes into account the in­
fluence of a number of these specific demographic
and social factors. For example, projections of
the proportion of persons enrolled in schools in the
various young ages are used to subdivide the fu­
ture population of young persons into those who
are expected to be in school and those not in school.
The population of married women in ages 20 to 44,
by age, is grouped by those who are expected to
have children of preschool age and those with no
children under 5, on the basis of projected trends
in fertility and child spacing. Similarly, proj­
ected marital status distributions of older adult
women are used to provide the future numbers in
each marital category within each age.

For each of these subgroups, the projected labor
force participation rates are applied to their re­
spective future populations and the resulting labor
force summed to provide the total labor force for
each age-sex group and for all ages.

Sources of Data
The monthly statistics on the labor force pre­
pared by the Bureau of Labor Statistics and based
on the Current Population Survey of the Bureau
of the Census is the source of the basic historical
data on labor force participation rates by age and
sex used in projecting the labor force. Historical
data on labor force activity by various categories
within several of the age-sex groups were obtained
from the recurring supplementary labor force sur­
veys also based on the Current Population Survey.
These include information from the October sur­
veys of the employment of school-age youth and
the March surveys of the marital and family
characteristics of workers.
The population projections are the latest avail­
able projections made by the Bureau of the Census
and published in their Current Population Re­
ports, Series P-25. Projections of school enroll­
ment and marital status, by age, are based on pub­
lished and unpublished data of the Bureau of the
Census. Data used in projecting the proportion
of women in each age group by presence of chil­
dren under age 5 years include published and un­
published data on birth rates, by age of mother
and order of birth, from the Division of Vital
Statistics of the Public Health Service; fertility
and marriage data from reports of the Bureau of
the Census, Current Population Reports , Series
P-20, and data from the decennial censuses of

Technical References

1. Cooper, Sophia, “ Interim Revised Projections of the U.S. Labor Force, 1965-67,” M onthly L ab o r
Review , October 1962, pp. 1089-1099, reprinted as Special Labor Force Report No. 24.
2. ----- , and Denis F. Johnston, “ Labor Force Projections for 1970-80,” M onthly L ab or Review ,
February 1965, pp. 129-140, reprinted as Special Labor Force Report No. 49.
3. P o p u latio n a n d L ab or Force P rojections fo r the U .S ., 1960 to 1975 (BLS Bulletin 1242, 1959).


— S o ph ia . C o o p e r
Office of Manpower and Employment Statistics

Chapter 7. Industry-Occupational Matrix
The Bureau of Labor Statistics has developed a
comprehensive set of data on the occupational em­
ployment composition of industry sectors repre­
senting the entire economy as of 1960. These data
are set up to form a table, or matrix, with 156
specific occupations or groupings of occupations
cross-classified with 137 industries. Thus, it is
possible to see the occupational pattern of each
industry—i.e., what proportion each occupation is
of total employment in an industry. Looked at
another way, the tabulation makes it possible to
determine how total employment in an occupation
is distributed by industry.
Initial work on the Industry-Occupational
Matrix grew out of concern by the Department of
Defense for anticipating the economic problems
that might arise from various defense programs.
The first set of tables related to 1950 and were pre­
pared by the Bureau as a part of the inter-industry
program of the early 1950’s, sponsored by the U.S.
Department of the Air Force. That program was
terminated in 1953, but the 1950 matrix and its
1960 successor continue to provide the basic in­
formation for emergency manpower planning, now
carried on by the Office of Emergency Planning.
In recent years, the strong interest in training new
workers, retraining workers displaced by automa­
tion, and providing information to high school
counselors and to students making career deci­
sions, has focused increased attention on the need
for estimates of the numbers presently employed
in detailed occupations and the likely future em­
ployment requirements by occupation. The In­
dustry-Occupational Matrix provides a systematic
approach to developing the desired information.

Sources of Data
Data for the 1960 Industry-Occupational Matrix
were brought together from a wide variety of
sources. A major source was the Occupation by
Industry report from the 1960 Census of Popula­
tion. The Monthly Report on the Labor Force

(MRLF) was the source for total employment,
employment for broad occupational groups and for
a few large, specific occupations.1 Other sources
of occupational employment data included the
Bureau of Labor Statistics annual surveys of oc­
cupational wage rates in metropolitan areas and
selected industries;2 regulatory agency statistics
on employment by occupation in the telephone,
railroad, and air tranportation industries; U.S.
Civil Service Commission statistics on employment
by occupation in the Federal Government; statis­
tics on selected professional occupations based on
licensure data and membership records of profes­
sional societies; and surveys of employers by the
Bureau and other agencies to obtain estimates of
employment in a limited number of highly im­
portant occupations such as scientists, engineers,3
teachers, and policemen.
In total, specific estimates from sources other
than the Census were entered in the detailed cells
of the matrix for about 16 million workers or onefourth of all those who were employed in 1960.
The remaining detail in the matrix was derived
by forcing 1960 population census estimates for
detailed cells (published in Occupation by In ­
dustry ) to agreement with control totals for oc­
cupational groups and industries from sources
other than the Census. The occupational control
totals were average annual employment by occupa­
tional group from the MRLF. Most of the in­
dustry employment totals are based on B L S
estimates of private wage-and-salary workers ad­
justed to include the self-employed, unpaid family
workers, and government workers, and to exclude
the secondary jobs of dual job holders. Total em­
ployment in agriculture and private households
was based on M RLF estimates. The adjustments
of the matrix to consistency with M RLF estimates
of total employment and industry employment
estimates, derived as described above, brings the
matrix for 1960 into agreement with data which
are the basis for projections of total employment
and employment by industry, being made by the
Bureau. Employment estimates in the matrix
1 See ch. 1.
2 See ch. 14.
3 See ch. 4.


may be modified as additional information is

Analysis and Uses
A basic objective of the project is to have avail­
able a comprehensive set of data on industryoccupational relationships which can be used in
projecting manpower requirements by occupation.
Although statistics on employment by occupation
are relatively thin, particularly between decennial
censuses, there is a great deal of information on
total employment in detailed industries. Each
industry utilizes a unique combination of occupa­
tional skills together with other factors of produc­
tion in its efforts to achieve least cost for its output.
Occupational patterns may be markedly different
from one industry to another. For example, em­
ployment in the insurance industry is primarily of
white-collar workers such as insurance agents,
office clerical workers, actuaries, and others. In
contrast, the workforce in restaurants is largely
made up of waiters, waitresses, cooks, and ownermanagers. Over periods as short as a decade, the
occupational structure of many industries is
relatively stable. Consequently, if good informa­
tion is available on the occupational composition
of individual industries for a base period, it can be
used together with the available statistics on chang-*
* See Chapter 30.


ing employment in each industry to develop esti­
mates of current employment by occupation for
later periods. Further, if projections of output
and employment are available by industry, the base
period occupational ratios applied to the industry
employment projections will yield initial estimates
of employment requirements by occupation for
future periods.
Although the occupational patterns of many
industries are relatively stable over periods of 10
to 15 years, it is clear that these patterns change
with the advance of technology and changes in the
supply of workers in each occupation. Hence,
information on how technology and labor supply
are changing the occupational pattern in each
industry is used to modify the initial estimates.
This improves the estimates of current employ­
ment by occupation and of future employment re­
quirements by occupation, developed by applying
base period industry-occupational ratios to in­
dustry employment estimates. Changing tech­
nology and other factors which affect skill require­
ments are being studied in an effort to estimate the
future occupational structure of each matrix in­
dustry. The adjusted occupational patterns are
being used, together with projections of employ­
ment by industry from the Government’s Economic
Growth Project,4 to estimate future employment
requirements for about 100 occupations.
— H arry G reenspan
Office of Manpower and Employment Statistics

Prices and Living Conditions
Chapter 8. Consumer Expenditures and Income
Consumer expenditure surveys1 are specialized
family living studies in which the primary em­
phasis is on collecting data relating to family
expenditures for goods and services used in dayto-day living. Expenditure surveys of the Bureau
of Labor Statistics also include information on
the amount and sources of family income, on
changes in savings or debts, and on major demo­
graphic and economic characteristics of family
The Bureau’s studies of family living conditions
rank among its oldest data-collecting functions.
The purpose of the first nationwide expenditure
survey in 1888-91,2 in line with the legislation
creating the Bureau, was to study the worker’s
consumption habits and living costs as elements
of production costs, with special reference to com­
petition in foreign trade. It emphasized the
worker’s role as a producer, rather than as a con­
sumer. Purposes and coverage changed in suc­
cessive surveys, and problems caused by rising
prices led to the second survey, during the year
1901. The index of prices of food purchased by
workingmen, with weights based on the 1901 data,
was used generally as a deflator for workers’ in­
comes and expenditures for all kinds of goods
until World War I. The third major survey,
spanning 1917-19, provided weights for comput­
ing a “cost-of-living” index, now known as the
Consumer Price Index (CPI). (See chapter 10.)
The next major study, for 1984-36, was made
primarily to revise these index weights and cov­
ered only urban wage and clerical workers.
However, in the severe economic depression of
the 1930’s, interest in consumer surveys expanded
from study of the welfare of selected groups to
general economic analysis. Thus, almost simul­
taneously with its 1934-36 investigation, the Bu­
reau cooperated with four other Federal agencies
in a fifth survey, the Study of Consumer Pur­
chases, in 1935-36, which undertook to show con­
sumption of all segments of the population, both
urban and rural.3 The Bureau’s sixth major sur­

vey, for 1950, covered all urban consumers. It
provided the basis for revising the Consumer Price
Index (CPI) and also supplied abundant material
for broader types of economic and market analysis.
The remainder of this chapter deals with the
1960-61 survey—the latest in the series 4 describing
changes in the consumption habits of the American

Description of Survey
The basic orientation of the Bureau’s most recent
survey was to obtain detailed information for re­
vising the CPI. The increasing need for consumer
expenditure and income data for other purposes
was taken into account in planning the survey of
urban families in 1960 and 1961. Then, in coop­
eration with the U.S. Department of Agriculture
(U SD A ), the 1961 coverage was extended to rural
areas. Thus, for the first time since 1941, informa­
tion on spending habits became available for a
cross-section of the entire noninstitutional popula­
tion in urban, rural nonfarm, and rural farm areas
of the United States. Concepts, techniques, and
publications for the 1960-61 survey were planned
to provide as much continuity and comparability
as possible with the Bureau’s 1950 and earlier ex­
penditure surveys.
1 In this chapter, the initials CES are used to refer to the
program of consumer expenditure surveys.
2 The State of M assachusetts conducted the first expenditure
survey in the United States. This investigation of living con­
ditions in 1874-75 undertook to measure the welfare of the
workingman’s family before and after migration to the United
States. For a more detailed account of expenditure surveys in
this country, see U.S. Bureau of the Census, Historical Statistics
of the United States— Colonial Times to 1957, Chapter G (1960).
8 The Bureau also cooperated with the U.S. Department of
Agriculture in a smaller scale nationwide survey of urban and
rural families in 1941-42 to obtain facts on which to base deci­
sions for the civilian economy during wartime. In addition, the
Bureau conducted a Survey of Prices Paid by Consumers in 1944
among a nationwide sample of urban fam ilies; as a byproduct
of the reports required for the analysis of prices, certain data on
family income, savings, and expenditures were obtained.
4 In addition to these 7 m ajor surveys, the Bureau has con­
ducted a number of expenditure surveys in cities selected for
specialized studies.


All data were collected through the voluntary
cooperation of families.5 The family, or consumer
unit (CU), referred to (1) a group of people
usually living together who pooled their income
and drew from a common fund for their major
items of expense,6 or (2) a person whose income
and expenditures were not pooled with others,
whether living alone or in a household. However,
never-married children living with parents always
were considered as members of the parents’
consumer unit.
Information was recorded for the family as com­
posed in the survey year, including part-year
members. Family members were not eligible for
periods in the survey year that they lived in mili­
tary camps, posts, or reservations; in institutions;
abroad (except on vacation, etc.); or were mem­
bers of another CU.7
A complete account of family income and out­
lays was compiled for a calendar year. This ac­
count included information to determine net
changes in the family’s assets and liabilities dur­
ing the year. The estimated value of goods and
services received as gifts or otherwise, without
direct expenditures by the family, was requested
also. To supplement the annual data, families
who prepared meals at home provided a detailed
7-day record of expenditures for food and related
items purchased frequently.
For selected items of clothing, housefurnishings,
and food, the record of expenditures was supple­
mented by information on quantities purchased
and prices paid. Characteristics of the housing
occupied by homeowners and renters and an inven­
tory of the major items of housefurnishings they
owned were recorded.
To permit more meaningful analysis of the
spending habits of American families, limited
demographic information was obtained. This in­
formation included the sex, age, years of school
completed, occupation, race, and marital status of
each family member.

Data Sources and Collection Methods
All data were collected by personal interview.
The B LS was responsible for collecting data from
all residents of urban places. The B L S and
USDA shared this responsibility in the rural areas


of Standard Metropolitan Statistical Areas
(SM SA’s)8 and the USDA had sole responsibility
for interviewing rural households in nonmetro­
politan areas.9

Field Organization
To reduce the size of the staff to be recruited
and trained and to utilize this staff over a longer
period, it was decided that the urban survey would
cover 2 years, 1960 and 1961. For similar reasons,
the surveys for each year were conducted in two
4‘waves.” As field work was completed in the
largest SMSA’s, supervisors were reassigned to
smaller places. Interviews for the 1960 and 1961
CES were conducted in the spring and summer
of 1961 and 1962, respectively.
The supervisory field personnel were recruited
by the B L S Regional Offices10 and brought to
Washington for 6 weeks of intensive training on
the purposes of the survey, survey techniques, and
schedule content. These supervisors went to an
assigned city where they, in turn, recruited inter­
viewers, whom they trained for about 8 days.

The detailed questionnaires used by the B LS
agents in interviewing families in the 1960-61 sur­
vey had been tested in surveys for 1959 in three
cities. They incorporated modifications based on
this experience. Three forms were used in the
nonfarm surveys. Schedule A was a two-page
form to determine the family’s eligibility for
the survey and, for families who refused or were
unable to participate in the survey, it provided a
record of minimum data for the analysis of non­
response. Schedule B, on which the interviewer
entered the complete annual record of the family’s
living arrangements, income, spending, and
5 See discussion of bureauwide policy on voluntary reporting
and confidentiality, under Introduction.
6 This category includes children temporarily away from home
at school or college.
7 The tabulations published in the CES reports listed in the
table at the end of this chapter include only full-year consumer
units, i.e., units with at least one member who was eligible over
the entire survey year. In addition, the Bureau obtained sched­
ules from approximately 400 part-year CU’s, for special analytical
8 See appendix B for description of SMSA classification system.
9 See discussion of sample design, p. 56.
“ The following description refers to procedures of the BLS,
but USDA procedures were similar.



changes in savings, consisted of 59 pages and
formed the basic framework of the survey. Sched­
ule 0 provided 15 pages for a detailed report of
the family’s purchases of food, beverages, tobacco,
personal care, and household supplies in the week
preceding the interview.11 The coding system for
summarizing and classifying the data was devised
while the 1960-61 schedules were being designed,
and tabulating codes were printed on the schedules.
The precoded data were then readily transferred
to punch cards and magnetic tapes for tabulating.
Information in schedule B was grouped in 23
sections, placed in a sequence so as to establish
and maintain rapport between the interviewer and
the respondent. The detailed checklists of items12
in each section were included not only to facilitate
recall, but to provide the specific information
needed to determine CPI weights. However, all
sections were rarely applicable to a single family.
For example, if the family were renters, the sec­
tions relating to homeownership could be omitted.
Families were encouraged to refer to records
whenever possible.
Beported receipts and disbursements were sum­
marized and reviewed in the field to determine the
completeness, consistency, and balance of the fam­
ily account. Families were reinterviewed when
the field supervisor deemed it necessary to clarify
11 A modification of schedule B, providing for detailed reporting
of farm receipts, disbursements, changes in farm assets, and
value of home-produced food, but otherwise paralleling the non­
farm schedule, was used by the U.S. Department of Agriculture
in rural farm areas. Schedule C wasi not used in interviewing
farm families.
“ For a discussion of global versus detailed questions and
“ free-listing” versus “ check-listing,” see report by Helen Humes
Lamale, Methodology of the Survey of Consumer Expenditures
in 1950, Philadelphia, Pa., University of Pennsylvania, Wharton
School of Finance, 1999, pp. 18-19.
“ Prior to editing the schedules for machine tabulation, the
schedules were reviewed in the Washington offices of the B L S
and USDA primarily to determine that entries conformed to the
survey concepts and methodology. This review w as concentrated
on sections of the schedule which had proved most difficult in
previous surveys, and on unusual situations which required
specialized instructions. Occasionally, schedules were returned
to the field for clarification or additional information, and some
schedules considered complete in the field were rejected in Wash­
ington. If a schedule met the test of the review and editing
instructions with respect to internal completeness and comsistency of expenditures with each other and the family’s reported
manner of living, the record was used even though there w as a
substantial lack of balance between the fam ily’s reported total
receipts and disbursements.
14 This procedure involves the probability selection of a sample
“ pattern” from a set of patterns purposively established so that,
taken as a group, they give each primary sampling unit its
proper chance of appearing in the final sample. The selection
of the city sample is described in “ Technical Note—The Revised
City Sample for the Consumer Price Index,” Monthly Labor
Review, October I960, pp. 1078-1083. (B L S Reprint 2352.)

ambiguous entries or to complete a record.13 On
the average, the interviewer spent 7 to 8 hours with
a family in a series of visits arranged at the
family’s convenience.

Separate stratified samples were selected for
urban areas, rural areas in metropolitan counties,
and rural areas in nonmetropolitan counties. A
three-stage sample design was used within each
area to obtain a sample of consumer units repre­
sentative of all U.S. consumer units as defined
for this survey.
In developing urban sampling plans, continued
representativeness of the sample for measuring
national changes in consumer prices was of para­
mount importance. Consideration of probable re­
sources led to setting 66 as the maximum number
of cities for the CES sample. Tests of the ef­
fectiveness of some of the more obvious modes of
stratification indicated that no elaborate stratifi­
cation was justifiable for so small a sample of
cities. In general, classifications by geographic
region and size of city seemed to be most effective,
especially since an important objective in selecting
specific cities was to achieve good geographic dis­
persion. For this purpose, the B L S utilized the
“controlled selection” procedure.14 The primary
sampling unit was the Standard Metropolitan
Statistical Area (SMSA) in the metropolitan seg­
ment of the United States and the individual
urban place in nonmetropolitan areas.

Sample Design for Three Urbanizations
In the first stage of the design for the urban
sample, all SMSA’s and nonmetropolitan urban
places were classified by population size and
region. A sample of 66 places listed in the table
was selected to represent all urban places in the
50 States. All of the 12 largest areas in the United
States automatically were included. For New
York and Chicago, the Standard Consolidated
Areas, rather than the constituent SMSA’s, were
used as primary sampling units. However, in the
collection and analysis of the data, the New YorkNortheastern New Jersey Standard Consolidated
Area was divided into two subareas—New York,
N.Y. and Northeastern New Jersey. All of these
largest areas were surveyed in both years with data

S urvey
( S um m ary


S a m p l e S iz e




C o n s u m e r E x p e n d i t u r e s , 1960-61

A v a il a b il it y o f D a t a f o r U r b a n a n d R u r a l A r e a s ,
l a t io n S t r a t u m , a n d S M S A o r O t h e r U r b a n P l a c e )
Number of—

Region, population stratum, SMSA, or other urban place



G e o g r a p h ic R e g io n , P o p u ­

Survey year






BLS report




Total urban and rural—United States..........................................
Northeast........................................ .............. ........................ .
North Central..............—............ ......... ............................... .
South................ .............. ........................................................
West............................. ..........................................................

i 2,869

i 2,228

Rural farm—United States...........................................................
Northeast.................. -.............. -.............................................
North Central.................... -........................... -.......................
South....................................................... ...........................—
West........... —................................. ........................................




Rural nonfarm—United States............. ................... ....................
Northeast--- -------- ----------------- ----------------------North Central..............-..................._................................... South.................................. ....................................................
West..................... ................................................................. -





Urban—United States....................................................................
Northeast........... ................. ............... —.............................. SMSA, population 1,400,000 and over
Boston, M ass......... ..................................................
New York, N.Y............................... .........................
Northeastern New Jersey..........................................
Philadelphia, Pa............................. ............— .........
Pittsburgh, Pa............................................................


i 9,476








237-7 and 57
237-4 and 54
237-13 and 63
237-8 and 58
237-11 and 61


SMSA, population 250,000 to 1,400,000
Buffalo, N.Y............................................................
Hartford, Conn.—......................................................







SMSA, population 50,000 to 250,000
Portland, Maine........................................................
Lancaster, Pa.......................................... ..................






Nonmetropolitan urban place, population 2,500 to 50,000
Burlington, Vt.......................................................... .
Kingston, N.Y...........................................................
Lewistown, Pa...........................................................
Athol, Mass............................................................... .
Millville, N J — .........................................................
Southbridge, Mass.......................... .......................... .















237-5 and 55
237-21 and 71
237-1 and 51
237-15 and 65


SMSA, population 250,000 to 1,400,000
Indianapolis, Ind........................................................
Dayton, Ohio....................................................... .......
Wichita, Kans.................................... ....................... .







SMSA, population 50,000 to 250,000
Cedar Rapids, Iowa................................................... .
Champaign-Urbana, HI.............................................
Green Bay, Wis.... ..................................................







Nonmetropolitan urban place, population 2,500 to 50,000
Devils Lake, N.Dak____________________ _____
Findlay, Ohio____ ____________ __________ —
LaSalle, 111..................................................... .............
Niles, Mich___________ _____ ____ __________
Owatonna, Minn............................... ........................
Cambridge, Ohio......... ..............................................
Crookston, Minn____________________________
Logansport, Ind..........................................................
Manhattan, Kans...... .............. ...................................
Menasha, Wis................................. ...........................















237-16 and 66
3 and 53


SMSA, population 250,000 to 1,400,000
Atlanta, Ga...................................
Dallas, Tex..... ....... .....................
Nashville, Tenn............................







SMSA, population 50,000 to 250,000
Austin, Tex........................... .......
Orlando, Fla........................ ........
Baton Rouge, L a . .. ......................
Durham, N.C__............................







North Central............ -......... — ..............................................
SMSA, population 1,400,000 and over
Chicago, 111..... ..........................................................
Cleveland, Ohio.........................................................
Detroit, Mich............................................................. .
St. Louis, Mo..... ........................................................ .

SMSA, population 1,400,000 and over
Baltimore, Md...............................
Washington, D.C..........................

See footnotes at end of table.


















(S um m ary


S a m p l e S iz e


C o n s u m e r E x p e n d i t u r e s , 1960-61— Continued]gj

a n d A v a il a b il i t y o f
l a t io n S t r a t u m , a n d

D a t a f o r U r b a n a n d R u r a l A r e a s , b y G e o g r a p h ic R e g io n , P o p u ­
S M S A o r O t h e r U r b a n P l a c e ) — Continued

Urban—United States—Continued
Nonmetropolitan urban place, population 2,500 to 50,000
Cleveland, Tenn . . . _
Griffin, Ga .
. _ _
_ .. .
Me Allen, Tex _
Reserve, La
... .
TTnion, S.C
_ ... ..... _ . ____
Viekshnrg, Miss _ .... _
. _ _ __
Florence, Ala
Gainesville, Tex
Mangnm, Okla
Martinsville, Va
. _
Okmulgee, Okla
Sehring, Fla __
........ . _
. . . .







237-22 and 72
237-2 and 52













SMSA, population 250,000 to 1,400,000
Seattle, Wash
Denver, Colo.
__ _
Honolulu, Hawaii
. _




SMSA, population 50,000 to 250,000
Bakersfield, Calif _













Eureka, Calif ... . ....
Orem, Utah .

BLS report



Nonmetropolitan urban place, population 2,500 to 50,000
Anchorage, Alaska— ......................... .......................................
Gallup, N.Max ... _ _ .._
Klamath Falls, Oreg . ... _ ...... ......




SMSA, population 1,400,000 and over
Los Angeles-Long Beach, Calif.................................................
San Francisco-Oakland, Calif............................. .................. —


Survey year

Number of—
Region, population stratum, SMSA, or other urban place





* Asterisk indicates year of survey.
i Includes Anchorage, Alaska, which was surveyed for 1959.
* Published by the U.S. Department of Agriculture.
* Supplements 2 and 3 not to be published; for a limited time photocopies of
tables may be obtained from the Bureau of Labor Statistics at a nominal cost.
* No supplements available.
8 Survey for 1959.

N o te : The Bureau also has published reports for the following special-city
surveys: Cincinnati, Ohio, 1959 (Report 237-28); Fairbanks, Alaska, 1959
(Report 237-30); Ketchikan, Alaska, 1960 (Report 237-31); Juneau, Alaska,
1960 (Report 237-32); Las Vegas, Nev., 1962 (Report 237-33); Houston, Tex.,
1963 (Report 237-83); Kansas City, Kans.-Mo., 1963 (Report 237-79); Mil­
waukee, Wis., 1963 (Report 237-80); Minneapolis-St. Paul, Minn., 1963
(Report 237-81); and San Diego, Calif., 1963 (Report 237-82).

collected from half the sample of consumer units
each year. Half of the remaining sample of
smaller SMSA’s and urban places was surveyed
each year, i.e., for 1960 and 1961.
At the second stage, a sample of housing unit
(living quarter) addresses was obtained in each
city and in the entire urban part of each SMSA
from the Comprehensive Housing Unit Survey
(CHUS) conducted by B L S in the fall of 196015
or 1961. The B L S housing unit surveys were
based on area block samples designed to represent
all noninstitutional living quarters (including non­
transient accommodations in hotels and rooming
houses) in the urban area.

In the third stage, a subsample of addresses was
selected from the housing unit addresses obtained
in the CHUS or Census, arranged by a number
of characteristics, e.g., household size.
The first stage in selecting the rural sample in
metropolitan areas utilized all 34 SMSA’s selected
for the urban sample. In the second stage, B L S
conducted a Rural Housing Unit Survey (RHUS)
which consisted of a listing of housing unit ad­
dresses in a stratified sample of Census Enumera­
tion Districts (ED ’s) and a subsample of smaller
segments or blocks in the designated ED ’s. Each
housing unit was visited and classified as farm or
nonfarm, and as to whether the family included
a farm operator.16 In the third stage, subsamples
of rural nonfarm and rural farm housing unit
addresses were selected from the RHUS listing by
applying a ratio based on census data on rural
farm and rural nonfarm households in each

15 Because of the shortage of time, the B L S did not conduct a
CHUS in 1960 in places with population of 2,500 to 50,000. The
1960 sample for these cities was selected from 1960 Census of
Housing and Population listings of living quarter addresses
arranged by Enumeration D istricts (ED ’s).
16 These classifications, which were on the basis of Census
definitions, made it possible for the B L S to refer addresses of all
households meeting the Census definition of farm operator or
farm resident to the USDA for inclusion in the rural sample.


In the first stage of the USDA’s design for the
rural sample in nonmetropolitan areas , counties

were grouped by State Economic Areas17 into 126
strata equal in weighted counts of rural farm and
rural nonfarm dwellings, as the same sample of
counties was to be used for both farm and nonfarm
households. For each stratum, one county was
chosen at random with a probability proportional
to its weighted count. Counties were selected
from 41 States. At the second stage, within each
sample county, a selection of rural segments was
made separately from rural places (100 to 2,500
inhabitants) and the open country. Addresses of
all housing units in these segments were listed and
classified as farm and nonfarm. Farm operators
also were identified. In the third stage, sub­
samples of nonfarm and farm housing unit
addresses were selected from the survey listings.

Sample Size
The master sample for the total urban and rural
population included 17,283 living quarter ad­
dresses which were assigned to the interviewers.18
Usable schedules were tabulated for 13,728 con­
sumer units. The distribution of assignment
addresses and usable schedules by urbanization,
geographic region, and for individual metro­
politan areas or cities in the urban sample is shown
in the table.

Estimating Procedures
To describe the spending and saving of all fami­
lies in the United States, data from the CES
samples have been combined to obtain regional
and U.S. levels.19 This information was sum­
marized for each level of urbanization and for the
entire population by using a system of weights
based on the 1960 Census of Population.
To obtain the weights, adjustments were made in
the Census total of persons in the population on
April 1,1960, to correct for definitional differences
between the Census and the CES universe. The
institutional population and on-post military per­
sonnel, which were not included in the CES, were
deducted from the Census population. Since the
CES data apply to the full survey year and family
size is measured in year-equivalent persons, while
the Census data are a count of persons on April 1,
222-617— 66-----5


1960, the Census data were adjusted to take account
of births, deaths, and net civilian migration dur­
ing 1960. For the 50 States, the net effect of the
adjustments was to lower the population total from
179,325,671 to 177,391,360.
The total adjusted population was distributed
among the sampling strata in accordance with the
distribution of the unadjusted population. The
population represented by each surveyed area was
divided by the average family size in the area, as
determined from the survey, to obtain the total
number of families represented by each area. The
estimated number of consumer units in the universe
for the United States was 55,306,253.
The adjusted 1960 population was used as urban
weights for both 1960 and 1961. Weights were
computed for 67 urban strata,20 including Anchor­
age, Alaska, which was surveyed for 1959. (See
table.) Rural nonfarm weights were computed
for 42 strata—34 SMSA’s and a farm operator and
nonoperator stratum for each of the four regions.
The rural farm sample was designed to be self­
weighting within regions. Sample averages for
the four regions were combined to U.S. levels by
the application of weights consistent with those
weights used in the urban and rural nonfarm
parts of the CES.
In applying the weights to the stratum aver­
ages, to obtain U.S. and/or regional averages, the
blow-up factor for each class (e.g., income group,
family-size class, etc.) was the number of consumer
units in the universe represented by each sample
family in a stratum multiplied by the number of
families in the sample for that class. The result­
ing numbers of consumer units became the multi­
pliers in calculating stratum aggregates which
17 See appendix B.
18 The address following each address in the master sample was
picked as an alternate. The alternate address was substituted
if the master address could not be located or the unit was
vacant, if no one was a t home after at least 2 visits, or if the
occupant refused or was unable to give the minimum informa­
tion required for classifying the family.
If the interviewer found more than one consumer unit living
at an address, each unit was included in the sample. In hotels
and apartments, the address given to the interviewer indicated
a specific housing unit or units within the structure.
19 The samples were not designed to provide tabulations by
20 F or a single year such as 1961, the city weights differed
from the 1960-61 weights, in that cities surveyed in that year
carried the entire weight for their respective region city-size
stratum in the 1961 tabulations. In combining 1960 and 1961,
each year's sample represented approximately half of the ad­
justed population. The 1959 data for Anchorage were weighted
into the combined 1960-61 tabulation for the West and the
United States, but not into the tabulations for 1960 or 1961.



were combined to regional and U.S. levels. The
regional and U.S. averages were obtained by divid­
ing the sum of the stratum aggregates by the num­
ber of consumer units in the universe for the class.

Analysis and Presentation
Data were tabulated separately for each city and
region listed in the table. All dollar values (in­
come, expenditures, and changes in savings) were
shown as averages per family (i.e., consumer unit).
As city tabulations were completed, they were
examined for reasonableness and internal consist­
ency, and were compared with tabulations for
other cities in the 1960-61 CES sample and with
information from independent sources—princi­
pally the 1960 censuses of population and housing.
Similar analytical comparisons were made for the
regional and U.S. tabulations at each level of
urbanization. For the urban sample, trends since
1950 were analyzed for cities which were in the
CES sample for both periods. Each report con­
tained brief analytical and interpretative text, plus
definitions and statements on methodology.
The basic reports (see column 5, table) pre­
sented averages for major components of family
accounts for consumer units classified by five
characteristics: Family income after taxes, family
size, age of family head, occupation of the head,
and housing tenure. Supplement 1 presented the
same information, classified by four additional
characteristics: Education of the head, race, fam­
ily type, and number of full-time earners. At the
regional and U.S. levels, families also were classi­
fied by whether they lived inside or outside metro­
politan areas; and inside SMSA’s, by central city
and other location.
Data for eight family characteristics in the
above summaries were cross-classified (two vari­
ables) with each of other selected characteristics,
as follows:
1. Family size with income, age of head, family type and
location and size of place.
2. Age of head with income, occupation of head, and
3. Occupation of head with income, race, and tenure.
4. Education of head with income, and occupation of head.
5. Race with income, and tenure.
6. Number of full-time earners with income.
7. Tenure with income.
8. Family type with income, and occupation of head.

The two-variable tables are published as Supple­
ment 2 to the basic reports.
Supplement 3 presents in detail the components
of consumer expenditures, income, and changes in
savings, which were summarized in the basic re­
ports and Supplements 1 and 2. To illustrate, the
category “automobile transportation” is broken
down into 10 subgroups of expenditures. These
detailed tabulations provide data for consumer
units cross-classified by family size and income
after taxes, and by family size and location of the
family’s residence inside or outside SMSA’s.

Uses and Limitations
From its inception, the 1960-61 CES was
planned to serve a variety of purposes. To this
end, questionnaires and tabulation plans were
circulated among a number of Federal agencies
for comment. The B L S also formed the Con­
sumer Expenditure Advisory Committee, repre­
senting academic, private research, and marketing
users, and consulted with this Committee on a wide
range of decisions affecting the CES. To date,
the results have been used for only some of the
contemplated purposes.
As stated earlier, the primary objective of the
1960-61 CES was to revise the CPI (chapter 10).
The Bureau also uses the data to revise and expand
its work in deriving standard quantity budgets for
selected types of families (chapter 9). As time
and resources permit, the Bureau proposes to draw
upon this fund of consumer information to develop
a broad program of family living conditions
The U.S. Department of Commerce relies on
these family expenditure studies as the sole source
of information for revising its benchmark esti­
mates for a number of components in the household
sector of the national accounts. The Internal
Revenue Service used the published 1960-61 data
to revise its tables of State sales tax payments, for
guidelines to taxpayers in filing their Federal in­
21 The Bureau has initiated a series of special analytical re­
ports (B L S Report 238-1, et. seq.) based on the 1960-61 CES.
A list of reports in this series is available upon request.
The Bureau also prepares methodological monographs for the
expenditure surveys which compare national aggregates of in­
come, expenditures, and savings derived from the CES with
aggregates developed for the national income and product ac­
counts by the U.S. Department of Commerce and with data from
other sources.


come tax returns. Currently the Department of
Health, Education, and Welfare is studying the
data in connection with numerous welfare
A number of universities utilize the Bureau’s
basic data in a variety of research projects. The
potentialities of the survey results in consumer
market analysis are evident from reports issued
by the National Industrial Conference Board, the
Chamber of Commerce of the United States, and a
variety of trade publications and business
Data obtained from a sample survey as complex
as the survey of Consumer Expenditures are sub­
ject to many types of errors. These include
sampling, recording, and processing errors, and
errors due to the refusal or inability of some fami­
lies to give the information requested.
All data were reviewed, edited, and screened to
minimize processing errors. Chance variations
due to sampling can be measured statistically, and
the B L S has published preliminary rough esti­
mates of sampling error in the urban sample, in its
report for the urban United States. Preparation
of more detailed estimates for both the urban and
rural segments will depend upon the availability
of resources. Each report contains a section alert­

ing users to the cautions that must be exercised in
using averages based on small samples and pro­
vides either the actual number or the basis for
determining the number of families on which the
averages in each table are based. Approximately
78 percent of the national sample of urban and
rural families furnished usable schedules and
some of the nonrespondents supplied limited infor­
mation on family characteristics.22 Among the
participating families, inaccurate reporting is a
source of error, despite continued research in
schedule design and intensive training of the inter­
viewers. Such inaccuracies result from memory
errors, misunderstanding of a question or reluc­
tance to answer it, and incorrect entries by the
interviewer. Although the B L S and USD A have
accumulated substantial knowledge about such
reporting errors and will continue research in
this field, these errors cannot be quantified
22 This information is used in the monographs on methodology
(see footnote 21) to evaluate the nature of the sample losses
due to nonresponse. It has been the Bureau’s practice not to
attempt to introduce in the basic CES tables adjustments for
nonresponse or for the underreporting which is common in vir­
tually all compilations of income statistics. Such adjustments
present many difficulties when data are cross-classified as ex­
tensively as are the CES tabulations.

Technical References

1. Brady, Dorothy S. and Williams, Faith M., “ Advances in the Techniques of Measuring and Esti­
mating Consumer Expenditures,” J o u r n a l o f F a r m Econ om ics , May 1945, pp. 315-344.
A review of changes in the direction of family expenditure surveys as the role of the consumer
gained importance in economic theory in the mid-1930's. Discusses problems of concepts,
definitions, sampling, and data collection that emerged as emphasis shifted from the analysis
of expenditures of selected occupational groups to the interrelationships of expenditures, income,
and savings of all types of families throughout the United States.
2. Ferber, Robert, “ Research on Household Behavior,” The A m erican Econom ic Review, March 1962,
pp. 19-63.
A survey of the main empirical research since World War II on the determinants of house­
hold spending and saving behavior. Includes an extensive bibliography.
3. International Labour Office, F a m ily L iv in g Stu d ies— A Sy m p o siu m , Studies and Reports, New
Series No. 63, Geneva, 1961.
A collecton of recent case studies selected to draw attention to different types of family
living studies and to their methods and problems. The 15 chapters include surveys from coun­
tries throughout the world.
4. Lamale, Helen H., “ Changes in Concepts of Income Adequacy over the Last Century,” The
A m erican Econom ic Review , May 1958, pp. 291-299.
Relates changes in criteria for income adequacy to economic and social developments in
three broad periods: 1860 to 1900— the “Subsistence” or “ Break-Even” concept; 1900 to 1935—
the “ Living-Wage” concept; since 1935— the “Social” concept.




Technical References—Continued

5. Lamale, Helen H.

Stud y o f Consum er E xp en ditures, Incom es an d S av in g s— Methodology o f the Survey
o f Consum er Expen ditures in 195 0 . Philadelphia, Pa., University of Pennsylvania, Wharton School







of Finance and Commerce, 1959.
A comprehensive statement of the purposes, procedures, and reliability of results of the 1950
survey, with summaries of earlier surveys and research which influenced the 1950 methodology.
Includes facsimiles of questionnaires used in BLS expenditure surveys for 1901, 1917-19, and
Stigler, George J., “ The Early History of Empirical Studies of Consumer Behavior,” The Jo u r n a l
o f P o litic al Econom y , April 1954, pp. 95-113.
Examines the relationship of the empirical work on consumer behavior to the theoretical
work on income and demand prior to the twentieth century.
United Nations Statistical Office, H andbook o f Household S u rvey s: A P r ac tic al G uide fo r In q u irie s
on Levels of L ivin g, Provisional Edition. (Studies in Methods, Series F. No. 10.) New York,
United Nations, 1964.
Presentation of standard concepts, definitions, classifications, and tabulations, for household
surveys with the aim that their adoption or adaptation would enhance the value of the surveys
for national purposes and facilitate international comparison of results.
U.S. Bureau of the Census, H istorical S tatistics o f the United States— C olonial Tim es to 1957 , (1960)
Chapter G. The same, C ontinuation to 1962 an d R evision s , 1965.
A compilation of tabular data from major Government and private statistical reports on
consumer income and expenditures in the United States. Text describing each report is keyed
to the tabulations.
U.S. Department of Labor, How A m erican B u y in g H ab its Change (1959).
A popular-style book dovetailing various studies to yield a picture of changes in the con­
sumption habits of the American people between 1875 and 1950. Includes a chapter of technical
comments on consumption statistics and an extensive bibliography.
Williams, Faith M. and Zimmerman, Carle C., Stu d ie s of F a m ily L ivin g in the United States an d
Other C oun tries: A n A n a ly sis o f M ate rial an d Method , U.S. Department of Agriculture, Mis­
cellaneous Publication No. 223, December 1935.
Analytical annotated bibliography of approximately 1,500 studies of family living made in
52 countries. Includes statements on the history and methodology of consumption studies since
their inception.
A continuation of this bibliography, covering 1946 and later years, is published in Miss
Williams chapter, “ International Comparisons of Patterns of Family Consumption,” in Consum er
Behavior— Research on Consum er Reactions (Lincoln H. Clark, ed.) New York, Harper and
Brothers, 1958.
Zimmerman, Carle C., Consum ption an d S ta n d ard s of L ivin g, New York, D. Van Nostrand Co.,
Inc., 1936.
A definitive analysis of family budget studies, with pertinent tabular materials, and dis­
cussion of psychological, social, and economic concepts and theories of consumption.


ath ryn

K. M

urph y

Office of Prices and Living Conditions

Chapter 9. Standard Budgets
“Standards of living” refer to the goals of con­
sumers and workers with respect to their consump­
tion of goods and services, use of leisure time, and
conditions of work. Standard budgets measure
the total costs or amounts of income required to
achieve the levels and manner of living implicit
in these goals.1 The cost estimates are developed
by translating the generalized concept of a living
standard into a list of commodities and services
which can be priced. In theory, the list is based
wholly on objective judgments that the items and
quantities specified are essential in maintaining
the living standard. Thus, standard budgets are
normative, or benchmark, estimates of living costs.
They do not represent the ways in which family
incomes should be spent, nor do they show how
average families actually spend their incomes.
The first standard budgets prepared by the Bu­
reau of Labor Statistics were developed for the
specific purpose of evaluating the living conditions
of cotton-mill workers in Fall River, Mass, and
the South in 1908-09. These budgets described
two standards of living—a minimum, including
only bare necessities; and a fair standard, includ­
ing some allowance for comfort. Another budget
was developed in 1919, defining a standard of
health and decency. In the late 1930’s, B LS coop­
erated with the Works Progress Administration
in pricing two budgets: A maintenance budget,
described as above the minimum subsistence level
but not approaching a satisfactory American
standard of living; and an emergency budget,
derived by cutting the maintenance budget for
emergency conditions “with the least harm to the
individuals and the social group.”
The information available at the time concern­
ing the requirements for nutritionally adequate
diets and healthful housing was incorporated into
the definitions of the living standard in each of
these budgets. For other components of family
living, however, the “requirements” were formu­
lated primarily on the basis of the personal judg­
ment of the budget makers.2 It was not until 1946,
when the Bureau compiled the City Worker’s
Family Budget, that an objective method was

developed for the derivation of all segments of a
standard budget.3 The same method, with some
refinements, was used in 1959, in the interim re­
visions of the City Worker’s Family Budget and
the Budget for a Retired Couple,4 and it is this
methodology which is described in the remainder
of this chapter.

Description of the Budget
All normative estimates of living costs must be
based on specific family situations. The construc­
tion of a standard budget, therefore, requires a set
of assumptions, i.e., specifications, which must be
stated explicitly at the outset. These relate to the
age, size, and type of family; the manner of living
appropriate for the specified family composition
and the locality in which the family resides; and
the relative position of the living standard on the
consumption scale established by society, which
ranks various consumption patterns in an ascend­
ing order from mere subsistence to plenitude in
every respect.
Family composition has a significant effect on
spending patterns, manner of living, and family
needs. For the City Worker’s Family Budget, it
was specified that the family consists of an em­
ployed husband, age 38, with a wife not employed
outside the home, and two children, a girl age 8
and a boy 13. This family type represents a
middle stage in the life cycle, and it has been
widely used as the unit for other budgets compiled
for earlier periods. The family in the Budget for
1 In 1954, at the request of the United Nations Economic and
Social Council, a committee of experts from six different countries
(including the United States) recommended that the following
distinction be maintained between the terms “level” and “ stand­
ard” of living: The level of living relates to the actual living
conditions of a people. The standard of living relates to the
aspirations or expectations of a people, that is, the living condi­
tions which they seek to attain or regain, or which they regard
as fitting and proper for themselves to enjoy. The Bureau’s use
of these terms conforms with this recommendation. However,
laymen and the general public frequently refer to “ the high
levels of living” actually achieved by the average American
worker as the “high standard” of American living.
2 Published sources for these budgets are described in Technical
Reference No. 2 (listed on p. 67).
3 See Technical Reference No. 10. For description of a com­
panion budget for an elderly couple, see Technical Reference
No. 11.
4 See Technical Reference Nos. 5 and 7.




a Retired Couple consists of a husband and wife,
age 65 or over, who are assumed to be self-support­
ing, in reasonably good health, and able to take
care of themselves. This family unit, which has a
markedly different pattern of living and needs
from the younger family, has been the subject of
special concern in national policy formation over
the last two decades. Budget quantities and
budget component cost estimates for other family
types cannot be derived as fractions or multiples
of the quantities or cost estimates for food, shelter,
clothing, transportation, etc. for the four-person
family or the retired couple.5
Both types of families were assumed to live in
a large city or its suburbs. Assumptions were
made concerning, also, the living arrangements
and tenure6 of the families; inventories of housefurnishings, household equipment, and clothing;
means of transportation; ownership of life insur­
ance, provisions for medical care, savings posi­
tions, etc. Generally these assumptions are based
on the prevalence of ownership of particular types
of assets in the urban metropolitan population,
and the availability of goods and services provided
by governments for collective consumption or pro­
vided under collective bargaining agreements
between employers and unions.
The standard of living represented in both
budgets lies somewhere between a minimum and a
liberal standard. It is described as “modest but
adequate” according to prevailing standards of
what is needed for health, efficiency, the nurture
of children, and for participation in social and
community activities. The social concept of what
constitutes an adequate living standard, or goal,
changes over time as modes and manner of living
change. Furthermore, in a complex and dynamic
society, at any given point in time, there is no
single living standard which can be considered as
an appropriate norm for all purposes.7 For at
least the last two decades, the “modest but ade­
quate” standard reflected in the budgets for a fourperson family and a retired couple has been
interpreted to be more appropriate than a mini­
mum standard as a measure of income adequacy
for self-supporting families. This standard has
been considered too high, however, as a norm for
dependent families receiving public assistance, and
too low for some of the other purposes for which
measures of income adequacy are needed today,

Data Sources
Budget quantities and pricing specifications
were derived from two sources: (1) Scientific or
technical judgments concerning the requirements
for physical health and social well-being; and (2)
analytical studies of the data reported in the
Bureau’s surveys of consumer expenditures,8 to
determine by objective procedures the choices of
goods and services made by consumers in successive
income groups.
Scientific standards for nutritionally adequate
diets for individuals in different sex-age groups
have been developed by the Food and Nutrition
Board of the National Research Council, and
translated by the U.S. Department of Agriculture
into food plans at different cost levels. These food
plans were used as the basis for the food-at-home
component of the budgets.
Housing standards established by the American
Public Health Association and the U.S. Public
Housing Administration were adopted for the
budgets. These standards relate to sleeping space
requirements, essential household equipment,
plumbing facilities, adequate supplies of utilities
and heat, and the structural condition and neigh­
borhood location of the dwelling units. Fuel
requirements were derived by analyzing actual fuel
purchases of families in the specified types of
dwellings in relation to degree-days to provide an
adjustment for differences in climate. Estimates
6 Extensive analyses of consumption data dating back over
more than a century have provided a variety of measuresi of
general welfare, e.g., the relative adequacy of diets, the propor­
tion of income spent for various categories of goods, or the pro­
portion of income saved. These measures, either singly or in
combination, have been used as the basis for determining scales
of equivalent income for families of different size. One such scale,
published by B L S in November 1960, is described in a Technical
Note: Estim ating Equivalent Incomes or Budget Costs hy Family
Type (see Technical Reference No. 8). The scale is based on the
assumption that families spending the same proportion of income
on food have attained equal levels of living. While the scale is
useful in estimating equivalent costs of goods and services, or
net income requirements after income taxes and occupational
expenses, it cannot be applied to individual items or major
components of budget costs.
9 Although there has been a substantial increase in homeownership among city families at all income levels since prewar
years, resources available for the interim revision of the budgets
did not permit the development of satisfactory concepts, pro­
cedures, and data for defining comparable dollar estimates of
budget costs for homeowners. Therefore, the interim budgets
include estimates for rental housing costs only.
7 For additional discussion of the relativity of living standards,
see Technical Reference Nos. 3 and 4.
8 For a description of the Bureau’s surveys of consumer ex­
penditures, see chapter 8.


of electricity and other utility services required
for the appliances specified for the budgets were
furnished by utility companies and associations.
The widespread use of insurance to cover the
cost of major illness was accepted as a quasi­
standard for medical care, and a family member­
ship in a group hospitalization plan was specified
for this component of the budgets. Quantities of
medical care services not covered by insurance
were derived from data on utilization rates pro­
vided by the U.S. National Health Survey con­
ducted in 1957-58.
For the other components of the budgets (cloth­
ing, housefurnishings, transportation, personal
care, household operation, reading, recreation,
tobacco, education, gifts and contributions, and
miscellaneous expenses), there are no generally
accepted scientific standards. Therefore, a tech­
nique was developed which relied on the collective
judgments of consumers as to what is adequate.
Purchases were examined at successive income
levels to determine the income level at which the
rate of increase in quantities purchased began to
decline in relation to the rate of change in income,
i.e., the point of maximum elasticity. The average
numbers and kinds of items purchased at this in­
come level became the quantities and qualities
specified for the budgets.
Prices for items included in the budgets were
collected in the types of stores and professional
and service establishments customarily patronized
by city workers’ families. The prices, pricing
procedures, reporting stores, and service establish­
ments, and price calculation methods were those
used by the Bureau for the Consumer Price In­
dex,9 except that more price quotations were ob­
tained in some cases to permit calculation of average prices and different qualities were priced m
other cases to represent the budget level.10

Analysis and Presentation
In the methodology described previously, a
standard budget is the end result of complex sta­
tistical analyses and calculations; it is not simply
the product of a survey of how families at partic­
ular income levels actually spend their money.
The judgment of the budget maker is involved in
selecting among the family types and manners


and standards of living to be represented; in de­
termining the most appropriate sources of data
to be used in deriving budget quantities; and in
interpreting actual family consumption in terms
of living standards. The appropriateness of the
operating assumptions can be evaluated only with
reference to the purposes to be served by and the
uses made of the standard budgets.11
Standard budget cost estimates may be analyzed
in four general ways: (1) Budget costs are an­
alyzed in relation to income. However, a stand­
ard budget is a completely appropriate measure
of income adequacy only for the specific family
type for which it was constructed. This type of
analysis has been restricted, therefore, by the
development of budget cost estimates for only two
family types, coupled with the paucity of data on
income by family composition. (2) Budget costs
in one place are compared with costs in another,
i.e., the budget costs provide a basis for calculat­
ing an index of intercity differences in living costs.
The Bureau has provided this type of analysis
in conjunction with its published reports on the
budgets, and has recommended the method as the
only valid approach to the development of a gen­
eral purpose intercity index for cities with reason­
ably comparable patterns of consumption. (3)
Budget costs are compared over time to provide a
measure of change in living standards. The
sporadic character of the Bureau’s standard budget
research program imposes serious limitations on
this type of analysis, but potentially the program
provides a basis for evaluating changes in the levels
and standards of living of families from decade
to decade. (4) Finally, budget costs estimates
for different standards of living are compared
with each other to provide a measure of the aggre­
gate addition to income required to raise con­
sumption to particular levels. This type of
analysis remains theoretical, since cost estimates
for more than one living standard have not been
developed in this country since the early 1940’s.
9 For a description of the Consumer Price Index, see chapter

10 .

10 Technical Reference Nos. 5, 7, 10, and 11 provide additional
details on the methodology, as well as lists of goods and services
priced and dollar cost estimates, of the B L S budgets.
11 For a discussion of the uses of standard budgets, see Techni­
cal Reference No. 9. See also Technical Reference No. 6 for
comments on the needs for and limitations of existing budgets
for the elderly.



Uses and Limitations
Standard budgets are used as tools in economic
research to appraise the economic condition of the
population and to evaluate the need for, and the
effect of, specific laws and programs. Normative
estimates of living costs are needed, for example,
to measure the extent to which social security bene­
fits or payments under unemployment insurance
programs meet family needs; to estimate aggregate
requirements for consumer goods as a basis for
developing public policies; or to prepare sound
and valid estimates of the number of families liv­
ing in poverty. Standard budgets also provide
benchmarks for administrative determinations of
need or of the cost of an adequate living standard,
as required by a number of existing laws or poli­
cies of voluntary social, welfare, and educational
agencies, e.g., in the determination of eligibility for
public assistance; public housing; subsidized medi­
cal, mental health, or guidance services; or college
scholarship aid.
In addition to their primary use as tools in
evaluating income adequacy, standard budgets are
needed to measure place-to-place differences in liv­
ing costs, as a basis for family counseling, in wage
negotiations, and as an aid in consumer education.
No single estimate of living costs such as that
provided by the City Worker’s Family Budget or
the Budget for a Retired Couple, which have been
priced in only 20 large cities, can be used validly
for all of the purposes for which living costs esti­
mates are needed. For example, budget costs for
only two family types are not an appropriate basis
for making global estimates of need (for the urban
population in the United States, Northeast Region,
New York City, etc.), which should be based on

cost estimates or on some measure of equivalent
income requirements, for families of all sizes and
types. Furthermore, prices collected in 20 large
cities for a specific date are not representative of
prices in other large or medium-size cities; and
neither budget quantities or prices for large cities
are necessarily appropriate to the manner of living
in small cities. Similarly, the Bureau’s two stand­
ard budgets represent a specified standard of liv­
ing, which may be higher or lower than the stand­
ard appropriate for evaluating income adequacy
under the requirements of existing or proposed
policies, laws, or programs, or within the frame­
work of a specific research project. This deter­
mination can be made only by the policy maker,
administrator, or research worker—not by the
budget maker.
Intercity living cost indexes based on the B L S
budgets are designed to reflect differences in costs
of established residents in a community. Differ­
ences in housing costs, for example, are based on
the average costs of occupied rental dwellings and
are not a valid measure of the costs of vacant
rental units available to new residents or the costs
of maintaining an owned home. Differences in
the cost of food reflect not only differences in price
levels, but more importantly differences in re­
gional preference patterns in the choice of food to
meet the nutritional standard.12 The indexes,
therefore, are more appropriate for use as a re­
search tool in analyses of the relationship between
income and costs of established residents in dif­
ferent locations than as a measure of differences
in costs for families moving from one location to
12 For a measure of the effect on food costs of price-level dif­
ferences versus regional differences in the choice of foods, see
Technical Reference No. 1.


Technical References

1. Brackett, Jean C., “ Intercity Differences in Family Food Budget Costs,” M onthly L ab or Review ,
October 1963, pp. 1189-1194.
An analysis of the effects on food budget cost estimates of using for all cities a single set of
weights representing urban U.S. food patterns, or different weights for each city reflecting the
food preferences of the region in which the city is located. Also presents a discussion of the
conceptual implications of varying the weights in a place-to-place comparison of family living
2. Clorety, Joseph A., “ Consumption Statistics: A Technical Comment,” How A m erican B u y in g
H ab its C hange , chapter X, 1959, pp. 217-242.
Includes a section on “ Standard Budgets as Indicators of Progress” (pp. 232-242). Also
presents in summary form a representative cross-section of budgets compiled in this country
during the 20th century, showing average dollar cost figures for the total and for the major
components of each budget.
3. Lamale, Helen H., “ Changes in Concepts of Income Adequacy Over the Last Century,” J o u r n a l of
the A m erican Econom ic A sso c ia tio n May 1958, pp. 291-299.
An analysis of the relationship over time between actual levels of living in the United States
and the goals or standards of living which have been accepted in different historical periods and
for different purposes; and a discussion of the implications in this relationship for present-day
concepts of income adequacy.
4. Lamale, Helen H., “ Poverty: The Word and the Reality,” M onthly L ab o r Review} July 1965, pp.
Discusses the role of standard budgets in providing an intelligible definition of poverty,
for use in evaluating income adequacy for different family types and in different geographical
locations and for estimating the extent of poverty in the United States.
5. Lamale, Helen H. and Stotz, Margaret S., “ The Interim City Worker’s Family Budget,” M onthly
L ab o r Review , August 1960, pp. 785-808.
Estimates of the cost of a “ modest but adequate” standard of living for a husband, wife, and
two children (living in rented housing), at autumn 1959 prices, in 20 large cities and their suburbs
(Atlanta, Baltimore, Boston, Chicago, Cincinnati, Cleveland, Detroit, Houston, Kansas City,
Los Angeles, Minneapolis, New York, Philadelphia, Pittsburgh, Portland, Oreg., St. Louis, San
Francisco, Scranton, Seattle, and Washington, D.C.) Includes a detailed list of the goods and
services considered necessary by four-person families to maintain the specified living standard as
determined by levels of living actually achieved in the 1950’s, and describes how this representa­
tive list was developed and priced. (See Reference No. 10 for description of original BLS City
Worker’s Family Budget.)
6. Orshansky, Mollie, “ Budget for an Elderly Couple: Interim Revision by the Bureau of Labor
Statistics,” S o c ia l Security B u lle tin December 1960, pp. 26-36.
A summary report on “ The BLS Interim Budget for a Retired Couple” . (See Reference
No. 7.) Includes a discussion of various conceptual problems encountered in developing normative
living costs estimates for a retired couple, and some of the limitations of this particular budget
for the multitude of purposes for which budgets for older persons and families are needed.
7. Stotz, Margaret S., “ The BLS Interim Budget for a Retired Couple,” M onthly L ab or Review ,
November 1960, pp. 1141-1157.
Estimates of the cost of a “ modest but adequate” standard of living for a man age 65 or over
and his wife (living in rented housing), at autumn 1959 prices, in 20 large cities and their suburbs
(cities are the same as those listed in Reference No. 5). Includes a detailed list of the goods and
services considered necessary for retired couples to maintain the specified living standard as
determined by levels of living actually achieved in the 1950’s; and describes how this representa­
tive list was developed and priced. (See Reference No. 11 for description of original Budget
for an Elderly Couple.)
8. U.S. Department of Labor, Bureau of Labor Statistics, “ Estimating Equivalent Incomes or Budget
Costs by Family Type,” M onthly L ab or Review , November 1960, pp. 1197-1200.
Describes a scale for measuring the relative after-tax income required by families of differing
composition to maintain the same level of material well-being, or for estimating comparable
costs of goods and services for families of different age, size, and type. (Scale values cannot be
used to estimate relative costs of components of family budgets— food, housing, taxes, insurance,

222 617 66——6




Technical References—Continued

9. U.S. Department of Labor, Bureau of Labor Statistics, “ Estimating Equivalent Incomes or Budget
Costs by Family Type.” Report of the A dvisory Committee on S ta n d ard Budget Research , June
1963, 26 pp.
Recommendations formulated by a committee of experts with extensive experience in using
standard budgets on the needs for various types of budgets, general concepts of the standards of
living to be described by the budgets, and methodological and other problems associated with
estimating and publishing budget costs. Includes a selected bibliography on the major uses
of standard budgets.
10. ----- . “ W orkers’ B udgets in the U nited S tate s: C ity F a m ilie s an d S in g le P e rso n s, 1946 an d 19 4 7 ,”
(BLS Bulletin 927, 1948) 55 pp.
Concepts, definitions, and techniques used in developing the original City Workers Family
Budget for a four-person family, detailed list of goods and services priced, and 1946-47 cost
estimates for 34 cities. Also an historical survey of family budgets, and summary data on State
budgets for single women workers.
11. Federal Security Agency, Social Security Administration, “ A Budget for an Elderly Couple,”
S o c ia l Security B u lletin , February 1948, pp. 4-12.
Estimates of the cost of a “ modest but adequate” standard of living for a couple age 65 or
older, at March 1946 and June 1947 prices, in eight large cities. (Concepts and techniques used
to compile this budget were parallel to those employed in developing the original BLS City
Workers Budget. See Reference No. 10.)

—J ea n C. B rackett

Office of Prices and Living Conditions

Chapter 10. Consumer Prices
The Consumer Price Index was initiated during
World War I when rapid changes in living costs,
particularly in shipbuilding centers, made such an
index essential in wage negotiations. To provide
appropriate weighting patterns for the index,
studies of family expenditures were conducted in
92 industrial centers in 1917-19. The Bureau of
Labor Statistics began publication of indexes for
32 individual cities in 1919. Regular publication
of U.S. city average indexes was not begun until
1921, but indexes were estimated back to 1913.1
Since that time the weighting factors, the list of
items included in the market basket, and the cities
in which price data were collected for calculating
the index have been updated several times.
Because people’s buying habits had changed
substantially by the mid-1930’s, a new study was
made covering expenditures in the years 1934-36
which provided the basis for a comprehensively
revised index introduced in 1940 with retroactive
calculations back to 1935.
During World War II, when many commodities
were scarce and goods were rationed, the index
weights were adjusted temporarily to reflect these
shortages. Again in 1950, the Bureau made in­
terim adjustments, based on surveys of consumer
expenditures in seven cities between 1947 and 1949,
to reflect the most important effects of immediate
postwar changes in buying patterns.2 This ad­
justment was followed by the first comprehensive
postwar revision of the index, which was com­
pleted in January 1953.3 At that time, not only
were the weighting factors, list of items, and
sources of price data updated, but many improve­
ments in pricing and calculation methods were in­
troduced. Also, coverage of the index was
extended to small cities so as to represent all urban
wage-earner and clerical-worker families. The
most recent revision was completed in 1964, with

the introduction of new expenditure weights based
on spending patterns in 1960-61, and updated
samples of cities, goods and services, and retail
stores and service establishments.
The manner in which the index has been used
and its acceptance by the public have changed
from time to time. It has seen many appraisals,
criticisms, and investigations. Perhaps the most
far-reaching study was conducted during World
War II by the President’s Committee on the Cost
of Living.4 The House Committee on Education
and Labor conducted a detailed examination of the
index in 1951.5 The most recent study was made
by the Price Statistics Review Committee, ap­
pointed by the National Bureau of Economic
Research, at the request of the Office of Statistical
Standards of the Bureau of the Budget, to review
all government price statistics.6
As a result of these investigations and the
Bureau’s continuing efforts to improve the index,
changes in coverage, collection, and calculation
procedures have been introduced at various times.
Examples of these changes include the addition
of medium and small cities to the city sample in
1953, the extension of coverage to include single
workers in 1964, and institution of direct pricing
of restaurant meals in 1953.
1 Collection of food prices back to 1890 had been initiated in
1903. During the course of the 1917-19 expenditure survey,
retail prices for other articles were collected in 19 cities for
December of each year back to 1914 and in 13 other cities back
to December 1917 only. Retail prices of food and wholesale
prices of other items were used to estimate price change from
1914 back to 1913.
2 See Interim, Adjustment of Consumers’ Price Index (B L S
Bulletin 1039, 1951).
3 See Consumer Prices in the United States 195S-58 (BDS Bul­
letin 1256).
4 Report of the President’s Committee on the Cost of Living,
Office of Economic Stabilization, Washington, 1945.
5 Consumers’ Price Index— Report of a Special Subcommittee
of the Committee on Education and Labor. House of Representa­
tives, 82/1, Subcommittee Report No. 2, Washington, 1951.
6 Government Price Statistics— Hearings before the Subcom­
mittee on Economic Statistics, Join t Economic Committee, Con­
gress of the United States, 87/1, P art 1, Washington, Jan u ­
ary 24, 1961.




Description of the Index
Concept and Scope.1 The Consumer Price Index

(CPI) is a statistical measure of changes in prices
of goods and services bought by urban wage
earners and clerical workers,8 including families
and single persons. The index often is called the
“cost-of-living index,” but its official name is Con­
sumer Price Index for Urban Wage Earners and
Clerical Workers. It measures changes in prices,
which are the most important cause of changes in
the cost of living, but it does not indicate how
much families actually spend to defray their liv­
ing expenses. Prior to January 1964, the com­
plete name for the index was: Index of Change in
Prices of Goods and Services Purchased by City
Wage-Earner and Clerical-Worker Families to
Maintain Their Level of Living.
The Consumer Price Index is a weighted aggre­
gative index number with “fixed” or “constant”
annual weights, or it often is referred to as a “mar­
ket basket” index. Thus, in the Consumer Price
Index the procedure is to measure price change
by repricing at regular time intervals and com­
paring aggregate costs of the goods and services
bought by consumers in a selected base period.
The quantities of these goods and services are kept
constant except at times of weight revisions. Since
new weights are introduced without affecting the
index level, any change in aggregate costs is due to
price change. The quantities represent not only
annual consumption of the goods and services
actually priced for the index but also consumption
of related items for which prices are not obtained,
so that the total cost of the market basket repre­
sents total consumer spending for goods and
The index represents price change for every­
thing people buy for living—food, clothing, auto­
mobiles, homes, housefumishings, household sup­
plies, fuel, drugs, and recreational goods; fees to
doctors, lawyers, beauty shops; rent, repair costs,
transportation fares, public utility rates, etc., in­
cluding all taxes directly associated with the pur­
chase of an item and its continued ownership. It
deals with prices actually charged to consumers,
including sales and excise taxes, since these are
an inherent part of the market price the consumer
must pay for goods and services subject to such
taxes. It also includes real estate taxes on owned
homes which are part of the price of homeowner

ship. However, it does not include income or other
personal taxes, since they are not associated with
prices of specific goods and services although they
may have an indirect impact.9 Since 1953, it has
treated the purchase of a home in the same way
as the purchase of such durable goods as automo­
biles, refrigerators, etc.
In the 1964 revision, the index coverage was
extended to include single consumer units10 in
addition to families of two or more.11 The average
size of families represented in the index is about
3.7 persons, and their average family income in
1960-61 was about $6,230 after taxes. The average
income after taxes of single persons represented
in the index was about $3,560.
Weighting Structure . The annual consumption
patterns represented in the index since January
1964 were determined in the Survey of Consumer
Expenditures (C E S )12 in 66 Standard Metropol­
itan Statistical Areas (SM SA’s) and smaller cities
7 See article by Sidney A. Jaffe, ‘‘The Statistical Structure of
the Revised CPI” , Monthly Labor Review, August 1964, pp. 916924.
8 The definition of wage earners and clerical workers is based on
the occupational classification used by the Bureau of the Census
for the 1960 Census of Population and listed in the Alphabetical
Index of Occupations and Industries. The group includes crafts­
men, foremen, and kindred workers, such as carpenters, book­
binders, e tc .; operatives and kindred workers, such as apprentices
in the building trades, deliverymen, furnacemen, smelters, and
pourers, e tc .; clerical and kindred w orkers; service workers, ex­
cept private household, such as waitresses, practical nurses, e tc .;
sales w orkers; and laborers, except farm and mine. It excludes
professional, technical, and kindred workers, such as engineers
and teachers; farm ers and farm m anagers; managers, officials
and proprietors, except fa r m ; private household w orkers; and
farm laborers and foremen. A consumer unit included in the
1960-61 Survey of Consumer Expenditures was classified in the
index group if more than half the combined income of all
family members w as obtained in a wage-earner and clericalworker occupation and a t least one family member was a full­
time earner (i.e., worked 37 weeks or more during the survey
9 For a more detailed discussion, see “ Taxes and the Con­
sumers’ Price Index,” Monthly Labor Review, Jan uary 1953, pp.
53-57. (Reprint 2090.)
10 A person living alone or in a household with others from
whom he was financially independent; i.e., his income and ex­
penditures were not pooled. Terms such as “ single persons,
single workers, singles, etc.” used subsequently, refer to single
consumer units and not to unmarried persons.
11 From Jan uary through November 1964, the Bureau computed
a separate index applying only to families of two or more, for
comparability with the previous series. Since the same items
were priced for families and singles, the overall movement of
the separate index was identical with the index including single
workers and it was discontinued.
32 The Survey of Consumer Expenditures isi discussed in
Chapter 8 of this bulletin. The selection of the expenditure sur­
vey and CPI city samples is described in detail in an article by
Marvin Wilkerson, “ The Revised City Sample for the Consumer
Price Index,” Monthly Labor Review, October 1960, pp. 10781083. (Reprint 2352)

T a b l e 1.

C i t i e s , P o p u l a t io n W e ig h t s ,


P r ic in g S c h e d u l e

fo r t h e

R e v is e d C o n s u m e r P r i c e I n d e x
Pricing schedule a

City and size stratum

Other items






A. Standard metropolitan statistical areas of 1,400,000 or more in 1960:
Baltimore, Md................................................................................................
Boston, Mass................. ................................................................ .............
Chicago-Northwestern Indiana4.*.......... ......... ............. ..............................
Cleveland, Ohio............................................................................................


1A, 2B
1A, 2B
1A, IB, 2A, 2B
1A, 2B


Detroit, Mich.................. .................................. .................................... ......
Los Angeles-Long Beach, Calif.................... ...............................................
New York-Northeastern New Jersey4...............-......................................
Philadelphia, Pa................................................ ...........................................


1A, 2B
1A, 2B
1A, 2B
1A, 2B


Pittsburgh, Pa.............................................................................................
St. Louis, Mo................................................................................ .............
San Francisco-Oakland, Calif........................................................ ............
Washington, D.C..........................................................................................


1A, 2B
1A, 2B
1A, IB, 2A, 2B
1A, IB, 2A, 2B

B. Standard metropolitan statistical areas of 250,000 to 1,399,999 in 1960:«
Atlanta, Ga..................................................................................................
Buffalo, N .Y ..................................................................................................
Cincinnati, Ohio-Ky... ................................................................................
Dallas, Tex.....................................................................................................
Dayton, Ohio.................................................................. ....................... ......


1A, 2B
1A, 2B
1A, 2B
1A, 2B

Denver, Colo__.......................... ...................... ...... ...................... ...............
Hartford, Conn..............................................................................................
Honolulu, Hawaii....... —........................................................ ......................
Houston, Tex............................................. .................................................
Indianapolis, Ind...................... ...................................................................


1A, 2B
1A, 2B

Kansas City, Mo.-Kans__................................................................. ...........
Milwaukee, Wis__________ __________________________________
Minneapolis-St. Paul, Minn___________ _____ __________________


1A, 2B
1A, 2B


Nashville, Tenn....................... ................................ .......... .......................
San Diego, Calif..................... .................. ..................................................
Seattle, Wash............ .....................................................................................
Wichita, Kans............................................ ............. ....................................


1A, 2B
1A, 2B



C. Standard metropolitan statistical areas of 50,000 to 249,999 in 1960:
Austin, Tex...................................................................................................
Bakersfield, Calif.................................... .....................................................
Baton Rouge, La___.............................................. ............... .......................
Cedar Rapids, Iowa.................................... .................................................
Champaign-Urbana, 111...............................................................................


Durham, N.C_._......................................... .................................... ............
Green Bay, Wis____ ________ _____ ___________________________
Lancaster, Pa.............................................................................. ................
Orlando, Fla........................ ......................................................................
Portland, Maine— ...................................................................... ................

A 9"R
' 11aA, 1T


A 0"R





1A, 2B

1A, 2B

1A, 2B
1A, 2B


D. Urban places of 2,500 to 49,999 in 1960:
Anchorage, Alaska____________________________________________
Crookston, Minn............................................ ............... ............ ............
Devils Lake, N.Dak..... ...................... .......................... ...... .......................
Findlay, Ohio...................................................................................... .........
Florence, Ala..................................................................................................





Kingston, N .Y ....... .......................................................................................
Klamath Falls, Oreg__________________________________________
Logansport, Ind............................ .................... ..........................................
Mangum, Okla............................................................................... ...............
Martinsville, Va................................................................ ...........................


McAllen, Tex__ _________________ _____ ______________________
Millville, N .J.................................................................................................
Niles, Mich....................................................................................................
Orem, Utah.. — ................................................................ .......................
Southbridge, Mass................................................................................... ......


Union, S.C.................... ................................... ...........................................
Vicksburg, Miss.................................................................................... .........


1The 18 largest Standard Metropolitan Statistical Areas as defined for
the 1960 Census of Population were selected on a certainty basis and repre­
sent themselves only in the population weight patterns. The other sample
selections carry not only their own population weights but also prorata
shares of the population weights of all cities in their region in the same
population class.
2 Item samples are identified as samples “ 1” and “ 2.” Outlet samples are
identified as samples “A” and “ B .” The determination as to the extent
of sampling within an area depended on plans for publishing separate area
indexes and on plans for developing estimates of sampling error and its



























1A, 2B
1A, 2B





1A, 2B
1A, 2B







* Foods, fuels, and several other items are priced every month in all cities;
Prices of a few items are collected semiannually or annually in all cities.
Prices of other goods and services are obtained on the schedule indicated:
M=Every month.
l=January, April, July, and October.
2=February, May, August, and November.
3=March, June, September, and December.
4 Standard Consolidated Areas.
* Population weights revised for this group beginning January 1966*



covering the period 1960-61, except for Anchor­
age, Alaska, which was surveyed for 1959. E x­
penditure records were obtained from the 4,343
urban families of two or more persons and from
the 517 single workers included in the survey who
were classified as wage earners or clerical workers.
Sixteen cities in the smallest size class which
were included in the 1960-61 survey are not in­
cluded in the CPI sample for pricing. However,
in the weight derivation, expenditures by con­
sumer units in these small cities were included
with those for the 16 small cities priced for the
index, so that the resulting weights are based on
the total sample of 32 small cities.
In establishing index weights from the detailed
expenditure data, about 400 items were selected
objectively to compose the “market basket” for
current pricing, beginning with the January 1964
“new series” indexes. Not all items are priced in
every SMSA or city. In order to make possible
estimates of sampling error, two subsamples of
items have been established. These are priced in
different areas and in different outlet samples, as
indicated in table 1. The population weights
shown in the table are used to combine price data
for the 56 individual areas in the CPI sample into
a U.S. city average. They were derived from the
1960 Census of Population but adjusted to repre­
sent the wage-earner clerical-worker coverage of
the CPI. For the 18 largest SMSA’s, which are
included in the sample with certainty, the weights
are based on their respective populations only. For
the remaining cities, which were selected by prob­
ability sampling methods to represent all other
urban places, the weights represent not only the
specific city population, but also the population
of all cities in the same region and size class. Thus,
every city in the same region and size class (other
than the 18 largest) has identical population
The list of items priced includes all the most
important goods and services and a sample of the
13 Six additional B size Standard Metropolitan Statistical Areas
were added to the national index in January 1966. Since they
were selected outside the probability framework, they were as­
signed only their own population weight which was substracted
from that of other stratum B cities in the same region.
14 Minor weight revisions are introduced by linking.
18 For a more complete discussion, see The Consumer Price
Index: Pricing and Calculation Procedures, unnumbered paper
by Doris P. Rothwell, Bureau of Labor Statistics, U.S. Depart­
ment of Labor, 1964.

less important ones. In combination, these repre­
sent all items purchased. Table 4 contains a com­
plete list of the items priced for the index. The
content of this market basket in terms of items,
quantities, and qualities is kept essentially un­
changed 14 in the index calculation between major
revisions so that any movement of the index from
one month to the next is due solely to changes in
prices. A comparison of the total cost of the mar­
ket basket from period to period yields the measure
of average price change.
In the selection of the item sample for the re­
vised CPI, except for the choice of the particular
quality or variety of the item to be priced (speci­
fication), probability sampling techniques were
used, as described later. The more important items
are included in the sample with certainty. The
remaining items within a given expenditure class
were selected with probability. The relative im­
portance of a certainty item represents consumer
spending for that item only. “Probability” items
represent all other items within the expenditure
class. The total weight of these items is divided
equally among the probability items within an ex­
penditure class. Table 2 shows relative im­
portances in the national index as of December
1963. Individual relative importances are not
shown in the table for probability items; rather
their combined importance is shown as “other
priced items” in each expenditure class.

Data Sources and Collection Methods
Prices are obtained in the 56 area CPI sample by
personal visit to a representative sample of nearly
18,000 stores and service establishments where
wage and clerical workers buy goods and services,
including chain stores, independent grocery stores,
department and specialty stores, restaurants, pro­
fessional people, and repair and service shops715
Rental rates are obtained from about 40,000 ten­
ants. Reporters are located both in the city
proper and in suburbs of each urban area. Cooperation is completely voluntary.
Prices are collected in each urban location at
intervals ranging from once every month to once
every 3 months, as indicated in table 1, with a few
items surveyed semiannually or annually. Be­
cause food prices change frequently, and because

T a b l e 2.


C o n su m er P r ic e I n d e x (N ew S e r ie s ) 1 R e la t iv e I mportance of M a jo r G r o u p s , S p e c ia l G r o u ps and
I n d iv id u a l I te m s S e l e c t e d w ith C e r t a in t y 2 D e c e m b e r 1963

Percent of
all items
December 1963

M ajor Groups

Medical care _

___ _________________________

Reading and recreation__________________________


S pecial Groups
All items less
All if.Aims less food _

_____ _______________

Durables _____________ _______________________
_______ ___________________________
Apparel commodities__________________ ____
^Apparel commodities less footwear__________
Nondurables less food and apparel_______ _____

Other services _

______ ____ ____________ ___

4. 56

I ndividual I tems
____ _ __ _ ________ _
Food at home _ ________ ___ _ _ ____ ______
Cereals and bakery products.. __
------- --Cereals_
______ ______________ ____
Bakery products_______________ __ ____
White bread_____ _______ _______ _
Other priced items_________ ____ _
Meats, poultry, and fish-__________ ______
Meats. ____ ___ „____ ____ ______
Beef and veal_____ _ _ _ __ ____ .
Hamburger_______________ _____
Steak . . . ___ ________ ____ _
Other priced items. _____ ______ _ _
Pork _________________ ____ . Pork chops_______ ___ _________
Bacon___ ________ ______________
Other priced items____ _______ . . .
Other meats______ _ __ __
__ ____
Poultry __ _ ____ . _ _ ________
Frying chickens___________________
Other priced items.. _____ _ __ ___
Fish_____ ___ _________________________
Dairy products__________ _ _______________
Mnk, fresh (grocery)__ _____________ ____
Milk, fresh (delivered).. . . . _____ _______
Butter_________________ ______ ____ __
Other priced items_______ __ __________
Fruits and vegetables_____ _
__ _____ _____
Fresh fruits______________________ ___
Apples________ ________ _ _ ______
Bananas______ ______________ ______
Oranges______ ______ _ _ ____ _ __ __
Other priced items__ _ _______ _ ____
Fresh vegetables______ ____ ___ __ ____
Lettuce________ ____ ____ ___________
Potatoes____________ ______ .. . ..
Tomatoes_______. . . ______ ____ _____
Other priced items___ __ ___________
Processed fruits and vegetables.___________ _
Other food at home. ___________ _ ________ ___
Eggs_________________ _____________ _ _
Fats and oils____________ ____ ________
Margarine____ ____ ___
_ _______
Other priced items____ _ _ _____ _ __
Sugar and sweets_______ _______________
Nonalcoholic beverages__ _____ ________
Coffee, can and bag______ _ _____ ____
Other priced items______ __________ ___
Prepared and partially prepared food________
Food away from home____________ _ _ _____ _ _
Restaurant meals__________ _______ _________
Between meal snacks________________________
See footnotes at end of table.

1. 05
2. 21


Percent of
all items
December 1963

Housing_____ __ ________________________ _______ _
Shelter____ ___________________ _______________
Rent___________________________ __________
Hotels and motels___________________________
Homeownership_________________ ______ ____
Purchase and financing____________________
Home purchase______________________
Mortgage interest_____________________
Taxes and insurance______________________
Real estate taxes______________________
Property insurance_____ _________ ____
Maintenance and repairs__________ _______
Commodities_______ _________________
Fuel and utilities_________________________ ___
Fuel oil and coal__________ _______________
Fuel oil________________________________
Coal____ ______________________ _______
Gas and electricity.______ __________ ________
Gas___ _______________________________
Electricity_________________________ ____
Other utilities______________ ______ ___ ___
Water and sewerage___ __________________
Household furnishings and operation_______________
Textile housefurnishings______________________
Furniture_______ _________________ ___ __
Bedroom suite___________________ ___ ___
Living room suite________________________
Other priced items____ ____ ______________
Floor coverings______________ _______ ______
Rugs, soft surface________ _ _____________
Other priced items____ ____ ______________
Appliances__ __ _ ___
_ __
Refrigerators _ _ ___ _________________
Other priced items______________ ______
Other housefurnishings_______ _ ____________
Housekeeping supplies_______________________
Housekeeping services____ __________________
Domestic service________________________
Baby sitter__ ________________________
Postage___ _ _ _____ ___ ___ _ _ ____
Other priced items_______________________


Apparel and upkeep____________________________ ___
Men’s and boys’ apparel_______________ __________
Men’s apparel . . . . . . .
Suits, year round
_ _
Other priced items
Boys’ apparel______________________________
Women’s and girls’ apparel ______________________
Women’s apparel___________________________
Winter coats __
Street dresses___________________________
Hose, nylon____________________________
Other priced items_______________________
Girls’ apparel_____________________ ________
Street shoes, men’s_________________ ________
Street shoes, women’s___ _ _ _____
Other priced items__________________________
Other apparel
Commodities. _
Drv cleaning
Men’s suit__________________________
Women’s dress
Other priced items _
Private transportation__________________________
Autos and related goods____ __________________
Auto purchase.. ________________________
New cars
Used cars_______ ________ _________
Gasoline and motor oil____________________
Gasoline______________________ _____
Motor oil.............. ......... ...............................
Auto parts
Automobile services.
.... _ _
Auto repairs and maintenance
Other automobile
Auto insurance
Registration fees
______ _
Drivers’ license
_ _ __
Parkins? fees
Auto financing charges 3
Public transportation
Local transit ______________________________


Train fares

Airplane fares______________________________
Intercity bus fares...................... ...............................




T a b l e 2.



m eth o ds

C o n s u m e r P r i c e I n d e x ( N e w S e r i e s ) 1 R e l a t iv e I m p o r t a n c e o f M a jo r G r o u p s , S p e c i a l G r o u p s
I n d iv id u a l I t e m s S e l e c t e d w it h C e r t a i n t y 2 D e c e m b e r 1963— Continued
Percent of
all items
December 1963

Health and recreation..............................................................
Medical care______________ ___________________
Drugs and pharmaceuticals___________________
Over-the-counter items _
Professional services__
F am ily doctor, house visit
Fam ily doctor, office v isit_ __ __ _ _ __ _ ___

Optometric examination and eyeglasses___-___

Dentists* fees
Other priced items
H ospital services
H ealth insurance 4
Hospital services
Nonhospital services
Personal care . _
Toilet goods

Men’s haircut


_ . ...


Other priced items
Reading and recreation
Recreation .

Recreational goods_______________________
T V sets

Other priced items____ _______________


Percent of
all items
December 1963

Health and recreation—Continued
Reading and recreation—Continued

Recreational services _ _
M ovies ( indoor)
Bowling fees
Other priced items
Reading and e d u catio n ___
Newspapers _

_ _ _ _ _


College tuition________ ______ ___________

Other priced items _ _ _
Other goods and services _






Tobacco products___________________________
Other priced items_______________________
Alcoholic beverages__________________________
Whiskey and w ine__ _ _____ _ _______
Away from home__ ___ ___
_ _ ____
Personal expenses___________________________
Funeral services_____________________ ____
Bank service charges________________ _ _ __
Legal services_________ _____
___ __

Miscellaneous3 5__________________________________


i For a description of the new series, see The Consumer Price Index,
{Revised January 1964). A Short Description.
* The list of items priced includes all the more important goods and services
and a sample of the less important ones. In combination, these represent
all items included in the CPI. Weights for individual certainty items are
shown separately. Some of them, however, are represented by more than
one specification, but the weights for the individual specifications are not
shown. The remaining weight of each expenditure class having both cerainty and probability items was shared equally by the probability items as

of December 1963, except in a few cases where weights for duplicated items
have double weights.
* Not actually priced; imputed from priced items.
4 Cost of health insurance is imputed to price changes for representative
services plus the cost of overhead. For a more complete discussion, see
article by James C. Daugherty, “ Health Insurance in the Revised C P I/'
Monthly Labor Review, November 1964, pp. 1299-1300.
* Personal financing charges other than mortgage interest and auto financ­

foods are a significant part of total spending, food
pricing is conducted every month in each urban
location. Prices of most other goods and services
are collected every month in the five largest urban
areas and every 3 months in all other places. Pric­
ing of food is done each month on 3 consecutive
days early in the month; rents and items for which
prices are obtained by mail are reported as of the
15th of the month; pricing of other items extends
over the entire calendar month. The Bureau uses
mail questionnaires to obtain data on streetcar and
bus fares, public utility rates, newspaper prices,
and prices of certain other items which do not
require personal visit by Bureau agents. For a
number of items, e.g., home purchase, college tui­
tion, used cars, magazines, etc., data collected by
other Government agencies or private organiza­
tions are used.
To insure that the index reflects only changes in
prices and not changes due to quantity or quality
differences, the Bureau has prepared detailed spec­
ifications which describe the physical character­
istics of the items in the market basket. Specially
trained Bureau representatives examine merchan­
dise in the stores to determine whether the goods
and services for which they record prices conform

to the specifications. Where the precisely specified
item is not sold at a particular retail establishment,
the Bureau’s representative quotes prices and
obtains a detailed technical description of the item
nearest to the physical characteristics of the spec­
ification, in order to insure that prices will be
quoted on the same quality and quantity from
time to time. At the first pricing in an outlet the
agent selects the volume selling item meeting spec­
ification, making sure that it is regular merchan­
dise in good condition and available in a customary
assortment of colors, patterns, etc.
At the subsequent pricings the agent prices the
identical item if it is still available in a reasonable
assortment and selling in substantial volume. I f
it is not, she must substitute another volume selling
item, meeting specification, if possible. I f she
cannot, she prices an item deviating from specifi­
cation. Prices of substitute items meeting specifi­
cation are compared directly. Prices of deviating
items are introduced by linking or splicing in
such a way that the difference in price between
the specification and the deviating item is not re­
flected as a price change. I f it is possible to obtain
an estimate of the value of the quality difference,
prices of the previous item are adjusted by the


quality estimate and compared with prices for the
current month.
When the sample of reporters is changed for any
reason, prices from the new reporter also are intro­
duced by linking.

A complicated index such as the CPI must be
based on a whole complex of samples.16 A sample
of cities or areas is required in which expenditure
surveys and price collection will be conducted.
Within each area there must be a sample of fam­
ilies or consumer units, from which consumer ex­
penditures will be obtained. It is convenient, but
not essential, that the city sampling points be the
same for price collection as for the Consumer Ex­
penditure Surveys.
Further, since it is impossible to price all the
thousands of items which consumers buy, it is
necessary to select a sample of items for pricing,
to represent price movement of all items. Samples
of outlets are needed at each sampling point in
which price quotations are obtained for the selected
items. Finally, pricing usually is done at a
specific time of the month or quarter so there is,
in effect, a sampling of time.
In the 1964 revision, the Bureau of Labor
Statistics used probability sampling to a greater
extent than had been done previously, despite the
well-known difficulties involved. At the same
time, the Bureau attempted to include in the CPI
sampling design some method for obtaining an
estimate, even if only a crude one, of sampling
error. Probability sampling is a necessity, of
course, if this is to be done in a conventional man­
ner. However, even if probability sampling could
be followed rigorously through all the complicated
CPI structure, the mere computational load would
be so extensive that it would be impractical to com­
pute measures of error except by some “simple”
approach. The objective, therefore, has been ap­
proached by the “replication” method.
The sample design includes an ex post facto pairing of probability cities (or Standard Metro­
politan Statistical Areas), two replicated item
samples, and replicated outlet samples. In addi­
tion to the minimum program, designed to produce
an estimate of the total sampling error in the index


from all sources, the structure includes more ex­
tended replication in selected cities aimed at per­
mitting some evaluation of the components of the
error, that is, variation in sampling results due to
sampling of cities, items, and outlets.
A core sample of 50 SMSA’s (see
footnote 13) or smaller cities for the index, sup­
plemented by 16 additional D size cities for the
family expenditure surveys was the maximum size
consistent with available budget. These addi­
tional D size cities were surveyed because expendi­
ture patterns are more variable among small cities
than among large cities. The primary sampling
units (PSTPs) are Standard Metropolitan Sta­
tistical Areas as they were defined by the Bureau
of the Budget prior to the 1960 Census, except that
the Standard Consolidated Areas for New York
and Chicago were used, plus individual urban
places outside the SMSA’s. Because 1960 Census
data were not then available, the measure of size
used in sample selection was the estimated urban
population as of January 1,1959. The population
weights actually assigned are based on 1960 data.
The P SU ’s were stratified by broad region and by
size into 12 regional-size strata. The 12 largest
SMSA’s were selected with certainty, that is, they
represent themselves in the sample design. Since
Alaska and Hawaii have been added in the revised
CPI, one sample selection has been allocated to
each of these two States. The remaining 36 selec­
tions are allocated to the 12 regional-size strata on
the basis of relative population and relative costs
of pricing cities of different size. Four size strata
are defined as follows:
A. The 12 largest SMSA’s on the basis of urban
population, in effect those with population over
B. Other large SMSA’s with urban population
greater than 250,000;
C. SMSA’s with urban population of 50,000250,000; and
D. Nonmetropolitan urban places with popula­
tion less than 50,000.
The method of selection used is known generally
as “controlled selection” which was described by
Koe Goodman and Leslie Kish in the September
City Sample.

MFor a more detailed discussion, see Sampling Aspects of the
Revised CPI, unnumbered paper by Marvin Wilkerson, Bureau of
Labor Statistics, U.S. Department of Labor, 1964.



1950 issue of the Journal of the American Statis­
tical Association (pp. 350-372). This method
accomplishes a good geographic dispersion of
sampling points across the country.
After the initial 50-area sample was selected,
the B L S received funds to prepare city indexes for
six additional large SMSA’s—Cincinnati, Hous­
ton, Kansas City, Milwaukee, Minneapolis-St.
Paul, and San Diego—as part of a plan to publish
indexes for each SMS A with 1,000,000 total popu­
lation in 1960. These areas were added to the
national index in January 1966.
Samples of Consumer Units . The CES samples
were chosen as subsamples of housing units enu­
merated in advance Comprehensive Housing Unit
Surveys (CHUS) conducted in each area late in
the year preceding the actual survey date.17 The
CHUS also serve as the source of the samples of
rental dwellings for measuring price change in
rents, and of owner occupied units for measure­
ment of changes in property taxes. They also
provide data for weights for home purchase. The
actual size of the CHUS sample in an area is
determined primarily by the rental sample desired
and by the proportion of renters in the given area,
as estimated from Census data. The number of
addresses enumerated in the CHUS is usually
many times larger than it is in the samples for
surveys of consumer expenditures.
Sam pling of Item s . A classification system has

been developed to provide a logical publication
framework containing the traditional major ex­
penditure groups, subgroups, etc., but, in a broader
sense, to divide the thousands of goods and services
purchased by consumers into meaningful and
manageable components of the universe. It pro­
vides the framework for the selection of the item
sample and for the derivation of index weights.
Two levels of the classification system are of
critical importance. These are: (1) the item level,
and (2) the level which defines the finest stratifica­
tion for the item sampling; that is, the strata to
which allocations of items are made and within
which probability samples of items are selected.
The term “expenditure class” (EC) is given to this
17 The selection of the CES sample is discussed in chapter 8
of this bulletin.

level. The expenditure classes are primarily
groupings of items which serve similar human
needs. Items are grouped within an EC so that
they are as homogeneous as possible with respect
to their physical characteristics. It is not possible
to confine groupings to items which are similar
with respect to price movements.
Within an expenditure class base period expen­
diture weights will be held constant; that is, the
EC expenditures serve as a way of defining the
level of living which is to be held constant until
the next major revision of the CPI. The Bureau
plans to resample items within an EC between
major revisions whenever there is evidence of a
major redistribution of relative expenditures or
indications that the previous sample of priced
items does not adequately represent the class. The
connotation of “item” in the sampling frame is
necessarily fairly broad and the items are not of
equal homogeneity in the different classes. Gen­
erally the listing is above that of the final “speci­
fied-in-detail” items for which prices are collected.
For the most part no attempt has been made to
carry probability sampling to this ultimate stage.
There were about 1,800 line items in the expendi­
ture survey schedule. After extensive experi­
mentation, using expenditure data from a 1959
pilot survey in Cincinnati, a final sampling frame
containing 52 EC ’s and 812 items was developed.
The list of EC ’s and the number of items in each
are shown in table 3.
The first step in the selection of the item sample
for the revised index was to make a roughly opti­
mum allocation of the total number of items to be
priced to each EC. Factors considered were the
relative importances of the EC ’s and a rough
measure of variability of price movement.
As in past revisions of the CPI, the samples
were selected on a national basis. Selection of
independent samples, city-by-city, is not practical
since it would result in a huge list of items to be
priced in at least one city and an impossible burden
of writing and keeping up with changes in
The two replicated samples of items of the re­
vised CPI have been selected with “probability
proportional to size,” size being defined as the
relative importance of the expenditures for the
item to total expenditures for all items. The gen-




3. N u m b e r


I tem s


S a m p l in g F r a m e

class number


N um ber


I t e m s P r ic e d


E x p e n d it u r e C l a s s

Number of Number of Number of

________________ __________




Food__ ___ ____. ..
Food at home:
Cereals and bakery products:
EC! 1
Cereals and grain products _ _ __ _ _ _______ ______ _______________
Tin 9
Bakery products I
______ ___ _______________
Meats, poultry, and fish:
EC 3................
Beef and veal__________________________________ _________________
Pork. _____________ _____________________________ ____________
Other meats_______ _ ___________ _________ . . . _____
EC! 4
Poultry_____ _ _____ _ _ _ _ _ _ _ _ _ _ _ _ _ _______ _ __ _ _ _________
Fish__'._ ___________ _ _______ ____ _______ _ __ ___ ______________
Dairy products:
EC! fi
Dairy products___________ _ _ _ _ _ _ _ _
______ ____ _ __ _ __________
Fruits and'vegetables:
EC! 7
Fresh fruits _
_ ________ _ ______________
EC! 8
___ _________
Fresh vegetables
Tin q
Processed fruits and vegetables
_ __
Other food at home:
EC! in
Eggs _
____ ___________ __ _ ________ _____
Tin ii
___ ____ ________ _ _______ ____
__ _ __ _____ ___
Fats and oils.
Tin 12
Sugar and sweets______ ___________ _ _ __ ___ ____ _ _________________
E C 13_______
Nonalcoholic beverages... __ _____ _ _ _ _ _ _ _ _ ____ _______
EC! 14
Prepared and partially prepared foods___
_ _ _____ _______ ___________
Food away from home*.____ l . . . 1__
_ ____ _
__ . __ ______________










All items.__












Housing___________________________________ ___ __ _______ _ _____ _____ _________
ShelterEC 16...............
_______ _ __ ___________
Purchase and financing___
_ ___________ ______________
EC 17- ____
EC 18
Taxes and insurance__ _ _______ _____ _ _ _ _______ _____________
Maintenance and repairs:
EC 19
___ _______ ____ _____ __________ _______________
EC 20
_____ ____ ____ _____ _ ________ _______________
EC 21
Fuel and utilities__ _______ _ __ ___________________________ _______ _____ ___
Household furnishings and operation:
EC 22
Textile housefurnishings.— ____ ______ _______ _________ __________________
Furniture and floor coverings:
EC 23
____________ ____ _ __________ _ __ ______________
EC 24_______
Floor coverings__
_ __ __ ___ ____ __ _____________
EC 25_______
__ __________ ___ _______ _____ ______________
EC 26_______
Other housefurnishings__ _____ _ _ _ _ ________ ___ __ _ _____________
Household operation:
EC 27
Housekeeping supplies_____________ ________ _ _ _ ________ _______ _
EC 28
Housekeeping services___ _____ _._ __ _ ____ _____ _ ____ _______






















Apparel and upkeep____ _ _______________
____ ___________________ _ __
Men’s and boys’ apparel:
EC 29_______
Men’s apparel__
_ _______
. _ ___ _______________
EC 30_______
Boys’ apparel.
_ __ ______ _
_ ______ _______
Women’s and girls’ apparel:
EC 31...............
Women’s apparel, ______
_____ _______ _ _____________________
__ __
EC 32...............
Girls’ apparel___
___ ____ ______________ ______ _____ _______ _______
EC 33___ ___
Footwear___ ___
_______ ______
_ _ __ _ __________ __________
Other apparel:
EC 34_______
______ _______ __ ____ _______________ ___
EC 35..... .........
Services._ _____ _________________________________________________________












Transportation__ _____ _______
___________ _ _ _____________ _ _____________
Autos and related goods:
Auto purchase. _ _ ____________ ___ __________ _______________ _
Gasoline and motor oil._________
____ _
Auto parts, etc
Automobile services:
Auto repairs and maintenance_______ ____ ___ __________ ________ _______
Other automobile expenses
. ______ __________________










Health and recreation. ___________________________________________________________
Medical care:
Drugs and prescriptions_____________ _____________ __ ______ ____ _____
Professional services..
__________ ____ _____ _______
Hospital services and health insurance____ _ __ _ ____
__ _______ ______
Personal care:
Toilet goods____
____ _
_ __________
Reading and recreation:
Recreational goods._
______ ___ __ ____________
Recreational services ________ ___ _ ____________________________
Reading and education_________ _____ ___ ___________ ___
_ __
Other goods and services:
Tobacco products__ ______________________ ____________________ ____ _ _
Alcoholic beverages. _
__ ______________
Financial and miscellaneous personal expenses_________________















EC 36___ ____
EC 37...............
EC 38...............
EC 39_______
EC 40—_..........
EC 41_______

EC 42_______
EC 43_______
EC 44_______
EC 45_______
EC 46_______
EC 47_______
EC 48_______
EC 49_______
EC 50_______
EC 51_______
EC 52_______




eral procedure was to array items within a
stratum and by using a random start to make regu­
lar selections along the array. Each of the two
replicated samples thus contains “certainty
items;” that is, items which are certain of inclu­
sion because their relative importance is greater
than the selecting interval. The replicated
samples also contain some duplicates of items
selected but not with certainty. Table 4 contains
a listing of the items priced in each sample.
For the final selection, relative importances (in
the family expenditure pattern) for the condensed
sampling frame (52 EC ’s and 812 items) were
obtained from expenditure data for nine of the
cities surveyed for 1960. Ideally, of course, the
data should have covered all 66 cities, but such data
were not available in time for use in selection of
items. Expenditure data for these nine places
were weighted together to give preliminary esti­
mates of U.S. average expenditures. (Final
index weights of course are based on complete data
for all cities.)
The selection of one or more specifications or
“specified-in-detail” items to represent the items
selected from the sampling frame has been made
in most cases by commodity specialists from expert
knowledge of the item. Factors taken into con­
sideration are the importance and representative­
ness of particular qualities and the feasibility of
describing a selected item clearly enough to permit
repetitive price collection. In a few cases where
sufficient data existed, it is possible to make a
second stage probability selection of specifications.
Outlet Sam pling . The first big problem encount­
ered in attempting probability sampling of outlets
was to obtain information about the universe of
retail and service establishments in a given area.
Ideally, names and addresses of such places, infor­
mation as to type of store or outlet, some indica­
tion of volume of sales, and preferably fairly
specific information as to types of merchandise
carried would have been desirable.
Comprehensive establishment data were ob­
tained from a list of firms which report to the
Bureau of Old-Age and Survivors Insurance
(Social Security Administration, U.S. Depart­
ment of Health, Education, and Welfare). Using
sampling ratios furnished by BLS, master sam­
ples of retail and service outlets were selected by

BOASI. These were supplemented with listings
from other sources.
In the larger SMSA’s, a two-stage sampling pro­
cedure has been followed. Samples of neighbor­
hood and suburban localities and shopping
centers have been selected in which pricing outside
the downtown area is conducted. These were
selected with probability proportional to sales vol­
ume, using the best available sales data. The list­
ings of sample outlets were limited to those falling
within the sampled areas.
The number of food stores priced varies from
less than 10 in the smallest cities to about 80 in
New York. The number of quotations for non­
food items per city is quite small; the basic num­
ber in each outlet sample is four. This means
that for the cities in which both item samples are
priced eight is the maximum sample size even for
items appearing in both item samples. In a few
“A” cities, the sample sizes are set at 5 per sample
or a maximum of 10. At the U.S. level, however,
the number of quotations is sizeable.
In selecting the sample, allocations of quotations
were made for each item by type of outlet, based
on available sales data, “where bought” surveys,
etc. As a specific example, if eight quotations
are required for a particular woman’s shoe specifi­
cation, the allocation might be three quotations to
department stores, two to women’s specialty shops,
two to women’s shoe stores, and one to family shoe
stores. Specific allocations also are made by loca­
tion within the SMS A (central business district,
neighborhood centers, and suburbs) and, in
some cases, to multiunit and independent
In addition to the pricing in regular retail and
service outlets, there are a number of special items
whose nature requires separate samples of specific
types of “outlets;” for example, samples of phy­
sicians and other medical specialists, restaurants,
dairies, hotels, property owners, etc. Each of
these offers its own particular problems.
When the original samples, which were selected
in Washington, were sent to the regional offices
a great many practical problems were encountered,
and many expedients and compromises with strict
probability procedures were required to complete
the initiation of pricing for the revised index.
However, even though some deviations from prob­
ability sampling were inevitable and had been an-

T a b l e 4.

L is t of I tem s P r ic ed for th e R e v ise d C o n su m er P r ic e I n d e x a s of D e c e m b e r 1963
Priced items

Groups, subgroups, expenditure classes
Sample A


Food at home:
Cereals and bakery products:
Cereals and grain products.


Bakery products................


Meats, poultry, and fish:
Beef and veal........




Other meats.






Dairy products.


Fruits and vegetables:
Fresh fruits..........








Com flakes................ .......
Rice, long and short grain.
White bread......................
Whole wheat bread-.........
Layer cake, plain.............

Flour, white, all-purpose.
Cracker meal.
White bread.
Cookies, cream filled.
Cinnamon rolls, frosted.

Hamburger, preground........................
Steaks, round, bone-in . ............ ..........
Steaks, porterhouse, bone-in................
Rump roasts, standing____________
Chuck roasts, bone-in..........................
Veal cutlets, bone-in......... ...................
Pork chops, center c u t.......................
Bacon, sliced........................................
Pork roasts, loin halves.......................
Picnics, smoked...................................
Lamb chops, loin.................................
Salami sausage, sliced..........................
Frankfurters, skinless.........................
Frying chickens, ready-to-cook............
Chicken breasts, fresh.................. ........
Fillets or steaks, fresh or frozen *.........
Tuna fish, chunk style_____ _______
Milk, fresh, grocery.............................
Milk, fresh, delivered_____________
Milk, fresh, skim________________
Ice cream, prepackaged............... ......
Butter, salted______ ___ ________

Hamburger, preground.
Steaks, round, bone-in.
Steaks, sirloin, bone-in.
Rump roasts, standing
Rib roasts, bone-in.
Beef liver, sliced.
Pork chops, center cut.
Bacon, sliced.
Pork roasts, loin halves.
Ham, whole.
Bologna sausage, sliced.
Liverwurst sausage, sliced or whole.
Ham, canned, domestic or imported.
Frying chickens, ready-to-cook.
Turkey, fresh or frozen.
Shrimp, raw. frozen.
Sardines, Maine.
Milk, fresh, grocery.
Milk, fresh, delivered.
Milk, evaporated, canned.
Cheese, American process.
Butter, salted.

Apples, all purpose...............................
Bananas, yellow variety......................
Oranges, except Temple or King____
Grapes, Thompson seedless. ............ .
Grapefruit, fresh, pink or white..........
Orange juice, fresh...............................
Fresh vegetables.
Head lettuce........................................
Potatoes, white__________________
Asparagus, green..................................
Carrots, topped, prepackaged............
Cucumbers ......................................... .
Spinach, prepackaged______ _____ _
Processed fruits and vegetables— Pears, Bartlett, can or jar....................
Lemonade, concentrate, frozen............
Beets, sliced, can or jar._____ ______
Tomatoes, can or jar.............................
Dried beans, Navy or Great Northern.
Other foods at home:
Eggs, fresh, large, Grade A..................
Fats and oils.................
Margarine, colored...............................
Salad dressing, Italian..........................
Sugar and sweets..........
Sugar, white, granulated......................
Chocolate bar, plain m ilk..................
Nonalcoholic beverages.
Coffee, can or bag................................
Carbonated drinks, fruit-flavored____
Tea bags, orange pekoe and pekoe tea..
Prepared and partially pre­
Bean soup, canned, condensed............
pared foods.
Spaghetti, in tomato sauce, canned___
Mashed potatoes, instant__________
Potatoes, French fried, frozen. ............
Food away from home.
Restaurant meals:
Between meal snacks:
Coffee, cup......................................
Carbonated beverages, cup______
Frankfurter on roll..........................
Ice cream, dish............................... .
Rent...................................... .......... Rent of house or apartment........................
Hotel, motel room rates............................ .
Home purchase and financing... Home purchase -..........................................
Mortgage interest rates................ -............
Taxes and insurance................... Property taxes, residential........................
Property insurance rates:
Fire and extended coverage................
Comprehensive homeownership policy
Maintenance and repairs:
Commodities. ..........
Exterior house paint...................................
Furnace air filters.......................................
Packaged dry cement mix..........................

See footnotes a t end of table.

Sample B

Residing houses. .
Reshingling roofs.
Replacing sinks..

Apples, all purpose.
Bananas, yellow variety.
Oranges, except Temple or King.
Grapes, Thompson seedless.
Strawberries, fresh.
Watermelons, whole or sliced.
Head lettuce.
Potatoes, white.
Cabbage, all varieties except red.
Celery, Pascal, stalk.
Onions, yellow.
Peppers, sweet, green.
Fruit cocktail, canned.
Pineapple-Grapefruit juice drink, canned.
Orange juice concentrate, frozen.
Peas, green, can or jar.
Broccoli spears, frozen.
Eggs, fresh, large, Grade A.
Margarine, colored.
Salad or cooking oil, vegetable.
Grape jelly, pure.
Chocolate flavored syrup.
Coffee, can or bag.
Coffee, instant.
Cola drink, carbonated.
Chicken soup, canned, condensed.
Baby foods, strained.
Sweet pickle relish.
Pretzels, hard, salted.
Restaurant meals:
Between meal snacks:
Coffee, cup.
Carbonated beverages, cup.
Pie, slice.
Candy bar.
Rent of house or apartment.
Hotel, motel room rates.
Home purchaseMortgage interest rates.
Property taxes, residential.
Property insurance rates:
Fire and extended coverage.
Comprehensive homeownership policy.
Interior house paint.
Shelving, Ponderosa pine.
Shrubbery, evergreen.
Residing houses.
Repainting living and dining rooms.
Repairing furnaces.





L ist


I t e m s P r ic e d

fo r th e


e v is e d

C o n su m e r P r ic e I n d e x

Groups, subgroups, expenditure classes
Sample A



Fuel and utilities.

Household furnishings and operation:
Textile housefurnishings..............




Floor coverings.............


Appliances................ .


Other housefurnishings.


Housekeeping supplies.


Housekeeping services.


Apparel and upkeep:
Men's and boys' apparel:
Men’s apparel......... .


Boys’ apparel.................


Women's and girls' apparel:
Women’s apparel..........


Girls' apparel.




Other apparel:



See footnotes a t end of table.

as of


ecem ber

1963— Continued

Priced items

Fuel oil and coal:
Fuel oil, #2....................... .................................
Coal, anthracite or bituminous........................
Gas and electricity:
Gas, 3 bills per city__________ ______ ____
Electricity, 3 bills per city................................
Other utilities:
Residential telephone services_____________
Residential water and sewerage services_____

Sample B

Fuel oil, #2.
Coal, anthracite or bituminous.
Gas, 3 bills per city.
Electricity, 3 bills per city.
Residential telephone services.
Residential water and sewerage services.

Pillows, bed, polyester or acrylic filling..................
Curtains, tailored, polyester marquisette---------Drapery fabric, cotton or rayon/acetate........ .........
Bedroom suites, good or inexpensive quality____
Living room suites, good and inexpensive quality.
Lounge chairs, upholstered............................. .......
Sofas, dual purpose________ ______ _________
Sleep sets, Hollywood bed type______ ________
Aluminum folding chairs_____ ________ ______
Rugs, soft surface:
Broadloom, wool....................... .......................
Broadloom, nylon.................................. ...........
Rugs, hard surface______________ __________
Refrigerators or refrigerator-freezers, electric_____
Washing machines, electric, automatic____ _____
Ranges, free standing, gas or electric____ ______
Clothes dryers, electric, automatic.............. ...........
Room heaters, electric, portable.............................
Dinnerware. earthenware...................... .................
Carpet sweepers, manually operated___ _______
Venetian blinds, white, steel or aluminum slats...
Electric drills, hand held— ............ ......................
Detergent, liquid, laundry......................................
Laundry soap for fine fabrics..................................
Scouring pads, steel wool.........................................
Toilet tissue__________________ _____ ______
Domestic service, general housework___________
Baby sitter service..................................... ............
Postal services............ ...... ......................... ............
Laundry flatwork, finished service---------------Licensed day care service, preschool child______
Washing machine repairs.........................................

Sheets, percale or muslin.
Bedspreads, chiefly cotton, tufted.
Slipcovers, ready-made, chiefly cotton.
Bedroom suites, good or inexpensive quality.
Living room suites, good and inexpensive quality.
Dining room suites.
Sofas, standard, upholstered.
Box springs.
Rugs, soft surface:
Broadloom, wool.
Broadloom, nylon.
Tile, vinyl.
Refrigerators or refrigerator-freezers, electric.
Washing machines, electric, automatic.
Vacuum cleaners, cannister type.
Air conditioners, demountable.
Garbage disposal units.
Flatware, stainless steel.
Table lamps, with shade.
Lawn mowers, power, rotary type.
Nails, 8d (penny) common.
Detergent, granules or powder.
Air deodorizers, spray type.
Paper napkins, embossed.
Stationery, envelopes.
Domestic service, general housework.
Baby sitter service.
Postal services.
Laundry flatwork. finished service.
Reupholstering furniture.
Moving expenses.

Suits, year round weight, 2 qualities......................
Topcoats, wool.................................. ......................
Suits, tropical weight..............................................
Slacks, wool or wool blend......................................
Shirts, work, cotton.................................................
Shirts, sport, cotton, short sleeves..........................
Shirts, sport, cotton, long sleeves...........................
T-shirt____ ___ __________________________
Coats, all purpose, cotton or cotton blend_______
Dungarees, cotton or cotton blend..........................

Suits, year round weight, 2 qualities.
Jackets, lightweight.
Trousers, work, cotton.
Slacks, cotton or manmade blend.
Shirts, business, cotton.
Socks, cotton.
Handkerchiefs, cotton.

Coats, heavyweight, wool or wool blend, 2 quali­
Carcoats, heavyweight, cotton........................ ......
Skirts, wool or wool blend............................... ......
Skirts, cotton or cotton blend...............................
Dresses, street, chiefly manmade fiber, 2 qualities.
Dresses, street, wool or wool blend_____ _______
Dresses, street, cotton.............................................
Housedresses, cotton....................... ........................
Slacks, lightweight, cotton and carded cotton___
Slips, nylon.................................................... .........
Brassieres, cotton....................................................
Hose, nylon, full fashioned and seamless, 2 styles..
Anklets, cotton......................................................
Handbags, rayon faille or plastic..........................
Raincoats, vinyl plastic or chiefly cotton...........
Skirts, wool or wool blend................................. .
Slips, cotton blend..................................... _...........
Handbags, plastic....................................................

Coats, heavyweight, wool or wool blend, 2 qualities.
Coats, lightweight, topper.
Sweaters, wool or acrylic.
Dresses, cocktail, street length.
Dresses, street, chiefly manmade fiber, 2 qualities.
Dresses, street, wool or wool blendDresses, street, cotton.
Blouses, cotton.
Bathing suits, one piece.

Sport coats, wool or wool blend.
Undershorts, cotton.

Girdles, manmade blend.
Panties, acetate.
Hose, nylon, full fashioned and seamless, 2 styles.
Gloves, fabric, nylon or cotton.
Coats, lightweight, topper.
Slacks, cotton.
Shorts, cotton.
Dresses, cotton.
Robes, duster style, quilted tricot, or percale.

Shoes, street, oxford, 2 qualities...................... . Shoes, street oxford, 2 qualities.
Shoes, work, high.
Shoes, street, pump, 2 styles............................. Shoes, street, pump, 2 styles.
Shoes, evening, pump.......................... ............ Shoes, evening, pump.
Shoes, casual, pump........................................
Houseslippers, scuff...........................................
Sneakers, boys', oxford type............................. Shoes, oxford.
Dress shoes, girls’, strap.
Diapers, cotton gauze..............................................
Yard goods, cotton.............................................. .
Earrings, Pearl, simulated or imitation........ .........
Dry cleaning, men’s suits and women’s dresses__
Shoe repairs, women’s heel lift...................... .........
Laundry, men’s shirts.............................................

Wrist watches, men’s, imported movement.
Wrist watches, women’s, imported movement.
Zipper, skirt or neck placket.
Dry cleaning, men’s suits and women’s dresses.
Automatic laundry service.
Tailoring charges, hem adjustment.



a ble


L ist


I t e m s P r ic e d

fo r t h e


e v is e d

C o n su m e r P r ic e I n d e x

as of


Sample B

Sample A







Auto purchase

New cars:
Chevrolet, Impala, 2-door, hardtop................ .
Chevrolet, Chevelle, 2-door hardtop—..........
Ford, Galaxie 500,2-door hardtop................... .
Plymouth, Fury III, 4-door sedan...................
Rambler, Classic 660,4-door sedan................. .
Used cars:
2 years old, Chevrolet and Ford______ ____ _
3 years old........ do........................................... .
4 years old........ do........................................... .
5 years old........ do........................................... .
Gasoline and motor oil...................... Gasoline, regular and premium............................. .
Motor oil, premium................................................
Auto parts......................................... Tires, tubeless, retread...........................................
Automobile services:
Auto repairs and maintenance.. Chassis lubrication, complete...............................
Motor tune-up....................................................... .
Automatic transmission repair.............................. .
Other automobile expenses.
Auto insurance rates, liability and physical
Auto financing charges2..........................................
Auto registration and inspection fees......................
Driver’s license fees................................................
Parking fees, private and municipal.......................
Public transportation.
Local transit fares.......................... .......................
Taxicab fares............................................................
Railroad fares, coach...............................................
Airplane fares, chiefly coach....................................
Bus fares, intercity..................................................
Health and recreation:
Medical care:
Over-the-counter items:
Drugs and prescriptions.
Multiple vitamin concentrates.........................
Liquid tonics.....................................................
Cold tablets or capsules....................................
Penicillin G buffered tablets......................
Sulfisoxazole tablets.
Sedatives and hynotics:
Phenobarbital tablets.................................
Chlordiazepoxide-hydrochloride capsules..
Propantheline Bromide tablets_________
Cardiovasculare and antihypertensives:
Reserpine tablets........................................

Professional services.

Antiarthritics________________ _________
Cough preparations:
Terpin hydrate with codeine, elixir.
Family doctor, office visits. ....................................
Family doctor, house visits_____ ___ ________
Pediatric care, office visits__________ ____ ___
Psychiatrists, office visits......... ..................... .......
Routine laboratory tests____ _______________
Examination, prescriptions and dispensing of
Fillings, adult, amalgam, one surface.....................
Dentures, full upper_______ ________________

Hospital services and health in­
Hospital services.................. Daily service charges:
Semiprivate room___________ _____ _____
Private room........ .......... .............. .................
Hospital services:
Health insurance3.
Daily service charges, semiprivate room_____
Daily service charges, private room.................
Operating room___ ___ _________________
Nonhospital services:
Family doctor, office visit________________
Surgeon’s fees, tonsillectomy/adenoidectomy..
Personal care:
Toilet goods.
Personal care services.

See footnotes a t end of table.

1963— Continued

Priced items

Groups, subgroups, expenditure classes


ecem ber

Chevrolet, Impala, 2-door hardtop.
Ford, Falcon, Futura, 4-door sedan.
Ford, Galaxie 500,2-door hardtop.
Pontiac, Catalina, 4-door sedan.
Volkswagen, Deluxe, 2-door hardtop.
2 years old, Chevrolet and Ford.
3 years old,
4 years old,
5 years old,
Gasoline, regular and premium.
Motor oil, premium.
Tires, tubeless, new.
Water pump replacement.
Replacing muffler.
Front end alignment.
Auto insurance rates, liability and physical damagea
Auto financing charges.2
Auto registration and inspection fees.
Driver’s license feees.
Parking fees, private and municipal
Local transit fares.
Taxicab fares.
Railroad fares, coach.
Airplane fares, chiefly coach.
Bus fares, intercity.
Aspirin compounds.
Cough syrups.
Adhesive bandages, package.
Tetracycline capsules.
Secobarbital sodium capsules.
Meprobamate tablets.
Phenobarbital and belladonna extract.
Crystalline digitoxin tablets.
Chlorothiazide tablets.
Prednisone, tablets.
Family doctor, office visits.
Family doctor, house visits.
Obstetrical cases.
Chiropractors and podiatrists, office visits.
Herniorrhaphy, adult.
Examination, prescriptions and dispensing of

Fillings, adult, amalgam, one surface.
Extractions, adult.

Daily service charges:
Semiprivate room.
Private room.
Hospital services:
Daily service charges, semiprivate room.
Daily service charges, private room.
X-ray diagnostic series, upper G.I.
Nonhospital services:
Family doctor, office visit.
Surgeon’s fees, hemioirhaphy, adult.
Obstetrical cases.
Prescriptions and drugs____ _____________
Prescriptions and drugs.
Retained earnings (overhead)________________ Retained earnings (overhead).

Toothpaste, standard dentrifice..........................
Hand lotions, liquid____ _________ ______ ___
Face powder, pressed_________ _____________
Cleansing tissues.......................... .............. ............
Men’s haircuts____________________________
Shampoo and wave sets, plain....................... ........
Women’s haircuts_____________ _____ „_____

Toilet soap, hand milled.
Shaving cream, aerosol.
Deodorants, cream or roll-on.
Home permanent refills.
Men’s haircuts.
Shampoo and wave sets, plain.
Permanent waves, cold.



T a b l e 4.

L is t of I tem s P r ic ed fo r th e R e v ise d C o n su m er P r ic e I n d e x a s of D e c e m b e r 1963—
Priced items

Groups, subgroups, expenditure classes
Sample A

Health andrecreation—Continued
Reading and recreation:
Recreational goods---EC-47




Recreational services.

Reading and education.
Other goods and services:
Tobacco products__


Sample B

TV sets, portable and console......... .......................
Radios, portable and table models, AM band only.
TV replacement tubes...................... ......................
Sports equipment:
Golf balls, liquid center.................. .................
Basketballs, rubber or vinyl cover...................
Outboard motors....................................................
Tricycles... ......... ............... .............. ............ ........
Dolls........ ............ ..................................................
Stuffed animal........................................................
Dog food, canned and boxed....................................
Indoor movie admissions:
Adult__ ______________ _____ _____ ___
Children’s______________________ ______
TV repairs, picture tube replacement.....................
Bowling fees, evening......................... ....................
Golf green fees.___ ____________ ___________
Newspapers, street sale and delivery....... ..........
College tuition and fees, undergraduate.................
Magazines, single copy and subscription................
College textbooks, undergraduate...........................

TV sets, portable and console.
Radios, portable and table models, AM band only.
Tape recorders, portable.
Sports equipment:
Fishing rods, fresh water spincast.
Bowling balls.
Phonograph records, stereophonic.
Bicycles, boys’, 26".
Movie cameras, 8 mm, fully automatic lens.
Film, 35 mm, color.

Cigarettes, nonfilter tip, regular size, pack.............
Cigarettes, filter tip, king size, carton-..................
Cigars, domestic, regular size_______ _____ ___
Beer, at home, local and national brands_______
Whiskey, spirit blended and straight bourbon.......
Wine, dessert and table.—....... ............. ...............
Beer, away from home............................................

Cigarettes, nonfilter tip, regular size, carton.
Cigarettes, filter tip, king size, pack.
Cigars, domestic, regular size.
Beer, at home, local and national brands.
Whiskey, spirit blended and straight bourbon.
Wine, dessert and table.
Beer, away from home.

Indoor movie admissions:
Drive-in movie admissions, adult.
Bowling fees, evening.
Film developing, black and white.
Newspapers, street sale and delivery.
College tuition and fees, undergraduate.
Paperback books, not school or technical.
Piano lessons, beginner.


Alcoholic beverages.


Financial and miscellaneous personal
Funeral services, adult____ _____ ______ _____ Funeral services, adult.
Bank service charges, checking account................. Legal services, short form will.

i Two of the largest volume sellers among the following types of fish are
priced within each city, since within any given city, all varieties of fish are
not available: Frozen ocean perch and haddock; fresh cod, catfish, king
salmon, halibut, sole, and haddock.

2 Not actually priced; imputed from priced items.
* Four items are priced only for health insurance: Operating room, X-ray,
tonsillectomy, and retained earnings; prices for the remaining items are also
included as directly priced professional and hospital services.

ticipated, the final samples adhere to the original
basic structure to the maximum extent possible.
As a result, the main benefits of probability sampl­
ing have been achieved: lack of bias, representa­
tion of different types of outlets, sections of each
SMSA, etc.


Calculation Procedures

In the absence of major weight revi­
sions, and ignoring the problems of sampling, the
index formula is most simply expressed as:
^ iPoQ.o)
or by its algebraic equivalent, the dollar weighted
average of price relatives:



The index is a time series. As previously ex­
plained, it is a weighted average of price changes
for a sample of priced items, expressed as a rela­
tive of avearge prices in a reference base as 100.
Weights, which are based on annual consumer
expenditures, are kept constant from month to
month. The index measures changes as they oc­
cur. It is not adjusted for seasonal variation.18
The Bureau began publication of seasonally ad­
justed indexes in 1966, for selected components
which show a significant seasonal pattern of price

2 (M o)
This is the customary, oversimplified way of
writing a price index formula to show that the
q’s are held constant between major revisions.
In actual practice, the basic data for weights are
values which include allowances for unpriced

18 For a discussion of the problems involved in using varying
seasonal weights, see “ Use of Varying Seasonal Weights in Price
Index Construction,” by Doris P. Rothwell, in the Journal of
the American Statistical Association, March 1958, pp. 66—77.
19 Factors used to compute seasonally adjusted indexes are avail­
able on request.


items, and the current index is computed by a
chain computation procedure, as shown below:
2(Pi-s4o)^ (pWa) X100
2 (p 0qo) ^ ( p '- s ia )


where q is a derived composite of the annual quantities
pm-chased in a weight base period for a bundle
of goods and services to be represented by the
specific item priced
p and p ' are the average prices of the specific
commodities or services selected for pricing (the
superscript indicates that the average prices are
not necessarily derived from identical samples of
outlets and specifications over long periods)
i —s is the month preceding a weight revision
(most recently, December 1963)
i is the current month
a is the period of the most recent Consumer Ex­
penditure Survey (1960-61) from which the
revised weights are derived
o is the reference base period of the index (1957-59).

The (p0q0) or (p ^ sqa) base “weights” for a
given priced item are the average annual expendi­
tures in a weight base period represented by that
item and other similar non-priced items. Al­
though constant physical weights are implicit in
the index, in reality the constant q*s are not
calculated separately.
In actual practice, the base expenditure for each
item is projected forward for each pricing period
by the price relative for the priced item:
iPiia) = {Pi-lia)

(~ )

In practice, then, the index formula is as follows:
(Dec. 1963 (Change from Dec.



1963 to month



==^ ( P i- s q o ) ^ ( P i- iq a )


2(Poqo) * 2 ( p 't- sqa)


(Change from month
i— 1 to month i )

w - g-> ( £




Illustrative Calculation. Average price changes
from the previous pricing period to the current
month are expressed as relatives (or ratios) for
each item, and the price changes for the various
goods and services are combined, using weighting
factors based on the importance of the item in
consumer spending and that of other items which
it represents. This composite importance is called
the cost weight of the market basket item. There
is a set of separate cost weights for each of the 56


urban locations included in the index. The fol­
lowing hypothetical example for pork illustrates
the index procedure:

September October
cost weight
September October October-5September
(Sept. X

Pork chops__
Ham .








Identical results could be obtained for pork by
multiplying prices each period by the implied
physical quantities included in the market basket,
as the following illustrates.
Sample item
Pork chops _

Implied September September October

1. 00

8. 00

1. 02


The average change in pork prices is computed
by comparing the sum of the cost weights in Octo­
ber with the comparable sum for September, as
October cost weight________$33.85 X 100=102.6
September cost weight______$33.00
This means that pork prices in October were 102.6
percent of (or 2.6 percent higher than) pork prices
in September.
Although the second method may appear
simpler, in reality it is not. Deriving the implied
quantity weights is an extra operation, and these
implicit quantities change as revised samples are
linked in. Furthermore, the second formulation
greatly complicates the handling of the numerous
substitutions of reporters and items which
occur constantly in repetitive index work. Con­
sequently, the first method is the one actually used
for the CPI. The second illustration, however,
may assist the user to understand the meaning of
the index mechanism.
After the cost weights for each of the items have
been calculated, they are added to area totals for
commodity groups and all items. The U.S. totals
are obtained by combining area totals, with each
area total weighted according to the proportion of



the total wage-earner and clerical-worker popula­
tion which it represents in the index based on 1960
Census figures. Finally, the U.S. totals for the
current and previous months are compared to com­
pute the average price change.
Reference Base Period . Since 1962 the index has

been calculated on the reference base of 195759=100.20 This means that current prices are
expressed as a percentage of prices for the average
of the 3 years—1957, 1958, and 1959. An index
of 110 means that prices have increased 10 percent
since the base period; similarly, an index of 90
means a 10-percent decrease. The index can be
converted to any desired base period for which the
index is available. This is done by dividing each
index number to be converted by the index for the
desired base period. Tables of conversion factors
are provided on request for most series, enabling
users to convert indexes for prior periods on other
bases to the current 1957-59 base. Since the 195759 base was adopted in 1962, some indexes have
been continued on the bases of 1947-49=100 and
1939 = 100. These are calculated by the applica­
tion of appropriate conversion factors to indexes
computed on the 1957-59 base.
Imputation Procedures . Although prices are not

obtained in all 56 cities every month (see table 1
for pricing cycle), it is necessary to represent all
56 cities in each monthly index computation. Be­
tween quarterly survey dates, for every item ex­
cept new automobiles, the weights are held at the
level of their last pricing, which in effect means
prices are estimated unchanged. For new auto­
mobiles, a price change is imputed to the unpriced
cities on the basis of changes in cities surveyed
every month.
For food and apparel items which are sold only
at certain seasons of the year, the index calcula­
tion is made in the off-season as if prices of these
items changed proportionally with prices of items
of a similar nature which are available all year.
For example, prices for strawberries when not in
season are carried forward on the basis of changes
in prices of all other fresh fruits. When the item
returns to the market the current price is, in effect,
compared with the estimated price implicit in the
procedure described.

Average Prices . In the calculation of average

food prices for publication, the prices used in the
index are given special editing since they are not
necessarily restricted to a single specified quality
and size. Procedures have been devised to calcu­
late city and U.S. prices for publication which use
index values and price relatives extensively.
These procedures employ benchmark prices for
defined specifications for each of the 56 cities, in
which quotations not meeting the specified quality
are excluded. Benchmark prices are computed in
an independent operation, pooling prices for all
outlets rather than as an average of average prices
for the two subsamples. The benchmark prices
are then adjusted month by month by the price
changes reflected in the index. The first bench­
mark calculation was for April 1964, from which
date prices were estimated back to December 1963
and forward to December 1964. New benchmark
calculations are planned as of each January. City
prices are combined to U.S. averages by the use of
the 1960 index population weights.21
Average bills for specified quantities of gas and
electricity and average prices of fuel oil, which are
published for the largest cities, are the same as
those used in the index calculation. Since these
are for identical quantities and qualities from
month to month, no special editing is required.
Item Indexes . Indexes for selected items and
groups (commonly referred to as item indexes),
were published semiannually during 1964 and
1965. Quarterly publication was resumed in 1966.
Although the published indexes refer to March,
June, September, and December, some prices for
earlier months must be used in their calculation for
cities not priced in these months. In June, for
example, the indexes are based on (a) June prices
in the 5 cities surveyed every month and the 17
cities surveyed in June on a quarterly cycle; and
(b) April and May prices in the 17 cities surveyed
in each of those months on a quarterly cycle, car­
ried forward to June as in the composite index
20 The index base was 1947-49 = 100 from 1953 through 1961,
1935-39 = 100 from 1935 through 1952, and 1913 = 100 from
1913 through 1934.
21 For a more detailed discussion, see article by Doris P.
Roth well, “ Calculation of Average Retail Food Prices,” Monthly
Labor Review, January 1965, pp. 61-66.


Analysis and Presentation
The CPI is made available first at a press con­
ference, usually held near the end of the month
following that to which the data relate. On the
same day, the press release is mailed to a list of sub­
scribers who have immediate need for the data.
This release contains a brief description of price
changes during the month and several tables of
major group and subgroup indexes and percentage
changes from selected dates, for the U.S. city aver­
age and selected large metropolitan areas. It in­
cludes seasonally adjusted indexes for selected
components. A report containing the same text
with some additional tables is published about two
weeks after the date of the press release. In addi­
tion each of the Bureau’s six regional offices pre­
pares and mails a press release for each of the cities
in its region for which CPI figures are published.
These releases are timed to coincide with the na­
tional release. Other monthly reports contain
average prices of selected foods and fuels in the
largest metropolitan areas. A quarterly publi­
cation presents the U.S. city average indexes for
individual goods and services.
The CPI for the United States and for selected
areas is published also in the Monthly Labor R e­
view in the issue dated two months later than the
index. The annual Statistical Supplement to the
Monthly Labor Review contains indexes for indi­
vidual goods and services (item indexes) as well
as the relative importance of the items in the total
index as of December.
Average prices for foods and fuels are published
in Estim ated Retail Food Prices by Cities and
Retail Prices and Indexes of Fuels and Electricity .

Uses of the Index
The most widespread use of the CPI is in wage
adjustments and collective bargaining negotia­
tions. Although this was the primary reason for
its beginning, use of the CPI for this purpose
declined during the post-World War I and de­
pression periods. Its use in this way was revived
during World War II, but escalation by the index
did not receive widespread acceptance until the
principle was written into a contract between the
United Automobile, Aircraft, and Agricultural


Implement Workers of America and the General
Motors Corporation in 1948. The number of
workers covered by such contracts in 1965 was
about 2 million. However, movements of the
index have an indirect effect on wages and salaries
of many more workers.22
The CPI is used extensively to measure changes
in purchasing power of the consumer dollar. It
is the basis for most estimates of changes in real
earnings of labor, and for comparison with pro­
ductivity measures. Changes in purchasing power
are used for such diverse purposes as adjusting
royalties, pensions of government and non-govern­
ment workers, welfare payments, rental contracts,
and occasionally alimony payments.
One of the most important uses of the index is
as a guide to broad economic policy. It is one of
the most widely used measures of inflationary
pressures. During wartime periods the index and
its components have served an important adminis­
trative function in connection with determination
of policies concerning price control and subsidies.
In peacetime the index and its underlying sta­
tistics have played an important part in the gov­
ernment’s effort to maintain stable wage-price
relationships and to judge the advisability of mak­
ing monetary or tax adjustments. It is one of the
chief statistical tools for conversion of the national
accounts to constant dollars.

Limitations of the Index
The CPI is not an exact measure of price
changes. It is subject to sampling errors which
may cause it to deviate somewhat from the results
which would be obtained if actual records of all
retail purchases by wage earners and clerical
workers could be used to compile the index. These
estimating or sampling errors are limitations upon
the precise accuracy of the index rather than mis­
takes in the index calculation. The accuracy could
be increased by using much larger samples, but
the cost is prohibitive. Furthermore, the index is
believed to be sufficiently accurate for most of the
practical uses made of it. With the changes in
sampling techniques introduced in 1964, the Bu22 See article by Francis S. Cunningham, “ The Use of Price
Indexes in Escalator Contracts,” Monthly Labor Review, August
1963, pp. 948-952. Reprint 2424.


T a b l e 5.
D e f in it io n o f t h e I n d e x

S ummary of C h a r a c te r istic s o f th e CPI B e g in n in g 1964

Title.................... ................................

Consumer Price Index—U.S. City Average for Urban Wage Earners and Clerical Workers.

Formula (Simplified expression).


Base period.............................................
Duration_____ _____ ______ ______
Definition of index expenditure weights.

1957-59=100. Series also published on 1947-49 and 1939 bases.
January 1964 forward.
Average expenditures for urban wage-earner and clerical-worker consumers (including single workers)
derived from the 1960-61 Consumer Expenditure Survey in 66 urban places, adjusted for price changes
between the survey dates and 1963 except for 6 cities added in 1966.

(& )

P o p u la t io n C o v e r a g e o f
E x p e n d it u r e Su r v e y

Place of residence_________________
Family size_______ ____________ I_.
Occupation...... .................................
Length of employment____________
Income................................................ .
C it y

Urban places of 2500 or more in 1960; including Alaska and Hawaii.
No restriction; single consumer units included.
Wage-earner and clerical-worker families and single individuals living alone. (More than half of tota Ifamily
income from wage-earner and clerical-worker occupations.)
At least 1 family member or single consumer unit must have been employed for 37 weeks or more during the
survey year in wage-earner or clerical-worker occupations.
No criterion as to income except the qualification above.

C o verage

Population weights................... ............
Sample of priced cities........ ..................
Published indexes. .................. .............

Sa m p l e

Basis of sample selection.......................
Basis for allocation to priced items.......
Commodity coverage........... ............... .
Number of items priced.................. .....
Pricing cycle. ....................... .................

R epo rter

Based on 1960 Population Census; Alaska and Hawaii included. Proportion of population in wage-earner
and clerical-worker group covered by index was based upon BLS expenditure surveys.
50 metropolitan areas and cities selected originally to represent all urban places in the U.S. including Alaska
and Hawaii with populations of 2500 or more in 1960. Six additional areas added in 1966.
U.S. and 17 large metropolitan areas for families and single consumer units combined. Indexes for six more
large metropolitan areas available in the latter part of 1965.

C o verage

Number of reporters___________ I.
Number of quotations___________
Pricing technique..............................

Probability proportionate to importance in family spending.
Expenditures classified into 52 expenditure classes. Certainty items assigned their own importance; re­
mainder of expenditures assigned equally to probability selections within expenditure classes.
Goods and services purchased for family living, including necessities and luxuries; excluding personal in­
surance, income and personal property taxes but including real estate taxes and sales and excise taxes.
About 400 represented in U.S. index and published city indexes. Certainty items priced in all unpublished
cities; other items in 1 of 2 subsamples of unpublished cities.
Prices of foods, fuels and a few other items priced monthly in all cities, except for San Diego and Milwaukee
where all items are priced quarterly.
Prices of most other commodities and services priced monthly in the 5 largest cities, and quarterly in re­
maining cities.
In central cities and selected suburbs of 56 metropolitan areas (50 areas in 1964 and 1965).
About 1,775 food stores (1,525 for 50 areas), 40,000 tenants (34,000 for 50 areas), 16,000 other reporters of all
kinds (15,000 for 50 areas).
Over 1 million food prices per year; about 80,000 rent charges per year (68,000 for 50 areas); about 375,000
quotations per year for items other than food and rent (350,000 for 50 areas).
Personal visit of BLS agent except for a few items collected by mail or from secondary sources.
Specification pricing but agent is permitted to price deviations from specification under prescribed con

reau is attempting to measure the sampling error
in the index.23
Another kind of error occurs because people who
give information do not always report accurately.
The Bureau makes every effort to keep these errors
to a minimum, obtaining prices wherever possible
by personal observation, and corrects errors when­
ever they are discovered subsequently. Precau­
tions are taken to guard against errors in pricing,
which would affect the index most seriously. The
field representatives who collect the price data and
the commodity specialists and clerks who process
88 Preliminary estimates of sampling error were computed
and published in Measurement of Sampling Error in the Con­
sumer Price Index: F irst Results, by Marvin Wilkerson, paper
presented at American Statistical Association meetings, Decem­
ber 29, 1964. Additional estimates will be made available as
work continues on this project.

them are well trained to watch for unusual devia­
tions in prices which might be due to errors in
The CPI represents the average movement of
prices for urban wage earners and clerical workers
as a broad group, but not the change in prices paid
by any one family or small group of families. The
index is not directly applicable to any other occu­
pational group or to non-urban workers. Some
families may find their outlays changing because
of changes in factors other than prices, such as
family composition. The index measures only the
change in prices and none of the other factors
which affect family living expenses.
In many instances, changes in quoted prices are
accompanied by changes in the quality of con­
sumer goods and services. Also new products are


introduced frequently which bear little resem­
blance to products previously on the market;
hence, direct price comparisons cannot be made.
Quoted prices are adjusted for changes in quality,
whenever necessary data are available. Technical
specifications and highly trained personnel are re­
lied on to insure comparability of quality of items
compared from period to period.24 Nevertheless,
some residual effects of quality changes on quoted
prices undoubtedly do affect the movement of the
CPI either downward or upward from time to
Another important limitation of the index is
that it measures only time-to-time price change in
a given area. City indexes do not show intercity
differences in either prices or living costs. They

show only differences in rates of price change from
one time to another. Other types of measures are
required to show place-to-place differences in liv­
ing costs. The most recent such measure is “The
Interim City Worker’s Family Budget” which
shows the estimated dollar costs of a “modest but
adequate” level of living in 20 large cities and their
suburbs in the fall of 1959, which is described in
Chapter 9.
MHoover, Ethel D., “ The CPI and Problems of Quality
Change,” Monthly Labor Review, November 1961, pp. 1175-1185.
Reprint 2378, and Larsgaard, Olga A., and Louise J . Mack,
“ Compact Cars in the Consumer Price Index,” Monthly Labor
Review, May 1961, pp. 519-523. Reprint 2368.
25 See testimony of Ewan Clague in Hearings before the Sub­
committee on Economic Statistics, Joint Economic Committee,
Congress of the United States, 87/1, Part II, p. 588, Washing­
ton, May 1, 2, 3, 4, and 5, 1961.

Technical References

1. Cunningham, Francis S. “ The Use of Price Indexes in Escalator Contracts,” M onthly L ab or Review ,
August 1963, pp. 948-952. Reprint No. 2424.
Discusses the techniques of escalation using the two major price indexes published by the
Bureau of Labor Statistics— the Consumer Price Index (CPI) and the Wholesale Price Index
(WPI). Examines the basic elements of an escalator clause and procedures for carrying out
the agreement.
2. Daugherty, James C. “ Health Insurance in the Revised CPI,” M onthly L ab or Review , November
1964, pp. 1299-1300.
Explains and justifies the major change in the treatment of the health insurance component
of medical care as initiated in the recent revision of the Consumer Price Index. Compares the
former method of pricing actual premium rates with the new method of pricing the benefits
received for hospital and professional services combined with a measurement for retained earnings.
3. Hoover, Ethel D. “ The CPI and Problems of Quality Change,” M onthly L ab or Review , November
1961, pp. 1175-1185.
Explains and illustrates problems of quality measurement met in the index calculation
procedures. Defines quality as used by the BLS, specification pricing, direct price compari­
sons, and linking procedures. Concludes that there is no evidence to support the argument
that the index is not a true measure of price change because of not fully eliminating the effect
of quality changes.
4. Humes, Helen and Schiro, Bruno. “ The Rent Component of the Consumers’ Price Index.
Part I— Concept and Measurement,” M onthly L ab or Review , December 1948, pp. 631-637.
“ The Rent Component of the Consumers’ Price Index. Part II— Methodology of Measure­
ment,” M onthly L ab or Review , January 1949, pp. 60-68. Combined in Reprint No. 1947.
Part I discusses the basic concepts underlying the rent index. Part II explains the methods
of obtaining and calculating rental data, the “ new unit bias” which existed during World War II
and the problem of compensating for depreciation of quality caused by aging.
5. Jaffe, Sidney A. “ The Statistical Structure of the Revised CPI,” M onthly L ab o r Review , August
1964, pp. 916-924.
Describes the concept and formulation, population and expenditure coverage, statistical
techniques and problems of the revised index. Examines some operational aspects, especially
sample replication. Presents the index formula in general, simplified, and in operational form.
6 . ----- . The Consum er P ric e In dex— Technical Questions an d P rac tic al A n sw ers . Paper presented
before the American Statistical Association Meeting, Washington, D.C., December 30, 1959
(Mimeographed by the U.S. Department of Labor, Bureau of Labor Statistics). 27 pp.
General explanation and definition of the index concepts, coverage and calculation. Dis­
cussion of problem areas including sampling, seasonality, and alleged quality bias.




Technical References—Continued

7. Lamale, Helen Humes. “ Housing Costs in the Consumer Price Index, Part I,” M onthly L ab or
Review, February 1956, pp. 189-196. “ Housing Costs in the Consumer Price Index. Part II,
M onthly L ab or Review , April 1956, pp. 442-446. Combined in Reprint No. 2188.
Part I defines the housing component of the index and describes the derivation of expendi­
ture weights used in the calculation of the shelter index. Part II describes the procedures used
to measure changes in the prices of the various items of shelter cost.
8. Larsgaard, Olga A. and Mack, Louise J. “ Compact Cars in the Consumer Price Index,” M onthly
L ab or Review, May 1961, pp. 519-523. Reprint No. 2368.
Summarizes and explains the methodology used to link compact cars into the Consumer
Price Index in 1961. Discusses the historical treatment of quality changes in standard size
9. National Bureau of Economic Research. The P rice S ta tistic s o f the F ed eral Government: Review,
A p p ra is a l, an d Recom m endations, (Washington, National Bureau of Economic Research, General
Series, Number 73, 1961), 496 pp. Also appears in Government P rice S ta tistic s: H earin gs,
Subcommittee on Economic Statistics of the Joint Economic Committee, 87th Cong., 1st sess.,
Part 1, January 24, 1961, ([Washington,] U.S. Government Printing Office, 1961), 526 pp.
Report of the detailed investigation by the Price Statistics Review Committee of the N BER
in 1959 of the main price indexes compiled by the Federal Government: The Consumer Price
Index; the Wholesale Price Index; and the Indexes of Prices Received and Paid by Farmers.
Reviews and analyzes the various aspects of the indexes and presents general and specific recom­
mendations for improvements. Twelve staff reports appended.
10. Rothwell, Doris P. “ Calculation of Average Retail Food Prices,” M onthly L ab or Review, January
1965, pp. 61-66.
Explains the BLS methods of collecting prices, and computing indexes and avreage prices
for food items in the index. Emphasizes the unsuitable nature of index data for comparison
of prices between cities. Presents estimated retail prices of food from December 1963 through
November 1964, the cities covered, and the pricing diagram for food in the index.
11. ----- . The Consum er P rice In d ex: P ric in g an d C alculation Procedures, (Unnumbered paper, U.S.
Department of Labor, Bureau of Labor Statistics, March 1964), 22 pp.
Discusses sampling, pricing by specification; price collection, processing, and editing.
Describes index formula, calculation of price relatives, expenditure weights, indexes, aggregation,
and correction policy.
12. ----- . “ Use of Varying Seasonal Weights in Price Index Construction,” J o u r n a l of the A m erican
S ta tistic a l A ssociation , March 1958, pp. 66-77.
Describes a formula based on varying seasonal weights for month-to-month measurements
of price change which does not exhibit the “ biases” of chain indexes and which satisfies classical
index theory with respect to year-to-year comparisons. Results of experimentation with
alternative formulas are presented.
13. U.S. Congress, House of Representatives, Committee on Education and Labor. Consum ers'
P rice In dex, Special subcommittee of the Committee on Education and Labor, 82nd Cong.,
1st sess., Report No. 2 (1951), 39 pp.
Nontechnical summary of results of hearings on the reliability of the Consumer Price
Index. Presents details of history, uses, and method of construction of the index. Recommends
continued support of the index by the Congress.
14. U.S. Congress, Joint Economic Committee. Government P rice S ta tistic s: H earin gs, Subcommittee
on Economic Statistics, 87th Cong., 1st sess., Part I, January 24, 1961 (Washington, U.S. Gov­
ernment Printing Office, 1961), 526 pp. Part 2, May 1, 2, 3, 4, and 5, 1961. (1961), 265 pp.
Part 1 presents findings of an investigation by the Price Statistics Review Committee of
the National Bureau of Economic Research in 1959-60 of all government price statistics. Also
includes 12 staff papers on specific subjects. The detailed technical report includes recommenda­
tions for improvement of all indexes and, specifically for the Consumer Price Index, suggests
extended coverage to include single consumers, probability sampling techniques, establishment
of a research division, and regularly scheduled weight revisions. Part 2 presents testimony
before the subcommittee of members of the Price Statistics Review Committee, government
officials, and other interested parties concerning the committee report and recommendations.


Technical References—Continued

15. U.S. Congress, Joint Committee on the Economic Report.

The C onsum ers’ P ric e In dex— Report
of the J o in t Committee on the Econom ic Report on the C onsum ers’ P rice In dex o f the United States
B u re au of L ab or S ta tistic s, 80th Cong., 2nd sess. (1949), 20 pp.








Brief statement of the results of examinations of methodology, compilation, composition,
and presentation of the Consumer Price Index as of 1949. Extensive bibliography.
U.S. Department of Labor, Bureau of Labor Statistics. The Consum er P rice In dex (.Revised J a n u a r y
196If): A Short D escription , (U.S. Department of Labor, Bureau of Labor Statistics, September
A nontechnical description of the index, its scope and computation. Explains the market
basket, formula, uses and limitations of the index. Tables show cities included, population
weights, pricing schedules, groups of goods and services priced, their relative importance, and
the number of items priced as of December 1963.
----- . S easo n al F actors, Consum er P rice In d e x : Selected Series, J u n e 1 9 6 3 - M a y 1961, Bulletin 1366
(1963), 47 pp.
Provides basic data with which Consumer Price Index old series indexes can be adjusted
for seasonal variation. Users are cautioned that the 1964 revision may have a very different
effect on the series. Includes a description of the BLS method of computing seasonal factors,
a discussion of its application to consumer price series, comments on specific series and tables
providing indexes and seasonal factors for 66 selected series through May 1961.
----- . “ Taxes and the Consumers’ Price Index,” M onthly Labor Review, January 1953, pp. 53-57.
Reprint No. 2090.
Discusses the present treatment of taxes in the index and the specific taxes included. Jus­
tifies the BLS policy of continuing to exclude income taxes from the index and including sales and
excise taxes.
----- . In terim A djustm ent o f Consum ers’ P ric e In d ex: Correction o f New U nit B i a s in Rent Component
of Consum ers’ P rice In dex an d Relative Im portance of Item s, Bulletin 1039 (1952), 49 pp.
Military developments in Korea in 1950 emphasized and made urgent the need for reweight­
ing of certain segments of the index before the already initiated revision could be completed in
1952. The failure to reflect the difference between rents for new dwellings when they first enter
the market and comparable dwellings already on the market during and after the Second World
War is discussed and the method of adjustment presented. Tabulation of adjusted indexes from
1940-50, relative importances and weights generated by the interim adjustment also are presented.
----- . Consum er P ric e s in the United States, 1 9 5 3 -6 8 : P rice T rends an d In dexes, Bulletin 1256,
(1959), 126 pp.
Analyzes and explains retail price trends and their effect on the economy from 1953 to 1958.
Brief history of the index and comparison of features of the old index based on the 1934-36
expenditures survey, with the adjusted index based on the 1947-49 expenditures survey in 7
cities and the revised index based on the 1950 expenditures survey. Explains the 1952 revision
in detail. Presents historical indexes for various segments of the index.
U.S. Office of Economic Stabilization. Report o f the P resid en t’s Committee on the Cost of L ivin g,
(1945), 423 pp.
Summarizes the findings of the investigation in 1943-44 of the suitability of the Consumer
Price Index for measurement of the change in the cost of living during wartime. Includes
detailed discussions of the definition, scope, and statistical methodology of the index.
Wilkerson, Marvin. M easurem ent of S a m p lin g E rro r in the Consum er P ric e In d e x: F ir st R esults.
Paper presented at the American Statistical Association meeting, Washington, D.C., December
1964 (mimeographed by the U.S. Department of Labor, Bureau of Labor Statistics), 16 pp.
With the computation of the Revised Consumer Price Index, completed in 1964, a first
attempt was made to produce estimates of error for a comprehensive national price index. Pre­
sents a brief description of the replication design, estimates of error through October 1964. Dis­
cusses limitations of the error estimates and provides an interpretation of the results.



Technical References—Continued

23. Wilkerson, Marvin. S a m p lin g A spects of the Revised C P I (Unnumbered paper, U.S. Depart­
ment of Labor, Bureau of Labor Statistics, October 1964), 33 pp.
Explains two significant improvements in methodology made in the Consumer Price Index
at the time of the most recent revision: (1) application of probability sampling, and (2) estima­
tion of sampling error through a system of replicated samples. Tables present the sampling
frame for selection of the index item sample and pairing of index cities for replication computa­
24. ----- . “ The Revised City Sample for the Consumer Price Index,” M onthly L ab o r Review , October
1960, pp. 1078-1083. Reprint No. 2352.
Describes the selection procedures used to derive the core sample of 50 cities used in the
revised Consumer Price Index computation. Tables show probability patterns for the selected
areas and basic and alternate city samples for the Consumer Price Index and the Consumer
Expenditures Survey.

— D o r is


R o th w ell and C a r l y l e


S t a l l in g s

Office of Prices and Living Conditions

Chapter 11. Wholesale Prices
The Wholesale Price Index (WPI) is the oldest
continuous statistical series published by the Bu­
reau of Labor Statistics (B LS) and one of the
oldest in the Federal Government. It was first
published in 1902, and covered the years 18901901. The origins of the index are associated with
a resolution of the United States Senate in 1891,
which authorized the Senate Committee on
Finance to investigate the effects of the tariff laws
“upon the imports and exports, the growth, devel­
opment, production, and prices of agricultural and
manufactured articles at home and abroad.” 1
The index published in 1902 on the base 189099 was an unweighted average of price relatives
and included from 250 to 261 commodities. Since
that time, many changes have been made in the
sample of commodities, the base period, and in the
method of calculating the index. The first major
change was completed at the end of 1914, when a
system of weighting was introduced and the in­
dex was recalculated back through 1890.2 By 1940,
the number of commodities had increased to ap­
proximately 900, based on about 2,000 individual
price quotations. Then, in 1952, the most exten­
sive revision in the history of the index was com­
pleted.3 The number of commodities and quota­
tions was doubled, weights were based on 1947
Censuses and changes were made in the calculation
method. Some changes in classification were made
also, including expansion to the present 15 major
By January 1965, the number of commodities
had increased to more than 2,200, the number of
price quotations had increased to almost 6,800, and
the index had become increasingly representative
of general primary market price changes.

Description of Survey
Throughout its history, the Wholesale Price In­
dex has been a measure of price changes for goods
sold in primary markets in the United States.
“Wholesale” as used in the title of the index refers
222-617— 66----- 7

to sales in large quantities, not prices received by
wholesalers, jobbers, or distributors.
From its inception, the index has been considered
a general purpose index designed to measure the
general price level in other than retail markets.
From the beginning of the index, however, atten­
tion was directed to some specific needs of users,
and indexes for individual commodities and for
major commodity groups were published. As
early as 1903, two special group indexes by stage
of processing—Baw Commodities and Manufac­
tured Commodities—were published “to meet the
wishes of students of price statistics.” In recent
years, emphasis has been placed on the develop­
ment of more subdivisions within major groups
and special combinations of indexes such as by
Stage of Processing and Durability of Product.
Most of the quotations reported to the Bureau
are the selling prices of representative manufac­
turers or producers, but some prices are those
quoted on organized exchanges (spot prices) or at
central markets. Prices for imported commodities
are those received by importers—the first commer­
cial transaction in the United States. Since the
index is intended to measure “pure” price change,
that is, not influenced by changes in quality,
quantity, shipping terms, product mix, etc., com­
modities included in the index are defined by pre­
cise specifications which incorporate the principal
price-determining characteristics of the com1 Wholesale Prices, Wages, and Transportation, Senate Report
No. 1394, “ The Aldrich Report,” Senate Committee on Finance,
Congress of the United States, March 3, 1893, Part I, (52nd
Cong., 2nd sess.), Government Printing Office 1893; and Course
of Wholesale Prices, 1890-1901, Bulletin of the Department of
Labor, No. 39, March 1902, pp. 205—209.
The Aldrich Report included 230 series of price quotations com­
bined to an All Commodities index on the base 1860 covering the
period 1840-1891, but not all series of quotations covered the
entire period. This study was brought up to date in 1900 by the
Department of Labor, under the supervision of Roland P. Falkner,
Professor of Statistics of the Wharton School of Finance and
Economy, University of Pennsylvania. The index was an un­
weighted average of 99 price relatives, computed quarterly,
covering the period from 1890 through 1899 and on the base
January 1890-January 1892 = 100.
2 See also, Allan D. Searle, “Weight Revisions in the Wholesale
Price Index, 1890-1960,” Monthly Labor Review, February 1962,
pp. 175-182.
3A large number of the newly introduced commodity prices
were carried back to 1947. The presently published index con­
tains the new commodities for the period 1947—51 and displaces
the older less comprehensive index on the 1926 base published
for the same period.





transaction prices. Companies are requested to
report prices less all discounts, allowances, rebates,
free deals, etc., so that the resulting net price is the
actual selling price of the commodity for the
specified basis of quotation. The Bureau periodi­
cally emphasizes to reporters the need to take into
account all discounts and allowances. However,
list or book prices are used if transaction prices are
Prices are generally f.o.b. production or central
marketing point to avoid reflection of changes in
transportation costs. Delivered prices are in­
cluded only when the customary practice of the
industry is to quote on this basis and the Bureau
cannot obtain a price at the production point.
Subsidies to the producer and excise taxes are ex­
cluded since they are not considered part of the
price, but import duties are included as part of the
selling price of imported goods.
Although the same commodity is priced gen­
erally month after month, it is necessary to provide
a means for bridging over changes in detailed
specifications (or descriptions of items priced) so
that only real price change will be measured. An
adjustment is particularly important when new
commodities are introduced, but even when specifi­
cations of existing commodities are changed, care
is exercised to help insure that only price changes
influence the index. A new price series resulting
from a physical change in an article or a change in
selling terms is substituted for the earlier series by
direct comparison or by linking. The objective of
the linking procedure is to insure that the index
will reflect only those changes which are due to
actual price differences.6*8 Each time a change in

To the extent possible, the prices used in con­
structing the index are those which apply to the
first significant commercial transaction in the
United States. Transactions for the same item at
later stages of distribution are not included.
However, as raw materials are transformed into
semifinished and finished goods, the resulting prod­
ucts are represented.
With some exceptions, the prices refer to one
particular day of each month. In most cases, the
pricing date is Tuesday of the week containing the
15th; but for some commodities (farm products,
particularly) a day other than Tuesday is used
because it is considered more representative.5
The Bureau attempts to base the WPI on actual

4An example of a commodity specification for steel strip is :
“ Strip, cold-rolled, carbon steel, coils, No. 4 temper, No. 2
finish, No. 3 edge, base chemistry, 6" x .050", in quantities of
10,000 to 19,999 lb., mill to user, f.o.b. mill, per 100 lb.”
6The prices used in the index through 1951 were the simple
arithmetic averages of prices for all Tuesdays in the month.
8 The following example illustrates the linking procedure: The
September price for a certain machine used in the calculation of
the index was $2,347.50. In October, a new model of the machine
was introduced, priced at $2,562.60. The new model was con­
sidered essentially comparable with the old, except that it had
a more powerful motor and larger tires. These were valued at
$186.20 more than the value of those used on the former model.
For linking, the September price of the new model was estimated
at $2,533.70 ($2,347.50 September price of former model plus
$186.20 increase in value of motor and tires). The price com­
parison between September and October was based on the esti­
mated September price of $2,533.70 and the reported October
price of $2,562.60. Thus a 1.1 percent price increase was reflected
in the October index while the price change due to quality im­
provement (more powerful motor and larger tires) was not

modifies.4 So far as possible, prices are f.o.b.
production point, and refer to sales for immediate
delivery. Prices applicable to long-run contracts
are not included, nor are “futures.”

The Wholesale Price Index universe consists of
all commodities sold in commercial transactions in
primary markets of the United States, including
Alaska and Hawaii. Commodities produced in
the United States are included, as well as those
imported for sale. The universe covers manufac­
tured and processed goods and the output of indus­
tries classified as manufacturing, agriculture,
forestry, fishing, mining, gas and electricity, public
utilities, and goods competitive with those made in
the producing sector of the economy, such as waste
and scrap materials. All systematic production is
represented, but individually priced items, such as
works of art, are excluded. Also excluded are
goods transferred between establishments owned
by the same company (interplant or intra-com­
pany transfers). Goods sold at retail by pro­
ducer-owned retail establishments also are ex­
cluded because they conceptually belong to a retail
(consumers’) universe, rather than to primary
market transactions.
Civilian goods normally purchased by the Gov­
ernment are in the universe, but military goods are
not. Government sales of some commodities (e.g.,
electric power) are included if they can be con­
sidered competitive with free market sales.


the item priced occurs, the Bureau appraises the
significance of the specification change to ascertain
whether an actual price change occurred. I f the
specification change is minor and does not involve
price-making factors, the substitution is effected
by direct comparison, and any reported price
change between the old and the new specification is
reflected in the index. I f changes in specification
are major, and if either no real price change oc­
curred or no information can be obtained con­
cerning the value of the difference in specification
(perhaps indicative of a change in quality), the
substitution is made by linking and no change is
reflected in the index. In this case, any reported
difference in price level is not permitted to affect
the index level.
When differences are major, an attempt is made
to obtain data from the reporters on the value of
the additional (or deleted) features and to adjust
the price index accordingly. This is particularly
important in the case of some durable goods, such
as automobiles, which have periodic model
changes. Also, price increases which result from
the addition of features which formerly sold at
extra cost are not reflected in the index. Con­
versely, price changes attributable to deletion of
equipment which was formerly standard are not
treated as decreases.
In the event production of a specified commodity
is discontinued by a reporter, or its importance is
reduced, the Bureau collects price data for a simi­
lar or a replacement item. Prices are obtained for
the new and the discontinued series for a 1-month
overlap period. The index is extended by linking,
and the difference, if any, between the new item
price and the original price is taken as a measure
of the quality difference between the two items.
Linking is also used for the addition to or dele­
tion of commodities or groups of commodities
from the index and, the addition to or deletion
of a company report from the sample of companies
priced or on occasion a change in the source of
price. Whenever a new commodity is added to
an existing commodity group, linking of the new
item to any one of the existing items is not per­
tinent. Instead, the weights of the entire group
are redistributed to include the new item and the
link is made at the group level instead of at the
commodity level. A similar procedure is used to
handle items which drop out of the index.


Prices for individual commodities reported by
the individual companies are averaged (usually by
means of an unweighted average). Month-tomonth price change should be computed from
matched-company data. In order that a change
in the company-reporter sample itself not affect
the measure of percentage change, the change is
calculated for any 2 months from identical-com­
pany data. Thus, a new report affects the index
no earlier than the second month.

The classification system of the Wholesale Price
Index follows commodity lines. Products are
grouped by similarity of end-use or material
composition, rather than by industry of origin.
The WPI classification does not match the Stand­
ard Industrial Classification (SIC ), the Standard
Commodity Classification, the United Nations
Standard International Trade Classification
(SITC ), or any other standard classification.
Historical continuity and the needs of index users
have been important in developing the classifica­
tion. No single classification plan can meet all
of the requirements for wholesale price statistics,
but the plan adopted should be flexible enough
to facilitate regrouping of price series to make
special grouping indexes. In January 1965, the
index was made up of 15 major groups, 90 sub­
groups, 314 product classes, 37 subproduct classes,
and 2,213 items.7
To meet the needs of index users, a number of
special group indexes are calculated and published
each month. Among these are indexes by Stage
of Processing,8 Indexes by Durability of Product,
and Indexes of Construction Materials. In ad­
dition, there are about 25 other special group
Except for the Stage of Processing Indexes,
these special groupings consist of rearrangements
of the WPI data into different combinations of
price series, so that the appropriate prices and
weights are those of the WPI. The Stage of
Processing Indexes, however, regroup each item*
7See table for the major groups and subgroups Included in
the WPI.
*The broad stages of processing are: Crude materials for
further processing; Intermediate materials, supplies, and com­
ponents; and Finished goods. Each of these is subdivided



Wh o lesa le
R e l a t iv e I m p o r t a n c e


N um ber


I tem s

Decem­ Decem­
ber 1964 ber 1960

I ndex
fo r

M a jo r G r o u p s

January 1965

January 1965

quota­ Items

January 1965

quota­ Items

quota­ Items










Farm products
Fresh and dried fruits and vegetables_____
Livestock and live poultry


















processed foods














Textile products and apparel
Cotton products _




































Cereal and bakery products_____________
Meats, poultry, and fish
Hairy products and ice cream
. .... . _
Canned and frozen fruits and vegetables___
Sugar and confectionery
Packaged beverage materials
Fdibl(Tfats and oils i_ _
Miscellaneous processed foods



Wool products_______________________
Manmade fiber textile products__________
Silk products________________________
Apparel_______ _____________________
Miscellaneous textile products___________
Plastic honsefnmishings l

_ _


Hides, skins, leather, and leather products—


Fuels and related products, and power.........
Coal....... „ ...................... .............................
Coke___ ___
.. . _ _ _ _ _____


Chemicals and allied products . . .
Industrial chemicals
Paint and paint materials i _
Drugs and pharmaceuticals
__ _
F a ts and oils, inedible
. _ .__ _ .
Mixed fertilizer. i zer m ateri a,ls __
Other chemicals and allied products







Rubber and rubber products
Crude rubber. _
Tires and tubes. _ _
. _
Miscellaneous rubber products __





Dumber and wood products
Mill work
.... ‘




P ulp, paper, and allied products
W oodpnlp




M etals and metal products. .
Tron and steel
.... . _ ................... _.
NTon ferrous metals
__ .
M etal containers

H ides and skins
le ath er

Footwear __________________________
Other leather products_________________

Gas fuels. _ _
Electric power __
_ .
Crude petroleum i _
Petroleum products, refined.

__ _ _ _

_ _.
Paperboard____________ ___ _______ _
Converted paper and paperboard products..
Building paper and board


Plum bing fixtures and brass fittings
Heating equipment
Fabricated structural metal products

Fabricated nonstructural metal products__
See footnote at end of table.


S ubgro u ps

January 1965

All commodities............................................

P lant arid animal fibers...
_ ..... _
Fluid milk
__ _
F ggs
........... _
Hay, hayseeds, and oilseeds
Other farm products ...


Number of items and price quotations

in total 1958


P r ic e

R e p o r t in g S o u r c e


1.731 1
























341 1,161
45 1 144


144 1______



January 1965

quota­ Items































W h o l e s a l e P r i c e I n d e x — Continued
R e l a t iv e I m p o r t a n c e


N um ber



I tem s


in total 1958
Decem­ Decem­
ber 1964 ber 1960


R e p o r t in g S o u r c e

fo r

January 1965

January 1965

quota­ Items














Furniture and other household durables__
Household furniture............................ .
Commercial furniture.......... ................
Floor coverings......__........... ........ .......
Household appliances..... .....................
Television, radio receivers, and phono­
graphs_____________________ ______
Other household durable goods_____ ___ _








Nonmetallic mineral products.................
Flat glass.................................... .......
Concrete ingredients_________ _________
Concrete products____________ ________
Structural clay products....................... .
Gypsum products................... .............
Asphalt roofing.................. ........ ........
Other nonmetallic minerals................. .








Tobacco products and bottled beverages__
Tobacco products--............................
Alcoholic beverages.............................
Nonalcoholic beverages..........................








Miscellaneous products..........................
Toys, sporting goods, small arms, etc_____
Manufactured animal feeds....... ....... ......
Notions and accessories________________
Jewelry,* watches, and photographic equip­
ment.................... ....... ...............
Other miscellaneous products___ _______



















1 Subgroup index not published.
N ote: Relative importance represents the basic value weight of an item
or items multiplied by the relative of price change between the weight date
to a later date and the result expressed as a percentage of the total for all

priced in the WPI according to the amount of
processing, manufacturing, or assembling it under­
goes before entering the market. A commodity
may appear in several different categories in this
scheme. Thus, 25 percent of the fresh vegetables
(by value-weight) was assigned to crude food­
stuffs and feed-stuffs for further processing and
75 percent to consumer foods (as “finished” goods).
The value weights are the same as those of the
WPI and the allocations among the stages of
processing are from an inter-industry transaction
study made for the year 1947.9
9 See W. Duane Evans and Marvin Hoffenberg, “The Inter­
industry Relations Study for 1947,” The Review of Economics
and Statistics, May 1952, pp. 97-142; also, National Bureau of
Economic Research, Input-Output Relations; and Appraisal,
Studies in Income and Wealth, vol. 18, New York, N.Y., 1955.









January 1965

quota­ Items




S ubgro u ps

Number of items and price quotations

Machinery and motive products________
Agricultural machinery and equipment___
Construction machinery and equipment__
Metal -working machinery and equipment..
General purpose machinery and equipment.
Miscellaneous machinery___ ____ _____
Special industry machinery and equipment.
Electrical machinery and equipment.........
Motor vehicles____ __________________
Transportation equipment, other than mo­
tor vehicles 1....................................


M a jo r G r o u p s

January 1965

quota­ Items

January 1965

quota­ Items
















commodities. The differences between the relative importance as of Decem­
ber 1960, the date of last major weight change, and that of December 1964
are due principally to price change only.

Data Sources and Collection Methods
Price data are collected by mail questionnaire,
and reporting is voluntary and confidential. Most
prices are collected each month. For a few com­
modities, where price changes are infrequent, the
shuttle schedule is mailed quarterly, but monthly
prices are requested. Generally, the price data
used in the index are obtained directly from the
producing company, but some trade publications
are used when the publication generally is accepted
as reliable by the Bureau and the industry. For
fish and most agricultural products, the prices used



are those collected and published by other Govern­
ment agencies.
Price reporting is initiated, wherever possible,
by personal visit by a Bureau representative to the
prospective respondent. Pricing of additional
products from established reporters often is done
by mail. In any event, a detailed report describ­
ing all of the price-making characteristics of the
commodity is prepared for each new price series.
This commodity price information sheet (B LS
1810) is shown on pages 99 and 100. The form
becomes a part of the permanent record for the
series. After the initial collection of prices,
monthly information is collected by mail by means
of a shuttle schedule (B L S 473, shown on pages
101 and 102).

The price data are combined using weights based
on value of shipments. The major sources of the
value data arre:
Bureau of Census____ C en su s o f M a n u fa c tu re s
C en su s o f M in e ra l In d u s tr ie s

Bureau of Mines_____ Various publications, e.g., M in­
e r a ls Y earbook,

Department of Agricul- Various publications, e.g., A griture.
c u ltu r a l S t a tis t ic s .
Bureau of Fisheries___Various publications, e.g., F is h ­
e r ie s o f the U nited S ta te s .

In addition, many other sources of data, such as
trade associations, are used. Import data are ob­
tained from reports of the U.S. Department of
Commerce, United States Trade with Puerto Rico
and with United States Possessions (No. 800)
and United States Im ports for Consumption (No.
F T 110).

The monthly index is based on a judgment
sample of commodities, a sample of specifications
(descriptions), and a sample of reporters. The
sample of commodities is chosen after a review of
the data of the industrial censuses and other sta­
tistics of value of transactions. Generally, the
commodities chosen are those of the largest ship­
ment values. New items are not added until they
have become established in the market.10 They

are added, normally, in December of any year,
with the first effect in the index in January.
Samples of specifications and of reporters are
selected after consultation with trade associations
or other industry representatives and with staff
of other government agencies. Individual com­
modity specifications are selected also on the basis
of net dollar sales. That is, the “volume seller” of
the industry (not of the company) is preferred.
The specification describes not only the popular
physical characteristics but the most common
quality, grade, level of distribution, and market.
However, terms of sale (discounts, etc.) are based
on the company’s own most common practice. For
some commodities, prices are quoted by producers
and sellers in terms of a single specification taken
as standard; all other prices are quoted as differen­
tials from the standard. This is true for some farm
products such as wheat and cotton. When there
is no standard commodity basis, the specification
to be priced is selected with the help of industry
The number of reporters is determined, to some
extent, by the variation of price movements among
them and the degree of price leadership. When­
ever possible, a minimum of three companies is
obtained, so that data for specified commodities
can be published without disclosure of information
supplied by individual companies. For commodi­
ties with more than one major production area
and a definite regional pattern, a larger sample
is selected. Among these commodities are waste
materials and such building materials as brick,
cement, and stone.
A comparatively small list of properly selected
commodities would produce a reliable index, if
only an All Commodities index were desired.
However, historically there has been great interest
in indexes for groups of commodities and for in­
dividual commodities. To meet these needs, the
Bureau has increased the sample in order to pro­
vide more detailed indexes as well as many spe­
cial-purpose indexes.
10 If new items are added before they become fully established,
the sharp price decline experienced by most products, as they
move from development to mass production, imparts a downward
bias to the index. Also, many new products turn out to be of
only transitory significance.


Estimating Procedures
Formula and Calculation
In concept, the Wholesale Price Index is cal­
culated according to a modified Laspeyres formula:
(1) /^[SQoPf/SQoPjXlOO, where P 0 is the
price of a commodity in the comparison period
and Pi is its price currently. Qa represents the
quantity shipped during the weight-base period.
An alternative formulation more closely ap­
proximates the actual computation procedure:

Ii=[X(QaP 0) (Pi/P0)/2QaP0]X 100

In this form, the index is a weighted average
of price relatives for each item (Pi/P0). The ex­
pression ( QaJPo) represents the weights in value
form and the “P ” and UQ” elements (both of which
originally relate to period “ a” but are adjusted for
price change to period uo” ) are not derived sep­
arately. Each value weight includes not only the
value of items priced but also the values of un­
priced items whose price movements are assumed
to behave similarly. When new weights are intro­
duced, the index with new weights is linked to
the index constructed with the earlier weights.
The weight adjustment itself therefore affects only
the later calculations of average price change.
When specifications or samples change, the item
relatives must be computed by linking (multiply­
ing) the relatives for the separate periods for
which the data are precisely comparable. (For a
somewhat more detailed treatment, see chapter 10,
Consumer Prices.)

Base Period
The Wholesale Price Index has been computed
on the base 1957-59 = 100 (average of 1957, 1958,
and 1959) since January 1962.11 Earlier bases were
1947-49,1926, and 1913. New items (or new index
groupings consisting primarily of new items) in­
troduced into the index after 1957 cannot be cal­
11 Rebasing factors for shifting the indexes from the 1957-59
base to the 1947-49 base are published in the January 1962
final report, Wholesale Prices and Price Indexes.
“ Before the 1952 revision (calculated back to 1947), priced
items in the index were weighted only by their own market values.
“ The censuses are, in general, collected at 5-year intervals.


culated on the 1957-59 base. Such indexes are
published with separate bases related to the date
of introduction.

The Wholesale Price Index weights represent the
total net selling value of commodities produced,
processed, or imported in this country (including
Alaska and Hawaii) and flowing into primary
markets. The values are f.o.b. production point
and are exclusive of excise taxes. The value of
interplant transfers, military products, and goods
sold at retail directly from producing establish­
ments also are excluded. Thus the definition of
the weights conforms to the universe definition.
Each commodity price series is considered repre­
sentative of a class of prices and is assigned its
own weight (the shipment value of the commodity)
plus the weights of other related commodities not
directly priced but whose prices are known or
assumed to move similarly.12 The assignment of
price movements for priced commodities to those
for which quotations are not obtained is referred
to as imputation. For some commodities—such as
ships and some kinds of custom-made machinery—
it is not possible to obtain direct measures of price
movement. The weights for such items are as­
signed to other commodities or groups of com­
modities for which prices are available. Usually,
this assignment is made to priced commodities
which have a similar manufacturing process, on
the assumption of similar price movements. Price
movements for attachments and parts for certain
machinery often are imputed to the machine itself.
The Bureau’s policy is to revise the Wholesale
Price Index weights periodically when data from
the industrial censuses become available.13 The
weights beginning in 1961 are based on the 1958
industrial censuses. Indexes for 1947 through
1954 are based primarily on the 1947 censuses. In
the January 1955 index, adjustments were made to
align the major group weight totals with 1952-53
average shipment values as reported in the Amnual
Surveys of Manufactures. Weights based on the
1954 Census shipment values were introduced in
January 1958. The 1963 Censuses of Manufac­
tures and Mineral Industries provide data for re­
vising these weights. Minor redistributions of the



weights are made from time to time to take care
of additions or deletions of commodities from the
The Bureau publishes the relative importance
of each item in the WPI rather than the actual
values used as weights. The relative importance
of an item represents its basic value weight used
in the index, including imputations, multiplied by
the relative of price change from the weight date
to a later date and the result expressed as a per­
centage of the total for all commodities or for some
index grouping.14

Imputing Missing Prices
Whenever price data are not available for a
particular month, it is necessary to estimate the
missing price for use in the calculation of the
index. For commodities in the farm products and
processed foods groups, out of the market season­
ally, the price in off-season is imputed from the
combined movement of the related commodities
for which prices are available for the two periods
being compared. For other commodities in the
first month of delinquency, prices are estimated by
a standard procedure. This procedure either holds
prices unchanged from the preceding month or
provides imputations based on the price movement
of other series in the cell (also taking account of
previous movements of the delinquent series), de­
pending on whether these other series have been
unchanged or have changed by more or less than
1 percent over the latest 3-month period.
Prices for some custom-made items are reported
to B L S as estimates. For example, prices for
fabricated structural steel for buildings and
bridges are obtained from producers who reprice,
each month, steel of the same specifications as used
in structures on which they had been engaged at
the time pricing for the WPI was initiated.
Elevators, normally sold including installation,
are reported f .o.b. plant—i.e., excluding transpor­
tation and installation cost—for use in the WPI.
Before the institution of pricing for collapsible
tubes, this commodity price was calculated on the
basis of the price for the component metal.
14 The use of relative importance data to construct indexes for
groups of products is discussed in B L S Bulletin 1214, 1957,
Wholesale Prices and Price Indexes, 195^-56. Relative im­
portances as of December for each item are published in annual
15 Wholesale Prices and Price Indexes, 1961 (B L S Bulletin
1382, 1964), p. 2.

Weekly Index
The weekly index represents a modified exten­
sion of the latest complete monthly index calcula­
tion. It is based on price data for a group of
over 250 commodities, mostly metals, petroleum,
and food product items which are generally charac­
terized by more frequent price changes than the
other commodities. Other commodities are esti­
mated unchanged. The average percent change
from the latest monthly index is converted to index
form for publication. No attempt is made to
maintain a continuous series by correcting these
indexes to new monthly levels.

Analysis and Presentation
The monthly Wholesale Price Index is published
first in a press release issued 4 to 5 weeks after the
pricing date (Tuesday of the week containing the
15th of the month). Indexes are shown for all
groups and subgroups as well as for All Commod­
ities, Farm Products and Processed Foods com­
bined, and All Commodities Other than Farm
Products and Foods. A brief description and
analysis of the causes of price movements are in­
cluded. The monthly detailed report, issued some
time after the press release, carries all data for
which wholesale price indexes are published, in­
cluding item indexes and all special group indexes.
Prices for many individual commodities also are
included. This report includes a more compre­
hensive analysis than that given in the press
release. Annual bulletins carry the data for the
year. In addition, numerous historical tabula­
tions at various levels of detail are available on
In the first month of publication, the monthly
indexes are preliminary and subject to revision.
In the following month, revisions and substitutions
of actual prices for estimated prices are incorpo­
rated. The indexes are published at this time as
final. Any revisions which come to light after the
final index has been published are made subject to a
standard correction policy which takes account of
the magnitude of the error.15
In addition to the weekly estimated All Com­
modities index described earlier, weekly indexes
for finished steel products and for petroleum
products are issued in separate releases.




B LS 1810

Budget Bureau N o. 44 R 602.5


Approval expires:




Washington, D .C . 20212

Code No. .. .

.... ................ . -


Firm name
Plant or division

Mfr. □

° ther

( Specify )


(Zip code)

(City and State)


Information authorised hy

Information furnished by
Mail schedule to

_ - Title _


A n n pa c c


(Zip code)

(City and State)


1. COMMODITY DESCRIPTION (include style no., model no., lot no., grade, brand, etc.)








A. C lass of seller________________ __________to C lass of customer___________________

(mfr., importer, etc.)

(wholesaler, user, etc.)

B. Prices are: Actual transaction prices* Z Z 5 L ist prices subject to discounts 1 j ;
L ist prices less discounts Z Z ; Other Z Z (specify) ------------------------------------------------------------------------------------------------------C. Unit quoted_______________________D. Size of order_____ ________________________
E. Shipping terms (f. o. b.; frt. allowed; e t c . ) _ ......................................... ................................. .................
F. Type of package used;



? Carton

G. Is refund allowed for returnable container?


Y es

5 Bag.



» Other---------------------------------


B " Y e s ” , explain---------------

♦A ctual sellin g prices to class of customer, for size of order, shipping terms, and discounts reported.

222-617— 66----- 8



A. Trade discount.
B. Quantity discount (based on size of order specified in 3D)
%of sales
C. Cash discount.

D. Seasonal discounts

E. Cumulative volume discount

F. Rebates (monthly)

G. Other discounts, allowances, free deals, etc. (explain fully).


A. These are included in prices quoted |~]
Not included Q
B. If tax is included, give example of how to calculate price excluding tax.
Other mfr.
(O .E.M . or
assem bler)





Sales (approx . % )

(specify )



BLS Representative

U ser



Budget Bureau No. 44—R 194.11.
Approval expires: 12.31-67

B IS 4 7 3
(R«v. 8 - 1 5 - 6 1 )

Burea u


La b o r


S t a t is t ic s


D .C .


D e ar S ir :
The p ric e d ata which you provide is used in com puting the W holesale
P r ic e Index which is the o ffic ia lly accep ted in d icato r of p r im a r y m ark et p ric e
m ovem en ts. The index is w idely used by in d u stry and govern m en t.
T h ese volu n tary r e p o r ts, su b m itted by you and other b u sin e ssm e n , a r e
the m a jo r so u rc e of in form ation used in p re p arin g th is index. The in form ation
you provide is s t r ic tly con fid en tial and open to in spection only to sw orn
em p lo y ees of the B u reau of L a b o r S t a t is t ic s .
P le a se u se the en clo se d envelope, which r e q u ir e s no p o stag e , fo r
retu rn in g this sch e d u le. Y our continued coo peration is g r e a tly a p p re c ia te d .

In the boxes provided on the other side, please be sure to indicate all changes in
that may have occurred since your last report.
Your cooperation in keeping all information current
is a great aid in computing a reliable, accurate Wholesale Price Index.



Code No.


...... ~


1 . COM M ODITY DESCRIPTIO N (Please indicate all changes.)

G ive date, nature, and
estim ated v alu e of chan g e




(Please indicate all changes. )•

Date and nature of change


C la s s o f selle r an d customer
S iz e o f order

Shipping terms_____________
O the r (Specify.)

3 . D IS C O U N T S , A L L O W A N C E S , AND T A X E S

In d icate a ll d is co u n ts , a llo w a n c e s , and ta x e s a p p lic a b le to
above b a s is of q u otatio n . T h is information is needed to a rriv e at the A C T U A L S E L L I N G P R I C E . ( P le a s e
in d ic a te a ll c h a n g e s .)
.............. .......................... 1

Date and nature of change


Q u an tity discount
Trad e discount

H av e a n y ind icated

C a sh discount

d iscou nts b e en d ed u cted
from the reported p ric e ?

S e a so n al discount
O ther discount
O ther charges

H a v e a n y of these

E xcise taxes

b e e n in c lu d e d ?


For the com m odity d e scribed in item 1, p le ase enter below the current price for the date in d icated , on the b asis quoted in
item 2.

P ric e a s of Sep t. 14, 1965





( i f any)


O c t. 12, 1965

A p r. 12, 1966

N ov. 16, 1965

May 17, 1966

D e c . 14, 1965

June 14, 1966

J a n . 1 1 ,1 9 6 6

Ju ly 12, 1966

F e b . 15, 1966

Aug. 16, 1966

Mar. 15, 1966

Sept. 13, 1966





( if any)









GPO 88 5-3 1 4


Seasonally adjusted indexes are not published.
However, because of the interest in the subject,
the Bureau has published seasonal factors for
about 180 commodites and commodity groups of
the W PI and for the All (Commodities index.16

Uses and Limitations
The Wholesale Price Index is used by govern­
ment and private research agencies for many pur­
poses, including market analysis, escalation of
long-term purchase and sales contracts, formula­
tion of monetary policies, and as an indicator of
economic trends.
A 1961 survey of users of the WPI revealed that
more than one-half use the All Commodities index
as a general economic indicator. About 40 per­
cent use that index or its components to compare
with their selling or buying prices. The survey
revealed that over 10 billion dollars (in terms of
unexpired value) in long-term contracts for pur­
chase of material or lease of industrial property
are escalated in accord with changes in the total
index or its components. Government agencies
and private research groups also use the component
series in deflating value data in preparation of the
gross national product estimates and in studies
of economic growth.
The index also is used by buyers and sellers of
commodities—purchasing agents and sales man­
agers. In most of these cases, it is not the All
Commodities index, but rather the group indexes
and the individual price series which are employed.
Buyers of commodities are able to check both the
amounts which they pay for goods and the general
movement of their purchase prices against the
index. The use of the index for checking absolute
price levels is limited substantially however. The
Bureau’s main goal has been to measure the direc­
tion and amount of change, and only incidentally
to measure actual selling prices.
The index, as a measure of general and specific
price trends, also is used widely in budget making
and review, both in government and in industry; in
planning the cost of plant expansion programs; in
appraising inventories; in establishing replace­
ment costs; etc. Components of the index also are


used in LIFO (Last-In, First-Out) inventory ac­
counting by some organizations.
While the WPI often is used to measure change
in purchasing power of the dollar, it should not
be used to measure change in general purchasing
power, prices at retail, securities prices, etc. Com­
parisons between the level of the WPI, the Con­
sumer Price Index, and the indexes of prices of
farm products show relative change from a base
period, but comparisons of the index levels should
not be used as a measure of the actual margins be­
tween farm prices and manufacturing or between
manufacturing and retail. Its commodity classifi­
cation structure should be borne in mind when
using it to measure price changes for industries,
many of which make divers products not classified
as their “primary” products.17
Again, as in other measures, the WPI has some
limitations even in the field for which it is con­
ceptually designed. Segments of the index are
used as deflators of gross national product data,
but gaps in WPI coverage leave considerable areas
for which deflators have not been provided.
The Wholesale Price Index is based on a pur­
posive, judgment sample. The All Commodities
Index can be assumed to be more reliable than a
component group index, in general. Also, it can
be assumed that the reliability of the index has
increased over time as the sample has expanded.18
As the economy has produced an increasing pro­
portion of fabricated finished goods (whose price
changes are relatively infrequent), over the years,
the WPI has become somewhat more stable. Cur­
rently, new products are added each year. In
earlier decades there were also major additions of
large numbers of new items at one time, in com­
modity areas previously under-represented. These
sudden expansions could have made it appear that
prices had suddenly stabilized.
To the extent that quality improves (or deteri­
orates) over the years, the index errs when no
adjustment is made. However, the Bureau makes
suitable adjustments whenever possible. Assum­
18 Seasonal Adjustment Factors, Wholesale Price; In d ex: Se­
lected Series 19^8-61 (B L S Bulletin 1379, 1963).
17 See chapter 12 on Industry-Sector Indexes.
18 The sample of priced items doubled in 1952 to about 1,859
items and has increased by about 400 since then.



ing quality improvement, the index would have an
upward bias if direct comparison were made
between unimproved and improved articles. If,
on the other hand, such changes were consistently
made by linking, a downward bias would result.

Since the Bureau has not adopted either method
exclusively, and in many instances tries to evaluate
the changes brought to its attention, the bias that
may exist is considered to be small. However, no*
measure of its magnitude is available.

Technical References

1. Cunningham, Francis S., “ The Use of Price Indexes in Escalator Contracts,” M onthly Labor
Review , August 1963, pp. 948-952. Reprint No. 2424.
A statement of the use of the Wholesale and Consumer Price Indexes in escalating purchase
and sales contracts and wages, with some specific suggestions and pitfalls noted.
2. Evans, W. Duane and Hoffenberg, Marvin, “ Input-Output Relations and Appraisal,” in Stu d ies
in Incom e an d Wealth , National Bureau of Economic Research, New York, 1955, Volume 18.
A statement of the conceptual framework, data, pricing problems and significance in economic
analysis of the U.S. Government’s interindustry statistical study of 1947.
3. Evans, W. Duane and Hoffenberg, Marvin, “ The Inter-Industry Relations Study for 1947,” The
Review of Econom ics an d Statistics, May 1952, pp. 97-142.
A description of scope, uses, and methodology of the U.S. Government’s interindustry
statistical study of 1947. Includes discussions of computational problems, areas of use, data
requirements, etc.
4. National Bureau of Economic Research. The P rice Statistics of the Federal Government: Review ,
A p p ra is a l, an d Recom m endations. (Washington, N BER General Series, Number 73, 1961),
496 pp.
An appraisal of price statistics of the Federal Government by the Price Statistics Review
Committee of NBER, covering uses, concepts, collection, and publication, sampling, and other
aspects of the Consumer Price Index, Wholesale Price Index, Index of Prices Paid by Farmers,
and other price measures.
5. Searle, Allan D., “ Weight Revisions in the Wholesale Price Index, 1890-1960,” M onthly L ab or
Review, February 1962, pp. 175-182.
History of weight changes and weighting concepts, from inception of the Wholesale Price
6. U.S. Congress, Joint Economic Committee. Government P rice Statistics: H earin gs: Subcommittee on
Econom ic Statistics, 87th Congress, 1st sess., Part 1, Jan. 24, 1961, 526 pp; Part 2, May 1-5
1961, 265 pp.
Part I presents the report, Price Statistics of the Federal Government, prepared by the
N B ER (q.v.); Part II contains statements of private and government economists including the
response by Ewan Clague, Commissioner of Labor Statistics.
7. U.S. Department of Labor, W holesale P ric es, 1890 to 1899, Bulletin 27, March 1900.
Describes an inquiry into the course of wholesale prices for the purpose of continuing the
study contained in the Report on Wholesale Prices, Wages, and Transportation made by the
Senate Committee on Finance, March 3, 1893 (pp. 237-313).
8. ----- . Course of Wholesale P rices, 1890 to 1901 Bulletin 39, March 1902.
Describes United States Senate Finance Committee index (pp. 205-211), and Department
of Labor index (pp. 212-243).
9. U.S. Department of Labor, Bureau of Labor Statistics, W holesale P rices an d P rice In dexes , 1954-56,
Bulletin 1214, 1957.
Method of Calculating Special Group Indexes, (pp. 12-13), Calculating Relative Importance
Data (p. 14), Description of Indexes by Stage of Processing (Economic Sector Indexes) (pp.
15-22); A Possible Effect on Weight Revisions (p. 7).
10. ----- . S easo n al A djustm ent F ac to rs; W holesale P rice In dex: Selected Series 1948-1961 (BLS Bul­
letin 1379, 1963).
Seasonal adjustment factors for 183 commodities and commodity groups, and description
of BLS seasonal adjustment method.
11. ----- . W holesale P ric es an d P rice In dexes, 1957 , Bulletin 1235, July 1958.
Indexes by Durability of Product (Economic Sectors by Durability of Product) (pp. 11-14).


Technical References—Continued
U.S. Department of Labor, Bureau of Labor Statistics, W holesale P ric e s a n d P ric e In dexes, 1958 ,
Bulletin 1257, July 1959.
Describes Supplementary Inquiry on Wholesale Price Reports (discount study) (pp. 10-12)
and January 1958 Revision of the Weighting Structure (pp. 14-16).
----- . Wholesale P ric e s an d P rice In dexes, 1961, Bulletin 1382, 1964.
January 1961 Revision of the Weighting Structure (pp. 14-16).
----- . Wholesale P rices an d P rice In dexes, January 1962 (final) and February 1962 (preliminary).
Describes New Base Period for Wholesale Price Indexes and Conversion Procedure (pp.
1-2) and Rebasing Factors (pp. 48-71).
U.S. Senate, Wholesale P rices, W ages, an d T ransportation, Senate Report No. 1394, Part I, “ The
Aldrich Report,” 1893.
Contains a summary of the complete Senate report on wholesale prices, on wages, and on
transportation made in response to a Senate resolution of March 3, 1891.

— A lla n D. S e a r l e , H e l e n F. H ald and B e n n e t t R. M oss
Office of Prices and Living Conditions

Chapter 12. Industry-Sector Indexes
During recent years a growing need for compre­
hensive measures of industrial prices, in addition
to the market oriented prices of the Wholesale
Price Index, has become increasingly apparent.
As a result, the Bureau initiated a program of
industry-sector price indexes based upon data
collected for the WPI.
An industry or sector price index is essentially
a composite index made up of price series that
match the economic activity of a defined industry
or economic sector. The Wholesale Price Index,
on the other hand, is compiled according to com­
modity rather than industry groupings. The
number of sectors or industries for which these
series are compiled depends somewhat upon re­
sources available for additional pricing.
A set of industry-sector price indexes covering
the years 1947 through 1953 was prepared in the
early 1950’s as part of the Bureau’s project on
interindustry economics. These indexes, generally
a regrouping of the Wholesale Price Indexes into
the interindustry (input-output) classification
structure, were designed for revaluing bills of
goods and industry outputs. In 1959, another set
of such indexes was compiled for the Bureau of
the Census in connection with that agency’s con­
struction of the 1958 production index benchmark.1
This second group of price indexes was used in
deflating values of shipments in those census
product classes where physical production data
were lacking or unsatisfactory. Again, these were
essentially indexes of commodity prices, classified
as primary to a given industry.2
The need for the Bureau to develop IndustrySector Price Indexes became increasingly appar­
ent in 1960 and 1961, when the Price Statistics
Review Committee of the National Bureau of
Economic Research recommended to the Bureau
of the Budget that the basic objectives of an in­
dustrial price program should be comprehensive­
ness, maximum detail in reporting, and groupings
most useful in economic analysis.3*8 The com­
mittee stated “. . . It seems desirable that the sub­
classification should aim at fitting into the Stand­
ard Industrial Classification.”

In 1962, the Bureau of Labor Statistics initiated
the development of industry and sector price
indexes. Because of its scope, the program is
viewed as a long-run program to be accomplished
in several stages. The first stage is being devoted
to the study of conceptual and data problems with
only a gradual expansion of commodity pricing.
The first indexes to be developed will be output
price indexes utilizing gross shipments weights.
Priority will be given to indexes for the manu­
facturing, mining, and agricultural divisions of
the Standard Industrial Classification. Input
price indexes, i.e., indexes representing the price
of industrial purchases, will come later. Eventu­
ally the work will expand to include sector in­
dexes with both value-added and net weights from
which intra-sector shipments are excluded.

Description of the Survey
An industry or sector price index is a composite
index derived from several series of prices that
closely match the economic activity of a specified
industry or industry sector. These indexes may
be either output or input price indexes based upon
either the products and services sold or the prod­
ucts and services purchased by an industry. An
output price index for a given industry represents
1 Neither of these earlier indexes were published, but were
made available to other government research agencies. See
Chapter 30, section on Economic Growth Studies, for background
on interindustry studies.
2 The classification of establishments into industries, in this
program, follows the guidelines established by the Bureau of the
Budget in its Standard Industrial Classification (SIC ) system,
as revised in 1957. Under this classification system, related
products or services are grouped together and given an industry
code number (consisting of 4 digits). Every establishment is
assigned to the industry in which its most important products or
services, in terms of values, are classified. Many industries
contain establishments which produce significant quantities of
goods and services that are classified in other industries. These
goods usually are referred to as “ secondary products.” See
appendix B.
8 Government Price Statistics, Hearings before the Subcom­
mittee on Economic Statistics, of the Join t Economic Committee,
Congress of the United States, P art I, January 24, 1961, page 64.
Also see report of United Nations Economic and Social Council;
Problems and Methods in the Gathering of Representative and
Comparable Wholesale Price Series, E/CN . 3/264, 15 March 1960,
Chapter II.


price indexes for the products of the industry,
averaged together according to the relative im­
portance of production of each product to the in­
dustry. An input price index for an industry con­
sists of an aggregation of price indexes for all
of the commodities and services purchased by the
industry, averaged together according to the rela­
tive magnitude of the purchases.
The Bureau’s work is being directed first toward
two sets of output indexes. One set is being
weighted by gross shipments of products “made
in the industry” to be used for deflating industry
shipments. The second set will be output price
indexes of product shipments classified on the in­
dustrial basis but weighted by shipments of the
product produced anywhere in the economy. The
principal use of the second set is for input-output

Ultimately, the scope of the universe will be
defined in terms of the Bureau of the Budget’s
Standard Industrial Classification (SIC) system
as revised in 1957, which covers all domestic eco­
nomic activity. This system groups together re­
lated products or services and assigns them in­
dustry and sector codes. Currently, however, the
scope of pricing is effectively restricted to the
commodities’ coverage of the WPI because of the
use of WPI price data.
I f price indexes are to parallel industry output
data, the indexes should cover the total output of
each industry including the value of interplant
transfers, the value of sales to all classes of custom­
ers, and the value of industrial services. They
should include the value of sales for export but
exclude excise taxes and costs of transporting fin­
ished goods to purchasers. This is consistent with
the “total activity” coverage of statistical series
on employment and production.
Input price indexes of materials consumed in
production should cover total materials inputs of
the industry. This would include imports for con­
sumption, and also transportation and delivery

Prices and Base Period
The prices used in the current Industry-Sector
Indexes are in general those used in the Wholesale


Price Index. In the Wholesale Price Index, pri­
mary market prices, f.o.b. production point, are
used. For the Industry-Sector Indexes, pricing
eventually should be extended to all classes of cus­
tomers, including retail, for use in the output in­
dexes. Buyers’ prices, including shipping costs,
should be used for input indexes and should repre­
sent the particular mix of products purchased by
the buying industry.
The current reference base period is the same
as that for other Federal Government indexes—

The Standard Industrial Classification (SIC)
of the Bureau of the Budget provides the frame­
work for the Industry-Sector Index classification
scheme. Within this framework, individual prod­
ucts are given a 7-digit code by the Bureau of the
Census.5 The product indexes are aggregated to
5-digit product classes and 4-digit industries.
Industry indexes can be aggregated to 3- and
2-digit levels as well. Four-digit industry indexes
also can be aggregated to fit the sectoring plan of
the latest Input-Output model.6

Sampling and Estimating Procedures
Currently the Industry-Sector program largely
depends on price data already available in the
Wholesale Price Index. However, extension of
industrial pricing is being geared to the new pro­
gram. A sampling plan has been prepared which
outlines industrial sectors that should have prior­
ity as new pricing is undertaken. As the result of
this analysis, pricing may be cut back in some sec­
tors to permit extension of pricing into inade­
quately covered sectors.
4 See Chapter 30, section on Economic Growth Studies, for a
discussion of the nature and uses of input-output data. See also
W. Duane Evans and Marvin Hoffenberg, “The Inter-Industry
Relations Study for 1947,” The Review of Economics and Star
tistics, May 1952, pp. 97-142; and “ Input-Output Relations and
Appraisal,” Studies in Income and Wealth, vol. 18, National
Bureau of Economic Research, New York, 1955.
5 The SIC provides no product codes.
6 The Input-Output model referred to is that compiled by the
Office of Business Economics, U.S. Department of Commerce.
See Goldman, Morris R., et al„ “The Inter-Industry Structure of
the U .S.” Survey of Current Business, November 1964, pp.



Price data used in computing an industry-output price index should be representative of the
output of the industry, that is, of the values of
products made in plants classified in the industry,
but should exclude prices of products primary to
the industry but made in plants classified in other
industries. Even though the products are the
same, the prices received by the primary and sec­
ondary industries may be different, sometimes be­
cause these industries sell to different types of
users. As a rough guide to the adequacy of sam­
pling, the immediate objective is to represent at
least 50 percent by value of the commodities in­
cluded in each 5-digit Census product class.7 This
percentage will be slightly less at higher levels
of aggregation. This criterion will be adjusted
to levels indicated by differences in variability of
price change among product classes as experience
makes possible the use of more sophisticated sam­
pling approaches.
Weights for the output indexes are 1958 value
of shipments obtained from the Census of Manu­
facturers, the Census of Mineral Industries, and
data of the U.S. Department of Agriculture.
Values include interplant transfer values, values
for goods produced and consumed in the same
establishment, and the value of goods sold for ex­
port. Values of imported commodities are not
included. The difference in the scope of the
weights, as compared with the WPI, stems from
the objective in this system to match price data
with the scope of domestic industry production.
Each priced product actually represents a class
of commodities and is assigned its own weight plus
the weights of other products not directly priced
in the index but whose prices are known or as­
sumed to move similarly. Values for unpriced
products which cannot be assigned to a specific
priced commodity are imputed to the average
movement of the product classes in which they
For use in deflating industry shipments, the 4digit (SIC) Industry Indexes will be derived
from 5-digit product class indexes weighted to­
gether by their shipments value for the particular
industry, i.e., the “made-in-the-industry” value.
For indexes to be used for deflating interindus­
try product flows, weights will be used at both

5- and 4-digit levels of aggregation that will corre­
spond to the production of goods made both within
and outside of the primary industries.

Formula and Calculation
A modification of the Laspeyres fixed-weight
formula is used. The underlying formula is:
I i= '2 ( P i/Po-PoQa)/2P0Qa where P jP o is an indi­
vidual product price index. P 0Qa, the value
weight (base year price times weight year quan­
tity), currently is the 1958 value of shipments
adjusted for price change from the year 1958 to
the average of 1957, 1958, and 1959. In succeed­
ing years, new weights will be introduced whenever
weights are revised for the comprehensive Whole­
sale Price Index.9
In actual practice the calculation may be some­
what more involved than indicated by the simple
formula presented previously. For example, in­
dexes used for deflating industry shipments values
should employ product weights based upon values
of commodities made within the same industry.
Since the Census data for such values are avail­
able only at the 5-digit (product-class) level, it
is necessary first to construct product-class indexes
based upon total output (wherever made) weights.
Then, using 5-digit made-in-the-industry weights,
the product-class indexes are combined to the 4digit industry level.

Analysis and Presentation
The published indexes for selected 5-digit prod­
uct classes and 4-digit industries are annual aver­
ages for the period 1957 through 1964, and cover
only a limited number of manufacturing and min­
eral industries. Future plans call for the publica7 As the component 7-digit products are usually quite homo­
geneous in physical structure, it can be assumed that one com­
modity specification priced directly will adequately represent an
entire 7-digit product. However, as a consequence of this assump­
tion, changes in classification detail at the product level from
one Census period to another may result in changes in reported
coverage without any corresponding change in sample.
8 This procedure is the same as that employed in the WPI.
However, as the product classes are defined differently, an un­
priced commodity may have a different price movement imputed
to it in the Industry Index program than it has in the WPI.
9 Much of the value used in weighting the WPI is derived
from the periodic Census of Manufactures and Mineral In­
dustries which are currently collected on 5-year cycles. See
Chapter 11, Wholesale Prices.


tion of monthly indexes as they become available,
and a bulletin containing historical indexes and
conceptual treatment.

Uses and Limitations
Price statistics organized along industrial lines
have particular relevance to studies of economic
growth, productivity, and other types of industrial
and economic analysis where the emphasis is on
industrial structure as distinct from market or
commodity-use classifications.
Whether an index meets a given specific need
depends largely upon its commodity coverage and
its weighting structure. An important use of an
output index weighted by gross shipment values
is to deflate value series in order to arrive at meas­
ures of output in constant dollars. Most measures
of output and productivity rely primarily upon
physical quantity data for the various products of
an industry, but in cases where quantity data are
not available, deflated values can be used if suit­
able price indexes are available for use as deflators.
Deflated value data also may serve as a check on
production indexes prepared from quantity data
and unit-value weights. There are many sectors
of the economy for which the analysis of industrial
output is severely limited because appropriate
price indexes are not available.
Essentially the process of deflation provides a
means of obtaining an estimate of quantity change
from available data on total dollar value and a
price index. I f the dollar values themselves are
divided by the price index, the resulting dollar
values express the sales value in terms of purchas­
ing power of the dollar as of the base period of the
index. Or an index of dollar volume can be
divided by the price index to obtain a production
The output indexes also may be used for com­
paring movements of prices with other industrybased statistical measures such as employment,
earnings, productivity, etc. Price indexes consist­


ent with total shipments weights will be useful
for deflating industry inputs. For example, the
appropriate index for deflating the value of alumi­
num purchased by an industry would be the index
whose components represent the entire economy’s
shipments of aluminum to buyers in this industry
rather than the aggregate output of the primary
aluminum industry.
Input price indexes will be especially useful to
research departments in private industry as well
as to public agencies in making cost studies. They
should be consistent in coverage with B L S series
on average hourly earnings, another important
element of cost. For contract escalation, they will
give index users a wider choice of indexes. Input
price indexes, however, are not available yet.
There are a number of uses which combine out­
put and input indexes. For example, gross output
price indexes and materials input price indexes
can be used to yield a measure of value added in
constant dollars.
The prices used in contracting the currently
published indexes are those regularly collected on
a monthly basis and used in the calculation of the
comprehensive Wholesale Price Index. These
prices generally are at the primary market level
but a few are at other levels. It must be assumed
that these price movements are similar to the mar­
ket level of sales represented by the Census data.
To include interplant transfer values and values
of goods produced and consumed in the same in­
dustry, it is necessary to assume that price move­
ments of goods in commercial markets represent
the price changes of goods not sold in commercial
Until additional pricing can be done, these new
indexes will be limited by the coverage—commod­
ity and class of customer—of the comprehensive
Wholesale Price Index.
10 It can be shown that division of the value index by the
Laspeyres (base-year-weighted) price index yields a production
index of the Paasche (current-year-weight) form. Division by
the Paasche price index, conversely, yields a quantity index of
the Laspeyres type. See Chapter 23, Output per M anhour:


Technical References

1. Evans, W. Duane and Hoffenberg, Marvin, “ The Inter-Industry Relations Study for 1947,” The
Review of Econom ics an d S ta tistic s , May 1952, pp. 97-142.
A description of scope, uses, and methodology of the U.S. Government's interindustry
statistical study of 1947.
2. Goldman, Morris R.; Marimont, Martin L.; and Yaccara, Beatrice N., “ The Inter-Industry Structure
of the United States,” Survey of Current B u sin e ss , November 1964, pp. 10-29.
A report showing preliminary results of the 1958 interindustry relations study and containing
tables of the percent distribution of 1958 gross output in 86-industry detail.
3. Moss, Bennett R., “Industry and Sector Price Indexes,” M onthly L abor Review , August 1965,
pp. 974-982.
Contains price indexes for about 50 4-digit (Standard Industrial Classification) industries,
together with a technical note on concepts, methodology, and uses.
4. National Bureau of Economic Research, “ Input-Output Relations and Appraisal,” S tu d ie s in Incom e
an d W ealth , vol. 18, New York, 1955.
A description of scope, uses, and significance in economic analysis of the U.S. Government’s
interindustry statistical study of 1947.
5. ----- . The Price Statistics of the Federal Government, New York, 1960.
An appraisal of price statistics of the Federal Government by the Price Statistics Review
Committee of NBER, covering uses, concepts, collection, and publication, sampling, and other
aspects of the Consumer Price Index, Wholesale Price Index, Index of Prices Paid by Farmers,
and other Price measures.
6. U.S. Congress, Joint Economic Committee. Government P rice S ta tistic s , H earin gs before the S u b ­
committee on Econom ic S tatistics , P a rt I an d P a r t I I , Washington, D.C., 1961.
Part I presents the report, The P rice Sta tistic s of the F ed eral Government, prepared by the N BER
(q.v.); Part II contains statements of private and government economists including the response
by Ewan Clague, Commissioner of Labor Statistics.

—B e n n e t t R . M oss
Office of Prices and Living Conditions

Chapter 13. Daily Spot Market Prices
As early as January 1934, at the request of the
U.S. Department of the Treasury, the Bureau of
Labor Statistics began the computation of a daily
commodity price index, using quotations for sen­
sitive commodities. It was released first to the
general public in January 1940. In 1952, in con­
nection with the revision of all its major price
index series, the Bureau issued a new Daily Index
of Spot Market Prices. The new index was not a
continuation of the old series, but was based on a
new sample of 22 commodities and was calculated
on a 1947-49 base; in contrast, the old index was
based on 28 commodities and was calculated with
August 1939 as base.
In January 1962, the 22-commodity index was
recalculated on a 1957-59 base to correspond to the
base period adopted for other Federal Government
general purpose indexes.

Description of Survey
The Daily Spot Market Price Index is a measure
of price movements of 22 sensitive basic commodi­
ties whose markets are presumed to be among the
first to be influenced by changes in economic con­
ditions. As such, it serves as one early indicator
of impending changes in business activity.
The commodities used are in most cases either
raw materials or products close to the initial pro­
duction stage which, as a result of daily trading in
fairly large volume of standardized qualities, are
particularly sensitive to factors affecting current
and future economic forces and conditions.
Highly fabricated commodities are not included
for two reasons: (1) they embody relatively large
fixed costs which fact causes them to react less
quickly to changes in market conditions; and (2)
they are less important as price determinants than
the more basic commodities which are used
throughout the producing economy.
A spot price is a price at which a commodity is
selling for immediate delivery. In the absence of
a spot price, a bid or an asked price may be used.

Some of the prices used are nommal prices in that
they are not actual transaction prices. Often they
are exchange prices—a price for a completely
standard commodity which eliminates the effect of
minor quality changes on actual transaction
prices.1 Trade publications may use this type of
price for commodities such as cocoa beans, coffee,
and wool tops. The price for print cloth is an
average of spot price and price for most distant
forward contract because it was determined that a
large part of the sales of print cloth are made on a
contract basis.
The 22 commodities are combined into an “All
Commodities” grouping, with two major sub­
divisions : Eaw Industrials, and Foodstuffs. Raw
Industrials include burlap, copper scrap, cotton,
hides, lead scrap, print cloth, rosin, rubber, steel
scrap, tallow, tin, wool tops, and zinc. Foodstuffs
include butter, cocoa beans, corn, cottonseed oil,
hogs, lard, steers, sugar, and wheat.
The items upon which the index is based are
classified further into four smaller groups: Metals,
Textiles and Fibers, Livestock and Products, and
Fats and Oils. However, some of the 22 com­
modities do not fall into one of these four group­
ings. For example, sugar is not included in any
special group. Furthermore, the groupings are
not mutually exclusive. Lard, for instance, is in
both the Livestock and Products Index and in the
Fats and Oils Index.

Data Sources and Collection Methods
The prices used in the index are obtained from
trade publications or from other Government
agencies. A listing of prices is compiled each
morning for the preceding day. Prices for cocoa
beans, steers, sugar, wheat, burlap, copper scrap,
cotton, lead scrap, print cloth (spot), rosin, rubber,
steel scrap, wool tops, and zinc, are of the same
specification and source as those used in the com­
prehensive monthly Wholesale Price Index.2
1 Exchanges which issue spot prices have committees to make
a determination of the spot price for the standard commodity.
2 See Chapter 11, Wholesale Prices.




Prices for butter, com, hides, hogs, lard, tallow,
and tin are either differently specified spot prices
or from different markets.

This product is then multiplied by 100 to obtain
the index number for each period. The calcula­
tion is made by means of logarithms. The formula
reduces to

Selection of Products

Log /&=:§ Log Pic—2 log Po+44

The criteria for the selection of commodities
were (1) wide use for further processing (basic),
(2) freely traded in an open market, (3) sensitive
to changing conditions significant in those markets,
and (4) sufficiently homogeneous or standardized
so that uniform and representative price quotations
can be obtained over a period of time.
Subject to these restrictions, efforts were made
to include representative sensitive commodities
from as large a segment of the economy as possible.
Also, the influence of international markets upon
the economy was taken into account by the in­
clusion of some key commodities (such as crude
rubber and tin) which are important in inter­
national trade. Both in the sample and in the
index structure, an attempt was made to prevent
price movements of agricultural products from
dominating the movement of the index.

Estimating Procedures
The daily index is an unweighted geometric
mean 3of the individual commodity price relatives,
i.e., of the ratios of the current prices to the base
period prices. The use of the geometric mean
has the advantage that the index is not dominated
by extreme price movements of individual com­
modities. Since extremely large movements may
be atypical, it was deemed better to minimize their
effects, even at the expense of losing the effect of
large representative changes. However, the fact
that each of the commodities is unweighted in the
index means that a price change for rosin, a com­
paratively unimportant commodity, has as much
effect as an equal percentage movement in the
price of a very important commodity such as
wheat, cotton, or steel scrap.
The computation procedure involves obtaining
for each commodity the ratio of its price in any
given period to its price in the base period and
taking the 22nd root of the product of these ratios.


/fc= Index for a given day
Pic= Price for a given day
P Q= Average (geometric) price in base
44= Logarithmic constant which when di­
vided by 2 equals log of 100.
Monthly average indexes are obtained accord­
ing to the previous procedure, except that /\= th e
geometric average of the daily prices over the
month. In maintaining the index over time, it may
be necessary to change commodity specifications
or substitute entirely new products. These
changes are handled by a statistical linking pro­
cedure so that only actual price movements are
reflected in the index.

Analysis and Presentation
Indexes and prices are published each day, Mon­
day through Friday, based on data for the preced­
ing day. They also are available in a weekly sum­
mary issued each Wednesday. In addition, the
weekly summary gives daily spot prices for eight
commodities which are not included in the index.
The daily spot market indexes and prices also are
presented in bulletins issued at approximately 3year intervals. Historical series on the 1957-59
base are available in mimeographed form on a
daily basis beginning with 1950; for Tuesday of
each week from July 1946 through 1949; and for
four selected dates—August 15,1939, December 6,
1941, August 17,1945, and June 28,1946.

Uses and Limitations
A survey of users in 1964 showed that the Daily
Index is frequently used as a general economic in-8
8 The geometric mean of n figures is the nth root of their
product. Thus, the geometric mean of the numbers 1.5, 2.0, and
9.0 is 3.0 (1 .5 X 2 X 9 = 27. 3V27 = 3). The arithmetic mean,
is 4.2.


dicator, for gaging the direction of basic prices,
for forecasting general price movements, and for
current prices of specific commodities. Other uses,
frequently mentioned, are for market research and
for comparing price trends with the user’s selling
or buying prices.
The Daily Index of Spot Market Prices differs
from the Wholesale Price Index in method of con­
struction and weighting, as well as in the sample
of items for which prices are included. While it
is independent of the monthly comprehensive
index, changes in the Daily Index or its compo­
nents may foreshadow turns in Wholesale Price
Indexes. However, the Daily Index is not a good
indicator of current price trends for the whole
economy. For this purpose, the comprehensive
Wholesale Price Index should be used. The Daily
Index is, by design, very sensitive to price changes
in basic commodities but, because of its unweighted
structure, the magnitude of changes in any of the
index groups cannot be used as a reliable measure
of the general price change of all commodities
within the groups.
For many of the 22 items, the commodity ex­
change prices are based upon transactions which
cover as little as 25 percent of the total sold in all
markets. In some cases, the price is set by a com­
mittee of experts from the commodity exchange
for a standardized commodity. Also, when there
are not enough transactions from which to obtain
an actual market price, a “nominal” spot price is
set. From this, it is apparent that the exchange
prices may not always be representative of the
large volume of private transactions occurring out­
side the organized market. However, it is believed
that the reported exchange prices generally are
used as the basis for private negotiations.


Composition of Grouping Indexes
M etals: Copper scrap, lead scrap, steel scrap,
tin, and zinc.
Textiles and F ib e rs: Burlap, cotton, print cloth,
and wool tops.
Livestock and Products: Hides, hogs, lard,
steers, and tallow.
F a ts and O ils: Butter, cottonseed oil, lard, and
Specifications for Commodities Included in the Index as of January 1965
Burlap......... ...... 10 oz., 40", in carloads, ex-dock or ware- New York,
house, duty paid, per yd.
Butter................ Grade A, 92 score, per lb...................Chicago.
Cocoa beans....... Accra, per lb....... ......... ......................... New York.
Copper scrap__ No. 1 copper, refiners’, buying price, New York,
delivered buyers’ works, per lb.
Corn..............___ No. 3 yellow, per bu.............................. Chicago.
Cotton___ ___ Middling, 1", per lb_______________ 15 markets.
Cottonseed oil—_ Crude, Valley, per lb____ ______ ____ Memphis.
Hides................. Packer, light native, cow, per lb_____ Chicago.
Hogs................... U.S. No. 1, 2’s, and 3's, 200-220 lbs., per Chicago.
100 lb.
Lard................. . Prime Steam, in tanks, per lb.......... .
Lead scrap........ Battery plates, smelters’ buying price, New York,
delivered buyers’ works, per lb.
Print cloth......... 39", 80 x 80 count, 4 yds./lb., spot and New York,
nearby, per yd.
Print cloth......... 39", 80 x 80 count, 4 yds./lb., most dis- New York,
tant contract, per yd.
Rosin.................. Gum, WG grade, carlots, per 100 lb___ New York.
Rubber............... Crude, natural, No. 1 Ribbed Smoked New York.
Sheets, per lb.
Steel scrap_____ No. 1 heavy melting, (dealer), con- Chicago,
sumers’ buying price, including brok­
erage, delivered, per gross ton.
Choice, 900-1100 lbs., per 100 lb______ Chicago.
Raw, 96°, duty paid, per 100 lb............. New York.
Packers’ Prime, inedible, per lb______ Chicago.
Grade A, spot delivery, ex-dock, per lb_ New York.
W h e atNo. 1 Dark Northern Spring, per bu.__ Minneapolis.
No. 1 Hard Winter, per bu_________ Kansas City.
Wool tops.
Certificated spot price, nominal, per lb. New York.
Slab, Prime Western, for prompt de- New York,
livery, delivered, per lb.

— H e l e n F . H ald

Office of Prices and Living Conditions

Wages and Industrial Relations
Chapter 14. Occupational Pay and Supplementary Benefits
The Bureau of Labor Statistics, for many dec­
ades, has conducted studies of wages by occupation
and industry, based upon employer records. The
Bureau’s first such study, growing out of a study
by the U.S. Senate in 1891, resulted in a wage rate
record extending back continuously to 1860. Sys­
tematic collection of wage data by occupation and
industry has continued since the turn of the cen­
tury ; changes in coverage has been dictated mainly
by government requirements. A large survey
program undertaken for the War Industries
Board in 1919 produced occupational pay rates by
industry and State, and (for some industries) by
city. Between 1934 and 1940, the selection of
industries studied was determined largely by
administrative needs under the National Recovery
Act, Public Contracts Act, and the Fair Labor
Standards Act, with emphasis on nationwide data
for relatively low-wage industries.
Survey activity shifted in the 1940-41 defense
period to heavy industries essential to war pro­
duction. Implementation of wage stabilization
policy during the war required a large-scale pro­
gram of occupational wage studies by industry
and locality. The emphasis on data by locality
has continued since 1945 within the framework of
industry studies generally designed to also yield
national and regional estimates. In addition, the
Bureau developed two new types of surveys.
Community wage surveys, initiated in the late
1940’s, were designed to meet the growing demand
for pay data related to office clerical and manual
jobs that are common to a wide variety of manu­
facturing and nonmanufacturing industries
within metropolitan areas. This survey program
was firmly established and temporarily expanded
for use in the wage stabilization effort during the
Korean emergency. The need for nationwide
estimates of white-collar pay in private industry
for use in appraising the Federal white-collar
salary structure resulted in a survey design that
would produce national averages, based on an area
sample. Data for individual areas studied also

serve the needs of the U.S. Department of De­
fense and other agencies in setting rates for “wageboard” employees.
Prior to 1960, studies in a very few professions
provided salary data. Beginning in that year,
salary surveys have been made on a nationwide
basis covering professional, administrative, and
technical jobs in a broad spectrum of industries.
Averages for these jobs, together with national
averages for clerical and drafting jobs included
in the community wage surveys, are utilized by the
administrative agencies directly concerned with
Federal pay matters.

Description of Surveys
Although differing in industrial, geographic,
and occupational coverage, the 3 types of surveys
described form an integrated program of occupa­
tional wage surveys based upon a common set of
administrative forms, manual of procedures, and
common concepts and definitions. Employer co­
operation in surveys is on a voluntary basis. Con­
fidential individual establishment data compiled
by the Bureau’s field economists are grouped in
published reports in a manner that will avoid pos­
sible disclosure of an establishment’s rates.
Establishments included in all surveys are classi­
fied by industry as defined in the 1957 edition of
the Standard Industrial Glassification Manual
prepared by the U.S. Bureau of the Budget.1
Survey reports identify the minimum size of estab­
lishment (measured by total employment) studied.
Definitions for Standard Metropolitan Statistical
Areas are employed in all programs.2
Industry wage surveys provide data for occupa­
tions selected to provide representativeness of the
range of rates, methods of wage payment, and of
men’s and women’s work activities. Consideration
also is given, in their selection, to the prevalence in
the industry, definiteness and clarity of duties, and
1 See appendix B.
2 See appendix C.


importance as reference points in collective
In addition to collecting straight-time first-shift
rates (or hours and earnings for incentive work­
ers) for individual workers in the selected occupa­
tions, surveys in most industries also establish the
wage frequency distribution for broad employ­
ment groups, i.e., production and related workers
or nonsupervisory workers. Weekly work sched­
ules, shift operations and differentials, paid holi­
day and vacation practices, and health, insurance,
and pension benefits are included in the informa­
tion collected, along with the provisions made for
other items, applicable to certain industries. The
studies also provide estimates of labor-management
agreement coverage, proportions employed under
incentive pay plans, and the extent to which estab­
lishments provide a single rate or range of rates
for individual job categories.
Fifty manufacturing and 20 nonmanufacturing
industries, accounting for about 20 million em­
ployees, are surveyed on a regularly recurring
basis. A majority are studied on a 5-year cycle,
but many comparatively low-wage industries are
on a 3-year cycle.
Nearly all of the manufacturing, utilities, and
mining industries are studied on a nationwide
basis and estimates are also provided for regions
and major areas of concentration. Surveys in
trade, finance, and service industries usually are
limited to a score or more of metropolitan areas.
Nationwide surveys generally develop separate
estimates by size of establishment, size of com­
munity, labor-management agreement coverage,
and type of product or plant group.
Community wage surveys provide data for
occupations common to a wide variety of indus­
tries in the communities surveyed. The more than
three-score occupational categories studied are
about equally divided between (1) office clerical,
draftsmen, and industrial nurses, and (2) main­
tenance, toolroom, power plant, and custodian and
material-movement jobs. Thus, they provide
representation of the range of duties and respon­
sibilities associated with white-collar, skilled main­
tenance trades, and other “indirect” manual jobs.
Weekly salaries collected for individuals in whitecollar jobs relate to regular straight-time salaries
that are paid for standard workweeks. Average
hourly earnings for maintenance and other manual
jobs relate to first shift hourly rates.


Industry divisions included are (1) manufac­
turing, (2) transportation, communication, and
other public utilities, (3) wholesale trade, (4)
retail trade, (5) finance, insurance, and real estate,
and (6) selected service industries. Establish­
ments employing fewer than 50 workers are ex­
cluded—with a minimum of 100 applying to
manufacturing, transportation, communication
and other public utilities, and to retail trade in the
dozen largest communities.
In addition to the all-industry averages and
distributions of workers by earnings classes, sepa­
rate data are provided for manufacturing and
nonmanufacturing in each area and, wherever pos­
sible, for individual industry divisions in the
nonmanufacturing sector. Among the 80 Stand­
ard Metropolitan Statistical Areas in this annual
survey program as of 1965, separate data are pro­
vided for transportation, communication, and
other public utilities in 72 areas, for retail trade
in 23 areas, for wholesale trade and finance, insur­
ance, and real estate in 16 areas, and for the
selected service industries in 8 large areas.
Data on weekly work schedules, paid holiday
and vacation practices, and health, insurance, and
pension benefits are recorded separately for non­
supervisory office workers and plant (nonoffice)
workers. Shift operations and differentials are
collected for plant workers in manufacturing.
Data on minimum entrance rates for inexperienced
office workers are collected in all industries. All
of these items are studied annually in six large
areas and biennially in other areas. This survey
program also has developed information on profitsharing plans, characteristics of sick leave plans,
wage payment systems, and other items related to
employee compensation.
The National Survey of Professional , Adminis­
trative, Technical, and Clerical P ay provides a
fund of broadly based information on salary levels
and distributions in private employment. The 75
occupation-work levels studied were selected from
the following fields: Accounting, legal services,
office services, personnel management, engineering
and chemistry, drafting, and clerical. Definitions
for these occupations provide for classification of
employees according to appropriate work levels
(or classes). Although reflecting duties and re­
sponsibilities in industry, the definitions were
designed to be translatable to specific pay grades
in the general schedule applying to Federal Classi­



fication Act employees. This survey, thus, pro­
vides information in a form suitable for use in
comparing the compensation of salaried employees
in the Federal civil service with pay in private
Average salaries, monthly and annual for all
occupations and also on a weekly basis for clerical,
relate to the standard salaries that were paid for
standard work schedules, i.e., to the straight-time
salary corresponding to the employee’s normal
work schedule, excluding overtime hours. Salary
distributions and averages are presented for men
and women combined. Since the clerical salary
data largely are derived from the community wage
survey program, separate estimates for men and
women are made available for metropolitan areas
studied in the program.
Industry divisions included are: (1) manufac­
turing, (2) transportation, communication, elec­
tric, gas and sanitary services, (3) wholesale trade,
(4) retail trade, (5) finance, insurance, and real
estate, and (6) engineering and architectural serv­
ices, and commercially operated research, develop­
ment, and testing laboratories.
Limited to the Nation’s metropolitan areas for
the years 1960 through 1964, the annual survey
was expanded in 1965 to include nonmetropolitan
counties. The minimum establishment size in­
3 The terms “ in scope” or “ within scope” are used throughout
this chapter to refer to the coverage of the particular survey
bein g d escribed.

4 An example of a job description :
Produces replacement parts and new parts in making repairs
of metal parts of mechanical equipment operated in an estab­
lishment. Work involves most of the following: Interpreting
written instructions and specifications; planning and laying out
of w ork; using a variety of machinist’s handtools and precision
measuring instruments ; setting up and operating standard ma­
chine to o ls; shaping of metal parts to close tolerances; making
standard shop computations relating to dimensions of work,
tooling, feeds, and speeds of m achining; knowledge of the work­
ing properties of the common m etals; selecting standard mate­
rials, parts, and equipment required for his w ork; and fitting
and assembling parts into mechanical equipment. In general, the
machinist’s work normally requires a rounded training in ma­
chine-shop practice usually acquired through a formal apprentice­
ship or equivalent training and experience.
5 In general, workers are included in a classification if the
duties as described are performed a major part of the time and
the remainder is spent on related duties requiring similar or
lesser skill and responsibility. However, in some jobs, particu­
larly office and skilled production-worker categories, workers may
regularly perform a combination of duties involving more than
one occupation. Unless indicated otherwise in the description,
in these situations consideration for classification purposes is
given to those elements of the job which are most important in
determining its level for pay purposes. Thus, a worker meets
the basic concept of the stenographer classification if taking of
dictation is a regular requirement of the job even though a
majority of time is spent on routine typing.

eluded in the survey was raised from 100 to 250 in
1961. Since the survey scope is subject to change,
users are directed to the Scope and Method of
Survey appendix in the reports for a description
of current practice.3
Concepts. The Bureau’s occupational wage sur­
veys summarize a highly specific wage measure,
the rate of pay, excluding premium pay for over­
time and for work on weekends, holidays, and late
shifts, for individual workers. In the case of
workers paid under piecework or other types of
production incentive pay plans, an earned rate is
computed by dividing straight-time earnings for
a time period by corresponding hours worked.
Production bonuses, commissions, and cost-ofliving bonuses are counted as earnings. In gen­
eral, bonuses that depend on factors other than
the output of the individual worker or group of
workers are excluded; examples of such nonpro­
duction payments are safety, attendance, year-end
or Christmas bonuses, and cash distributions under
profit-sharing plans.
Bates do not include tips or allowances for the
value of meals, room, uniform, etc. The earn­
ings figures, thus, represent cash wages (prior to
deductions for social security, taxes, savings bonds,
premium payments for group insurance, meals,
room or uniforms) after the exclusion of premium
pay for overtime, weekend, holiday, or late shift
Hours shown for salaried occupations relate to
standard weekly hours for which the employee
receives his regular straight-time salary.
Occupational classifications are defined in ad­
vance of the survey. Because of the emphasis on
interestablishment and interarea comparability of
occupational content, the Bureau’s job descriptions
may differ significantly from those in use in in­
dividual establishments or those prepared for other
purposes. The job descriptions used for wage
survey purposes are typically brief and usually
more generalized than those used for other pur­
poses. The primary objective of the descriptions
is to identify the essential elements of skill, diffi­
culty, and responsibility that establish the basic,
concept of the job.4
It should be recognized that although work ar­
rangements in any one establishment may not cor­
respond precisely to those described, those workers
meeting the basic requirements established for the
job are included.5


In applying these job descriptions, the Bureau’s
field representatives exclude working supervisors,
apprentices, learners, beginners, trainees, handi­
capped workers, part-time or temporary workers,
and probationary workers.
Paid holidays, paid vacations, and health, in­
surance, and pension plans are treated statistically
on the basis that these are applicable to all nonsupervisory plant or office workers if a majority
of such workers are eligible or can expect even­
tually to qualify for the practices listed. Data for
health, insurance, and pension plans are limited
to those plans for which at least a part of the
cost is borne by the employer. Informal pro­
visions are excluded.

Survey Methods
Planning . Consultations are held with appro­
priate management, labor, and Government rep­
resentatives to obtain views and recommendations
related to scope, timing, selection, and definitions
of survey items, and types of tabulations. Partic­
ularly in planning surveys in specific industries,
these discussions importantly supplement com­
ments and suggestions received from the regional
offices at the conclusion of the previous study. Re­
flecting its use in evaluation of Federal white-col­
lar pay, the design of the National Survey of Pro­
fessional, Administrative, Technical, and Clerical
Pay was developed in conjunction with the Bureau
of the Budget and the Civil Service Commission.
Changes in the survey scope, item coverage, and
job definitions are initiated by these agencies.
The industrial scope of each survey is identified
in terms of the classification system provided in
the Standard Industrial Classification Manual .
The scope may range from part of a 4-digit code
for an industry study to a uniform combination
of broad industry divisions and specific industries
for the community wage surveys or the salary sur­
vey of professional, administrative, technical, and
clerical j obs. The needs of maj or users are a maj or
consideration in designing the multi-purpose
occupational studies.
The minimum size of establishment included in
a survey is set at a point where the possible con­
tribution of the excluded establishments is re­
garded as negligible for most of the occupations


surveyed. Another practical reason for the adop­
tion of size limitations is the difficulty encountered
in classifying workers in small establishments
where they do not perform the specialized duties
indicated in the job definitions.
Considerations in timing of industry surveys in­
clude date of expiration of major labor-manage­
ment agreements, deferred wage adjustments,
seasonality of production (e.g., garments), and
interests of users. Wherever possible, community
wage surveys are timed to follow major wage set­
tlements as well as to meet the needs of govern­
ment agencies engaged in wage administration as
required by law.
The types of occupations studied and criteria
used in their selection were identified in the de­
scription of the various types of surveys. The job
list for each survey is selected to represent a rea­
sonably complete range of rates in the wage struc­
ture for the employment categories involved, i.e.,
production and related workers in a specific manu­
facturing industry or nonsupervisory office, main­
tenance, material handling, and custodial workers
in a metropolitan area. The established hierarchy
of job rates to be found within establishments and
industries permits the use of pay data for such
key or benchmark jobs for interpolating rates for
other jobs. Technological developments or user
interests may dictate changes in the job lists and
definitions. New definitions for jobs usually are
pretested in a variety of establishments prior to
their use in a full-scale survey.
Questionnaires. Two basic schedules are used
in obtaining data in all surveys. The first (B LS
2751A) includes items relating to products or serv­
ices, employment, shift operations and differen­
tials, work schedule, overtime premiums, paid holi­
days and vacations, insurance and pension plans,
union contract coverage, and other items applica­
ble to the establishment. The second (BLS
2752A) is used in recording occupation, sex,
method of wage payment, hours (where needed),
and pay rate or earnings for each worker studied.
Supplementary forms are used to meet particular
Collection. Bureau field economists collect data
by personal visit to each of the sample establish­
ments. Job functions and factors in the establish­
ment are carefully compared with those included
in the Bureau job definitions. The job matching
may involve review of records such as pay struc-



B IS2751A

+ expires
» N»9 -3 0 -6 6


Y our report w ill b e
h e ld in c o n fid e n c e





Establishm ent Sch edu led

Payroll period
196 ___ .

1 96 ___ .

(Street address)

1 96___ .

(County and State)
Name and title of authorizing official

Name and title of official supplying data

1 96 ___
1 96 ___
1 9 6 __

(Address of office from which information was obtained, if different from above)

B. C en tral O ffice (if an y)


(Street address)

(City and State)

A . Product or S ervice (to be used to assign industry classification)
Aporoxi mate

Approxi mate

% annual

1 96_____


% annual

196 _____

1 96 _____

% annual

B. Sco pe of O p eration s

3. Office Use Only
Schedule No.






6 -7


9 -1 0



SIC code




Total employment





2 4 -28



(2 >

Nonsupervisory production (plant) workers

Nonsupervisory office workers

Total employment






Other .(executive,
sup ervisory, etc.)






If the establishment lacks either production (plant) or office workers, leave corresponding spaces blank on a ll items.
A , Shift D a ta — Production (plant) w orkers only
1. Shift operation:
Does establishment operate extra shift?
Code block for each shift independently
Codes (Blocks 44 and 45)
1 . Establishment currently operating extra shift
2« Establishment not currently operating extre shift but has formal shift premium provisions
3 . Establishment not currently operating extra shift and has no formal shift premium provisions


2 . Shift employment (approximate)— To be answered only if extra shift is being operated:

D ay

Night (3d)

Evening (2d)

1 9 6 ___________________________ __________________________________ _____________________________
1 9 6 ___________________________ ___________________________________


19 6 ___________________________ _



Office Use Only— Weighted Shift Employment





4 6 -4 9

5 0 -53

3d Shift
(or other)

2d Shift





B L S 2752A

B o No. 4 4 -R 1 0 5 1 .
Approval exp ires 9 -3 0 - 6 6


Y our report w ili b e
h eld in c o n fid e n c e

Industry or survey

Establishm ent nam e

A rea _______________________________________________________________________________________________________________________________


Payroll period


(also including cost-of-living and annual improvement adjustments)


Approximate number
workers affected


(if any)




Production (plant)




Classes of workers
affected by
wage change



W ag e changes granted since

W a g e changes pending to______________________________________________________ 1 9 6 _____




Production (plant) workers:

Yes □

Office workers:

Yes |

Annual Improvement

No □

No |

Y „ Q

Yes |

No □

No |





Production (plant) workers --------------------------

Office w o rkers______________










Number of

Production (plant)


Card count








ture plans and organizational charts, company
position descriptions, interviews with appropriate
officials, and, on occasion, observation of jobs
within plants. A satisfactory completion of job
matching permits acceptance of company-pre­
pared reports where this procedure is preferred
by the respondent. Generally, however, the field
economist secures wage or salary rates (or hours
and earnings, when needed) from payroll or other
records and data on the selected employer prac­
tices and supplementary benefits from company
officials, company booklets, and labor-management
Community wage surveys in all except the large
metropolitan areas involve personal visits every
second year with partial collection by mail in the
intervening years. Establishments participating
in the mail collection receive a transcript of the
job matching and wage data obtained a year ear­
lier by the field economist, together with the job
definitions. The up-dated returns are scrutinized
and questionable entries are checked with the re­
spondent. Personal visits are made to establish­
ments not responding to the mail request and to
those reporting unusual changes from year-earlier
The work of all field economists is checked for
quality of reporting, with particular attention di­
rected to accuracy in job matching. The revisits
are made by supervisory and senior economists.
Systematic technical audits of the validity of sur­
vey definitions, made by staff with specialized
training, also are maintained for the technically
complex nationwide white-collar salary survey.

The sampling design employed is almost always
highly stratified. Before the sample is selected,
information on all known establishments that
might possibly fall within the scope of the survey
is compiled from lists provided by regulatory gov­
ernmental agencies (primarily State unemploy­
ment insurance agencies), supplemented by data
from trade directories, trade associations, labor
unions, and other sources.
Establishments then are stratified as precisely as
available information permits. Each geographicindustry unit for which a separate analysis is to


be presented is sampled independently. Within
these broad groupings, a finer stratification by
product (or other pertinent attributes) and size of
establishment is made. Stratification may be car­
ried still further in certain industries: textile mills,
for instance, are classified on the basis of integra­
tion, i.e., whether they spin only, weave only, or
do both. Such stratification is highly important
if the occupational structure of the various industry segments differs widely.
The sample for each industry-area group is a
probability sample, each establishment having a
predetermined chance of selection. In order to
secure maximum accuracy at a fixed level of cost
(or a fixed level of accuracy at minimum cost),
the sampling fraction used in the various strata
ranges downward from all large establishments
through progressively declining proportions of
the establishments in each smaller size group, in
accordance with the principles of optimum alloca­
tion. Thus, each sampled stratum will be repre­
sented in the sample by a number of establish­
ments roughly proportionate to its share of the
total employment. Though this procedure may
appear at first to yield a sample biased by the overrepresentation of large firms, the method of esti­
mation employed avoids the possibility of bias by
the assignment of proper weights to the sample
In the event a sample establishment within
scope is unable to supply usable data, a substitute
is assigned in the same industry-location-size class.
(Since no close relation exists between failure to
participate in these surveys and the items being
studied, little bias is introduced by this procedure.)
The overall nonresponse rate in published surveys
averages about 5 percent, and in exceptional cases
reaches 10 percent.
The size of the sample in a particular survey
depends on the size of the universe, the diversity
of occupations, and their distribution, the relative
dispersion of earnings among establishments, the
distribution of the establishments by size, and the
degree of accuracy required. Estimates of vari­
ance based on data from previous surveys are used
in determining the size of the sample needed.
As indicated earlier, community wage surveys
are limited to selected metropolitan areas. These
areas, however, form a sample of all such areas,
and, when properly combined, yield estimates of
the national and regional levels.



The sample of areas is based on the selection of
one area from a stratum of similar areas. The
criteria of stratification are region, type of indus­
trial activity as measured by percentage of manu­
facturing employment, major industries, and level
of earnings in manufacturing. Each area was se­
lected with probability proportionate to its nonagricultural employment.
The largest metropolitan areas are self-repre­
senting, i.e., each one forms a stratum by itself and
is certain of inclusion in the area sample. The
area sample contained about 80 percent of all nonagricultural employment of the metropolitan area
complex of the entire country in 1964.
In some strata it was impossible to give some
areas their proper chance of selection, because of
difficulties in making surveys in these areas pri­
marily due to predominance of single employers.
Then, too, some areas were chosen to represent some
strata because of interest in the area of itself.
These departures from a strict probability design,
covering only 14 percent of the total, are believed
to be negligible in their effect.

Estimating Procedures
Estimated average earnings (hourly, weekly,
monthly, or annual) for an industry or an occupa­
tion are computed as the arithmetic mean of the
individual employees5 earnings. They are not
estimated by dividing total payrolls by the total
time worked, since such information almost never
is available on an occupational basis.
All estimates are derived from the sample data.
The averages for occupations, as well as for in­
dustries, are weighted averages of individual earn­
ings and not computed on an establishment basis.
The proportion of employees affected by any
fringe provision likewise is estimated from the
sample; all workers in each establishment are con­
sidered to be covered by the predominant benefit
policy in effect, and the entire employment of the
establishment is classified accordingly.
As mentioned previously, the use of a variable
sampling ratio in different strata of the population
would result in biased estimates if straight addi­
tion of the data for the various establishments were
made. Therefore, each establishment is assigned
a weight that is the inverse of the sampling rate
for the stratum from which it was selected—e.g., if

a third of the establishments in one stratum are
selected, each of the sampled establishments is
given a weight of 3.
To illustrate the use of weights, suppose the uni­
verse were 7 establishments, from which a sample
of 3 was selected. Assume that establishment A
was drawn from a cell, or stratum, in which half of
the plants were used in the sample. It therefore,
is, given a weight of 2. Establishment B, on the
other hand, was taken with certainty (or a prob­
ability of 1) and is thus given a weight of 1.
Establishment C was taken from a group where a
fourth of all plants were used in the sample, and
hence is given a weight of 4. The following cal­
culations are made in estimating average earnings
for a given occupation.
Workers in occupation
in sample establish­
ments at specified rate



B ..................


Estimated universe




Estimates of total in stratum





1X30X 1.70
1X20X 1.95
4X10X 1.20
$258. 00

The estimated average hourly earning is thus
$258. 00 or $1.52.
A similar method applies to any characteristic
estimated from the sample. To estimate the pro­
portion of employees in establishments granting
paid vacations of 2 weeks after 2 years of service,
for instance, the establishments are classified ac­
cording to the length of vacation granted after 2
years’ service, establishment weights are applied
to employment, as in the previous example, and
the proportion of the estimated employment in
the 2-week category of the estimated total employ­
ment then is computed. Using the same three
establishments as in the previous example, this can
be illustrated as follows:


Actual total

B _________
Estimated universe.......... ............ .....................


after 2 years

200 1week.
500 2 weeks.
300 1week.

Thus, the estimated percentage of workers in
establishments granting 2 weeks’ vacation after 2
years of service is j- qqq or 50 percent.


When a large establishment within survey scope,
for which no substitute exists, is unable to supply
data, the deficiency is alleviated by increasing the
weight of the most nearly similar units. Should
any segment be affected by a substantial amount of
such noncooperation, the publication of materials
will be diminished by omitting separate presen­
tation of sectors seriously affected.
Where a sample of selected metropolitan areas
is used to represent the totality of such areas, a
second stage of weighting is used to expand the
individual area totals to region and/or national
estimates. Since, as indicated in the description
of the sampling method, each area represents a
stratum of similar areas, the total from each area
are weighted to the estimated stratum totals by
multiplying by the inverse of the chance of selec­
tion. This procedure provides the ratio of nonagricultural employment in the stratum to that
in the sample area (one in the case of the large
self-representing areas). Summing all such esti­
mated stratum totals yields the earnings and em­
ployment totals for the region and the country as
a whole.

Analysis and Presentation
Where an industry survey is designed to yield
estimates for selected States or areas, these are
published separately as information becomes avail­
able from all sample firms in the State or area unit.
Industry surveys limited to selected areas do not
provide a basis for the examinations of pay levels
by size of community, size of establishment,
product, or labor-management agreement cover­
age that generally are included in bulletin reports
on nationwide surveys. Eegardless of geographic
scope, industry survey reports record the incidence
of incentive pay plans and, to the extent possible,
average pay levels separately for time and
incentive workers.
Individual bulletin reports on individual com­
munity wage surveys are supplemented by a twopart summary bulletin. Part I compiles the re­
sults of individual area surveys made during a
fiscal year. Part II contains information on oc­
cupational earnings, employer practices, and sup­
plementary wage benefits for all metropolitan
areas combined and by industry division within
the four census regions.
222-617— 66----- 9


Wage-rate indexes are constructed for broad
occupational categories, e.g., office clerical work­
ers, skilled maintenance workers, and unskilled
plant workers. These indexes are published an­
nually, separately for all industries and manu­
facturing, for each metropolitan area studied,
and for all metropolitan areas combined and by
region. Area pay relatives for the three occupa­
tional categories are published annually, permit­
ting ready comparisons of average pay levels
among areas. Occupational pay relationships
within individual establishments are summarized
periodically, as are estimates of labor-manage­
ment agreement coverage.
Bulletin reports on the National Survey of
Professional, Adminstrative, Technical, and Cler­
ical Pay present occupational averages and dis­
tributions on an all-industry basis, nationwide and
separately for all metropolitan areas combined.
Average pay levels for industry divisions are
shown as percentages of the all-industry averages.
Salary trend estimates are published for each occu­
pation-work level. Special inquiries have devel­
oped estimates of the value of cash bonuses and
information on practices in the design and use
of salary structures relating to white collar
Industry and community wage survey reports
are issued throughout the year as the surveys are
completed. The bulletin on the National Survey
of Professional, Admistrative, Technical, and
Clerical Pay is made available in November.
Summaries of the data in the bulletins and spe­
cial analyses appear also in the Monthly Labor

Uses and Limitations
Occupational wage data developed in these sur­
veys have a variety of uses. They are used by
Federal, State, and local agencies in wage and
salary administration and in the formulation of
public policy on wages, as in minimum wage legis­
lation. They are of value to Federal and State
mediation and conciliation services and to State
unemployment compensation agencies in judging
the suitability of job offers. Knowledge of levels
and trends of pay rates by occupation, industry,
locality, and region is required in the analysis of



current economic developments and in studies re­
lating to wage dispersion and differentials.
Bureau data are used in connection with private
wage or salary determinations by employers or
through the collective bargaining process. To the
extent that wages are a factor, survey data also
are considered by employers in the selection of
location for new facilities and in cost estimating
related to contract work.
Occupational wage survey programs are not
designed to supply mechanical answers to ques­
tions of pay policy. As suggested earlier, limita­
tions are imposed in the selection and definition of
industries, of geographic units for which estimates
are developed, of occupations and associated items
studied, and in determination of periodicity and
timing of particular surveys. Depending upon
his needs, the user may find it necessary to in­
terpolate for occupations or areas missing from
the survey on the basis of knowledge of pay
Because of interestablishment variation in the
proportion of workers in the jobs studied and in
the general level of pay, the survey averages do
not necessarily reflect either the absolute or rela­
tive relationships found in the majority of estab­
lishments. To illustrate, employment in the spe­
cialized maintenance crafts tends to be concen­
trated in the larger establishments, whereas em­
ployment in custodial and material-movement jobs
is distributed more widely within an industry or
area. Thus, to the extent that pay rates in the
larger establishments vary from the average level,
the skill differential measure based on the survey
averages will differ to some degree from that ob­
tainable within each of the larger establishments.
The incidence of incentive methods of payment
may vary greatly among the occupations and
establishments studied. Since hourly averages
for incentive workers generally exceed those for
hourly-rated workers in the same job, averages for
some incentive-paid jobs may equal or exceed aver­
ages for jobs positioned higher on a job evaluation
basis but normally paid on a time basis. Whereever possible, data are shown separately for time
workers and incentive workers in the industry sur­
veys. Incentive plans (generally plant-wide in
application) apply to only a very small proportion
of the workers in the indirect plant jobs studied
in the community wage program.

Although year-to-year changes in averages for
a job or job group primarily reflect general wage
and salary changes or merit increases received by
individuals, these averages also may be affected by
changes in the labor force resulting from labor
turnover, labor force expansions and reductions
for other reasons, as well as changes in the propor­
tion of workers employed in establishments with
different pay levels. A labor force expansion
might increase the proportion of lower paid
workers and thereby lower the average, or the
closing of a relatively high-paying establishment
could cause average earnings in the area to drop.
Reliability of surveys. Results of the surveys
generally will be subject to sampling error. This
error will not be uniform, since, for most occupa­
tions, the dispersion of earnings among establish­
ments and frequency of occurrence of the occupa­
tion differ. In general, the sample is designed so
that the chances are 9 out of 10 that the published
average does not differ by more than 5 percent
from the average that would be obtained by enu­
meration of all establishments in the universe.
That error applies to the smallest breakdown pub­
lished. Hence, the error present in broader
groupings will be somewhat less.
The sampling error of the percentage of workers
receiving any given supplementary benefit differs
with the size of the percentage. However, the
error is such that rankings of predominant prac­
tices almost always will appear in their true posi­
tion. Small percentages may be subject to con­
siderable error, but will always remain in the same
scale of magnitude. For instance, the proportion
of employees in establishments providing more
that 4 weeks’ paid vacation to long-service em­
ployees may be given as 2 percent, when the true
percentage for all establishments might be only 1
percent. Such a sampling error, while consider­
able, does not affect the essential inference that
the practice is a rare one.
Estimates of the number of workers in a given
occupation are subject to considerable sampling
error, due to the wide variation among establish­
ments in the proportion of workers found in indi­
vidual occupations. (It is not unusual to find
these estimates subject to sampling error of as
much as 20 percent.) Hence, the estimated num­
ber of workers can be interpreted only as a rough
measure of the relative importance of various


occupations. The greatest degree of accuracy in
these employment counts is for those occupations
found principally in large establishments. This
sampling error, however, does not materially affect
the accuracy of the average earnings shown for the
occupations. The estimate of average earnings is
technically known as a “ratio estimate,” i.e., it is
the ratio of total earnings (not payrolls) to total
employment in the occupation. Since these two
variables are highly correlated (i.e., the errors
tend to be in the same direction), the sampling
error of the estimate (average hourly earnings) is
considerably smaller than the sampling error of
either total earnings or total employment.
Since completely current and accurate informa­
tion regarding establishment products and the cre­
ation of new establishments is not available, the
universe from which the sample is drawn may be
incomplete. Sample firms incorrectly classified


are accounted for in the actual field work, and the
universe estimates are revised accordingly. Those
firms which should have been included but were
classified erroneously in other industries cannot be
accounted for.
Since some measure of subjective judgment
enters into the classification of occupations and
other characteristics, there is some reporting vari­
ability in the results. A repetition of the survey
in any establishment with different interviewers
and respondents would undoubtedly produce
slightly different results. However, when spread
over a large number of establishments the differ­
ences, being random, would tend to balance out.
Hence, analyses based on a small number of re­
spondents must be used with care, even when all
eligible establishments are included. No evidence
of any consistent error has been uncovered.

Technical References

1. Cohen, Sam uel E ., “ Studies o f O ccupational W a g e s an d S u pplem entary Benefits.”

M o n th ly L a b o r

R eview , M a rc h 1954 (p p . 292-297).
A n earlier description of the m ethods o f w age surveys, sim ilar to the present article.
2. D o u ty , H . M ., “ S u rvey M e th o d s an d W a g e C om parison s.”

L a b o r L a w J o u r n a l, A p ril 1964 (p p .

A discussion o f the uses of w age su rve y results, an d the pitfalls to be avoided.

A short

discussion of the factors affecting s u rv ey m ethods is also included.
3. K an ninen, T o iv o P ., “ N e w

Dim ensions in B L S

W a g e S u rvey W o r k .”

M o n th ly L a b o r Review ,

O ctober 1959 (pp. 1081-1084).
A n outline of the occupational w age su rve y program s, as exp an d ed in fiscal 1960. Lists the
typ e o f su rv e y a n d cycle fo r each o f 70 industries studied separately, an d identifies the area sam ple
as origin ally determ ined fo r the la b o r m arket su rvey p rogram .

—Toivo P.

K a n n in e n

Office o f W a g e s an d In d u stria l Relations

Chapter 15. Employee Earnings and Hours Frequency Distributions
An extensive program of studies of the fre­
quency distribution of employee earnings has been
maintained by the Bureau of Labor Statistics since
1954, and since 1964 frequency distributions of
weekly hours of work have been a part of this pro­
gram. The program of studies was initiated to
provide knowledge about the internal structure
of earnings and hours which an average does not
reveal. For example, distribution studies show
whether earnings or hours are dispersed evenly
or tend to be clustered around one or more points;
they show the gap between the highest and low­
est paid workers; they reveal the proportions of
employees working short or long hours, and similar
information on individual earnings and hours.
They serve to locate the employed “poor,” i.e., those
not sharing in the general rise in living standards
available from work.
Although the need for such data had long been
recognized and some work had been done in the
area, the program did not receive its full impetus
until 1955 when Congress, in an amendment to sec­
tion 4(d) of the Fair Labor Standards Act, di­
rected the Secretary of Labor to include in his
annual report an evaluation and appraisal of mini­
mum wages established by the Act and his legisla­
tive recommendations.1 To meet these require­
ments, the program of employee earnings
distributions studies was established.

Studies of employee earnings and hours distribu­
tions generally include within their scope all nonsupervisory employees. No attempt is made to
classify employees by occupation, although at times
information is collected separately for some types
of employees. The data collected relate to straighttime hourly earnings, excluding premium pay for
1 The Bureau’s studies conducted in this area prior to 1954
frequently related to selected occupations, or were limited to
relatively narrowly defined industries.
2See technical references for a listing of publications con­
taining full descriptions of both types of studies.
3See appendix B.
4See appendix C.


overtime work and for work on weekends, holi­
days, and late shifts. Incentive pay, production
bonuses, and cost-of-living payments are included
in earnings, but nonproduction bonuses (e.g.,
Christmas bonuses) are not. Hours information
is collected for hours worked during the week and
for hours spent on vacations, holidays, or sick leave
for which pay is received. Earnings and hours
information is obtained for each employee, making
it possible to calculate average hourly earnings
and weekly hours of work for the entire group as
well as to show the distribution of employees by
earnings, hours, or both.
Two basic types of distributions studies are con­
ducted—industry and area.2 Industry studies may
have broad coverage, such as manufacturing or
wholesale trade, or they may be limited to specific
industries, such as motor carriers. While industry
studies usually are conducted on a nationwide
basis, they sometimes (especially in the case of
specific industry studies) are limited to one or
several areas in which the industry is concentrated.
Industries are defined on the basis of descriptions
in the current Standard Industrial Classification
Manual, published by the Bureau of the Budget.3
Area studies are limited in geographic coverage,
perhaps to a Standard Metropolitan Statistical
Area,4 to a county or a group of contiguous coun­
ties, or to a region or part of a region. This type
of study usually is conducted on a cross-industry
basis—that is, earnings and hours data are col­
lected for most industry divisions, including min­
ing; manufacturing; transportation (except rail­
roads) , communications and other public utilities;
wholesale and retail trade; finance, insurance, and
real estate; and services. Minor exceptions may
be made.

Data Sources and Collection Methods
Data generally are obtained from employer pay­
roll records, using one of two collection methods.
Establishments may be solicited for information
by mail or by personal visit. Those establishments
contacted by mail receive a questionnaire form and
a note requesting cooperation in the survey and


explaining the nature and purpose of the study.
Typically, the questionnaire requests information
on the product or service of the establishment and
number of employees, as well as other establish­
ment information which may be pertinent to the
survey.5 The respondent is requested to report
each employee’s earnings and weekly hours
worked—the primary focus of the survey. Pro­
vision is made for the reporting of earnings on an
hourly basis or on a salary or incentive basis. In
the last two instances, the respondent reports the
number of hours paid for during the salary or in­
centive period, thus permitting the computation
of average hourly earnings for employees paid on
other than an hourly basis.
Certain large establishments, or companies from
which data for several establishments are re­
quested, are visited personally by the Bureau’s field
economists who may prepare the data for the sur­
vey from company records, or arrange with the
company for the completion of the questionnaire
form. In addition, a sample of establishments
which do not reply to the mail questionnaire is
visited by Bureau field economists to obtain the
desired information.
To limit errors caused by incorrect reporting by
the respondent, questionnaires received by mail
are reviewed for reasonableness and consistency.
Where data are questionable, a letter is written to
the respondent asking him to review the item to
assure its accuracy.

Earnings and hours distribution studies are
conducted on the basis of a sample of all establish­
ments within the scope of the survey. The sample
generally is derived from State Unemployment
Insurance (UI) listings which show reporting
units with four employees or more by location,
number of employees, and industry classification.
In industries where establishments with fewer
than 4 employees are of numerical importance
(e.g., retail trade) the U I lists may be aug­
mented by sources such as other government
agencies, or trade directories.
The size of the sample depends on several fac­
tors, among which are the size of the universe, the
5 See illustrations, pp. 128-133.


distribution of establishments by number of em­
ployees, the relative dispersion of earnings among
establishments, the degree of accuracy required,
and the cost of obtaining the data. Estimates of
variance based on data from previous surveys also
may be used in determining the appropriate size
of the sample.
The sample usually is selected using a highly
stratified probability sampling design. Establish­
ments are first grouped, or stratified, according to
industry, geographic location, and employment
size. Establishments in specific industries or areas
for which earnings and hours data are to be pre­
sented separately are grouped independently of
establishments in other strata, and sampled
In order to obtain maximum accuracy per unit
cost, the number of establishments in the sample
is distributed among the various strata in the most
efficient manner, in accordance with the principles
of optimum allocation. A fraction of establish­
ments in each stratum is included in the sample,
with the sampling fraction diminishing as the
employment size of the strata decreases, so that
the probability of inclusion in the sample is greater
for the large than for the small establishment.
Frequently the entire stratum containing the larg­
est establishments is included in the sample.
The following example shows a hypothetical
universe of 47 establishments in three employmentsize groups. By applying the appropriate sam­
pling ratio to each size group (stratum), the
number of sample establishments is determined.
size group

Under 25...............
50 and over.......... .

Number of
in universe


of sample

Weight of
each sample


1 /



1 3

No assumption is made that the earnings and
hours structures of establishments not responding
to the mail questionnaire are similar to those of
establishments which do respond. Therefore, a
sample is taken of the nonrespondents following
a procedure similar to the one just described. E s­
tablishments in this subsample are visited by
Bureau field representatives in order to obtain the
required data.
Data are not always obtained for every estab­
lishment in the original sample. Generally, ap­
proximately 60 to 70 percent of the sample
establishments supply usable data to the survey.


Budget Bureau No. 4 4 -R 1 0 0 6 .5
Approval expires March 31, 1965

BLS 2834
Your report w ill be
held in confidence.

W a s h in g t o n ,

D.C. 20210



This report should cover m anufac­
turing plants, warehouses, and ceiw
tral o ffices in the county or area
designated to the left.


List separately all products and services for the establishment designated above.
The products listed should account for 80 percent or more of last y e a r’ s sales.
Include in sales all receipts of nonmanufacturing activities, if any.
percent of



Approxim ate
percent of


The information requested should -correspond to your payroll period (weekly, bi­
w eekly, or monthly) which includes March 12, 1964.
Indicate the dates fo.r the payroll period used.
F ro m ______________________________ , 1964 t o ________________________________ , 1964.




___________ .

Enter to tal number of em ployees (full-tim e and part-tim e) who received pay for any part of the payroll period.
Include nonsupervisory, executive, adm inistrative, and professional em ployees. DO NOT INCLUDE proprietors,
members of unincorporated firm s, or unpaid fam ily workers.


Nonsupervisory production and related employees_______________

___________ .

Enter number of em ployees (fu ll-tim e and part-tim e) who were engaged in production jobs below the supervisory
le v el and received pay for any part of the payroll period. Include em ployees engaged in fabricating, processing,
assem bling, inspecting, receiving, storing, handling, packing, warehousing, shipping, trucking, hauling, m ainte­
nance, repairing, jan ito rial, watchmen, product developm ent, auxiliary production for plant's own use (e. g . , powerplant), recordkeeping, and other services closely associated with the production operations. Also include working
forem en (who spend less than 20 percent of their tim e at supervisory duties), leadm en, and trainees.


Other nonsupervisory employees_________________________________

___________ .

Enter number of em ployees (fu ll-tim e and part-tim e) who were engaged in other than production jobs below the
supervisory lev el and received pay for any part of the payroll period. Include o ffice, tech nical, and cafeteria
em ployees, routemen, em ployees engaged in the installation and servicing of products m ade in the establishment,
and force-account construction em ployees engaged in construction of m ajor additions or alterations who are utilized
as a separate work force. DO NOT INCLUDE outside salesm en, executive, adm inistrative, professional, and su­
pervisory personnel.

Do you want a copy of the Bureau's report on this survey?_______ Yes I

I No |


Name and title of person furnishing data________________________________________________
(Please type or print)


R eg.







Indicate the premium overtime pay provision for production and other nonsupervisory employees in your establishment. If more than one provision exists,
indicate that which applies to the greatest number of employees. Enter ."none"
if such a provision does not exist.
(1) How many hours during a single day must an employee work before he is
eligible for premium overtime pay, and what is the premium rate (e. g. ,
8 hours - 1 V2 times regular rate)?
Production em ployees________________________________________________________ .
Other nonsupervisory employees ____________________________________________ .
(2) How many hours during a week must an employee work before he is eligible
for premium overtime pay and what is the premium rate (e. g. , 40 hours l l z times regular rate)?
Production em ployees________________________________________________________ .
Other nonsupervisory em ployees_____________________________________________.


How many of your production employees are paid on an incentive basis (e. g . ,
piecework, bonus)?______________________________________________


How many of your production employees were employed on late shifts during the
payroll period covered ?_________________________________________
How many received shift differentials?


(Please read carefully to avoid correspondence)

Separate sections are provided for reporting data for production and related workers
and for other nonsupervisory workers.
Earnings and hours should be reported separately for each employee unless these data are identical for two or more employees.
Exclude premium pay for overtime and for work on weekends, holidays, and late
shifts. Do not report aggregate earnings and hours for several workers. For convenience of reporting for employees paid on other than an hourly basis (e. g. , salary,
incentive), columns 5 and 6 are provided.
Instructions for reporting the necessary
data in each column are listed below and examples are shown on the enclosed sheet.
Column (1)— Indicate whether the employee is male (M) or female (F).
Com plete
colum ns 1, 2,
and 3 fo r a ll
non supervisory
em ployees
covered by th is

U se column 4
to report
earn in gs o f
em ployees
p a id on an
hourly b a s is .

U se colum ns 5
an d 6 to report
earn in gs o f
em ploy ees p a id
on a s a la r y or
in cen tiv e b a s is .

Column (2)— Use a separate line for each employee and enter "1, " unless
two or more employees of the same sex worked the same number of hours
during the selected week, and received identical hourly or salary rates.
Data are to be reported individually for each employee whose earnings
were based entirely or in part on commissions, bonuses, or incentives.
Column (3)— Enter the number' of hours paid for during the week of
March 9-15, 1964. Include hours paid for sick leave, holidays, vaca­
tions, etc. These hours should relate to a 1-week period regardless of
.the length of the payroll period.
Column (4)— Enter the base (straight-time) hourly’rate. Premium payments
for overtime and for work on weekends, holidays, and late shifts should
be excluded.
This column may also be used to report earnings of em­
ployees paid on an incentive or salary basis if average straight-time
hourly earnings are available.
Column (5)— Enter for each employee the total straight-time salary and/or
incentive earnings for the payroll period (weekly, biweekly, or semi­
monthly) which includes March 12, 1964. Include straight-time pay for
overtime, but exclude overtime premium.
Column (6)— Enter the number of hours paid for during the payroll period
(weekly, biweekly, monthly, or semimonthly) which corresponds to the
earnings reported in column 5. Include hours paid for sick leave, hol­
idays, vacations, etc.







' "
(as defined in 3C)
(as defined in 3B'
Use this .
Use this
column for
column for Use these columns
Use these columns
nonsuper­ for nonsupervisory
Com plete these columns
for nonsupervisory
C om plete these columns
visory em ­
visory em ­
for a ll nonsupervisory
for a ll nonsupervisory
em ployees paid
em ployees paid
other than on an
em ployees
other than on an
em ployees
paid on
paid on
hourly basis
hourly basis
an hourly
an hourly
Straight-tim e
Straigh t-tim e
salary or
salary or
paid for Straightincentive
paid for Straightincentive
paid for
paid for Sex
tim e
tim e
earnings for
earnings for
during (M or
(M or
payroll period
the week hourly
payroll period
the week
F) em ployees March
F> em ployees March 99which includes period
which includes period
15, 1964
March 12,
March 12,
15, 1964
_ (2)
_ (3)
..... (4)


BLS 2786

13 1
Budget Bureau No. 44—6615.
Approval expires 12-31-66.

(Rev. *66)

W A SH IN G T O N , D. C . 2 0 2 1 2

Your report will be
held in confidence



Individual Hours and Earnings



State Area sales SIC Wgt.

The d ata, ex cep t for Item 2 which r e la t e s to the
entire company, sh o uld cover a ll e sta b lish m e n ts
(re ta il sto r e s , w areh o u ses, central o f f ic e s , e t c .)
in the county or a re a d e sig n a te d to the left.

(C h eck appropriate b o x.)




#250,000 to

$500,000 to

or more

C h eck the block which in d ic a te s the annual g ro s s volume of s a l e s (e x c lu siv e o f e x c is e ta x e s a t the re ta il le v e l) fron. a ll
related a c tiv itie s o f the en te rp rise . Include re c e ip ts from sto re s co v ered by th is report a s w ell a s a ll other re la te d
a c tiv itie s . U se the la s t ca le n d a r or f is c a l year.



Please enter the information requested in the columns below for each separate establishment (retail store, warehouse, or
central office) covered by this report. Each retail store in a separate location is considered a separate establishment for
the purpose of this survey. However, if the records for main store and suburban branch are kept on a combined basis, they
may be considered as one establishment.
(a) Location: Identify each establishment by its street address and city.
(b) Type of Retail Activity: Enter for each establishment the major retail activity such as department store, drug store,
gas station, etc.
(c) Employment: Include all full-time, part-time, seasonal, and casual employees who received pay for any part of the
payroll period including June 13, 1966. Exclude employees, such as those in leased departments and demonstrators,
who received all or a substantial part of their pay from another employer.
Total—Enter total number of employees including officers and other principal executives, such as buyers, department
heads, and managers whose work is above the working supervisory level.
Nonsupervisory—Enter total number of employees below the supervisory level, such as salespersons, shipping and
receiving clerks, laborers, warehousemen,- caretakers, office clerks, driver-salesmen, installation and repairmen,
elevator operators, porters, janitors, watchmen, and other employees whose services are closely associated with those
listed above. Oo not include officers and other principal executives, such as buyers, department heads, and managers
. whose work is above the working supervisory level.
(d) Annual Gross Sales for the Establishment: Check the column which indicates the annual gross volume of sales (exclu­
sive of excise taxes at the retail level).
(street address
and city)


--------------m --------------- --------- -------------Employment
for payroll period
Type of
June 13, 1966
Total supervisory

Gross establishment
(check appropriate column)
Were last year’s sales—
$150,000 $250,000
$150,000 $250,000


Employment and earnings data reported should correspond to your payroll period (for example, weekly, biweekly, or
monthly) including June 13, 1966. Indicate the dates for the payroll period used. If die length of the payroll period
varies among employees, enter the dates affecting the greatest number.
From ______________________ _ 19___to_________________________ 19____

2 2 2 —6 1 7 — 6 6 ------ 1 0



This study is designed to provide information on hourly earnings and weekly hours of work for both male and
female nonsupervisory employees and working supervisors for a payroll period including June 13, 1966.
The number of employees in each establishment for which earnings and hours data are reported should corre­
spond with the number of nonsupervisory employees shown in item 3(c) on page 1. The information requested
should be reported separately for each establishment and the establishment identified. Earnings data for food
counter, cafeteria, or restaurant workers in Department, Drug, or Variety Stores should be entered only on the
blue supplement provided. Data for all other employees should be reported in Item $ of this form.
Report earnings and hours separately for each employee unless these data are identical for two or more em­
ployees. Do not report aggregate earnings and hours for several employees, fo r convenience of reporting for
employees paid on other than an hourly basis, columns 5 through 8 are provided. Data w ill not, however, be
published separately by various methods of pay. Instructions and examples for reporting the necessary data in
each column are listed below.

(Please read carefully to avoid correspondence)

Column (1)—-Indicate whether the employee is male (M) or female (F).

Complete col­
umns 1, 2, and Column (2)—Use a separate line for each employee and enter " 1,” unless two or more employees of the same
3 for all nonsu­
sex work the same number of hours during the selected week, and receive identical hourly or salary rates
pervisory em­
(see example A). Data are to be reported individually for each employee whose earnings are based entirely
ployees cov­
or in part on commissions or bonuses (see examples C, D, and E).
ered by this
Column (3)—Enter the number of hours worked during a single week including June 13, 1966. Include hours
report (see
examples A—E).
paid for sick leave, holidays, vacations, etc. These hours should relate to a 1-week period regardless

of the length of the payroll period.

Use column 4 to
report earnings Column (4)—Enter the base (straight-time) hourly rate. Premium pay for overtime work should not be reported.
o f employees
This column may also be used to report earnings of employees paid on other than an hourly basis if average
paid on an
straight-time hourly earnings are available. For employees paid a commission or bonus in addition to an
hourly basis
hourly rate, also complete columns 7 and 8 (see example D).
(see example A).
Use columns 5
and 6 to report
earnings of em­ Column (5)—Enter for each employee the straight-time earnings for the salary period (weekly, biweekly,
ployees paid on
monthly, or semimonthly) including June 13, 1966.
Include straight-time pay for overtime, but exclude
a weekly, bi­
overtime premium. Do not include "draws” against commission as salary.
weekly, monthly,
or semimonthly Column (6)—Enter the number of hours worked during the salary period (weekly,, biweekly, monthly, or semi­
basis (see ex­
monthly). Include hours paid for sick leave, holidays, vacations, etc. For employees paid a commission or
ample B).
bonus, also complete columns 7 and 8 (see example E).

Column (7)—Enter for each employee the total commission and/or bonus earnings, including "PM’s,” "Stims,”
or any special bonuses based on sales paid quarterly or oftener by the store. These earnings are to be

reported for the commission or bonus period including June 13, 1966.
If the commissions earned dur­
Use columns 7
ing that pay period are not representative of normal commission earnings, a longer period may be used. If
and 8 to report
earnings of non­
store employees receive both commission and bonus payments for an identical period of time, report the
supervisory em­
combined figure (see example D). If bonus payments cover a period longer than the commission period, add
ployees based
only the prorated amount of the bonus to the commission earnings that correspond to the commission period
entirely or in
(see example E).
part on com­
missions and
bonuses (see
Column (8)—Enter the number of hours worked during the commission or bonus period. (The hours should
example C).
refer to the total hours worked during the period (weekly, biweekly, monthly, or semimonthly) and not nec­
essarily only to those hours during which commissions or bonuses were earned.) For employees paid an
hourly rate or salary in addition to commissions or bonuses, it is also necessary to complete column 4, or
columns 5 and 6 (see examples D and E.)

NOTE: If all the information requested in Item 5 is contained in your payroll, you may, if you wish, send us a copy of your pay­
roll instead of completing Item 5. All other items on the form should be completed and the form should be sent with the payroll.
The copy of your payroll will be returned to you if you request it.
(See illustrations on next page)
A. Two women each worked 3 6 \ hours during the selected week, and each was paid a straight-time hourly rate of $1.05.
B. One man worked 40 hours during the selected week, and received a salary of $125, exclusive of premium pay for overtime,
for 88 hours worked during the salary period (V2 month).
C. One man worked 32% hours during the selected week and u>as paid on a straight commission basis, receiving $215.70
for 168 hours.
D. One woman worked 40 hours during the selected week and was paid an hourly rate of $1.25; she also received $35 in com­
missions and $7.50 in uPM*s" for 173.6 hours worked during the commission period (1 month).
E. One man worked 37*4 hours during the selected week, and was paid a weekly salary of $75; he also earned commissions of
$102 during a 1-month period (162 hours) and $150 in bonuses during a 3-month period. Only Vj of the bonus, or $50 is
reported so that the bonus period corresponds to the commission period.






Complete these columns for
each nonsupervisory employee.

Use this
column for non­
supervisory em­
ployees paid on
an hourly basis.

Hours worked
Number during a single Straight-time
week including hourly rate
June 13,
Illustrations of examples on page 2.
$1 05
B. M
3 Z .S
d. f
___ 3-7.S



E a t. C ity
s a le s siz e

Em p.

C U ts
em p.

Use these columns for nonsupervisory employees paid
other than on an hourly basis.

salary for
salary period
June 13, 1966

Hours worked
salary period


bonus pay

Hours worked



75.00____ |___ 37.5

____ L3LGQ___ ___ VfcT-O





Estimating Procedure
Although a greater proportion of large than of
small establishments is included in the sample, any
possible bias which might result from this differ­
ence is avoided by means of the estimating pro­
cedure. Each establishment in the sample is
assigned a weight which is the reciprocal of the
sampling ratio in the stratum from which it was
selected. That is, an establishment selected from
a stratum in which a sampling ratio of 1 out of 4
is used is assigned a weight of 4, so that it repre­
sents itself and 3 other establishments in the
stratum (see previous example). Data for each
establishment are multiplied by the weight as­
signed to the establishment. Thus, all establish­
ments, regardless of their size, are represented
appropriately in the final estimates.
An establishment in the subsample of nonre­
spondents is weighted to represent all nonrespond­
ents in the stratum. It is assigned a new weight—
the product of the original weight and the inverse
of the subsampling fraction. Thus, if a third were
subsampled of a group originally sampled at the
rate of 1 out of 2, the weight of 6 would be as­
signed. In the case of an establishment included
in the sample with certainty, another establish­
ment which is similar to the nonrespondent would
be weighted to represent it.
In industry surveys, estimated employment
totals derived from the weighting process are
further adjusted to the employment levels for the
payroll period studied, as reported in the Bureau
of Labor Statistics monthly establishment employ­
ment series.8 This adjustment is necessary to re­
duce the hazards of sampling, and because the
State U I listings, which constitute the universe or
a large part thereof, are prepared prior to the time
of the survey and thus do not account for establish­
ments opened or closed between the compilation of
the lists and the date of the survey.
Estimated average hourly earnings and weekly
hours of work are the arithmetic mean of weighted
individual employees5 earnings or hours. Gen­
erally, they are derived by totaling weighted indi­
vidual hourly earnings or weekly hours and di­
viding the sum by the weighted number of em­
ployees in the group. However, in industries such
as retail trade, in which hourly earnings frequently

vary by weekly hours of work and in which the
length of the workweek covers a broad range
among employees, the most representative group
average hourly earnings figure is considered to be
the quotient of total individual weekly earnings
divided by total individual weekly hours worked.*7
The following example illustrates the more com­
mon method of estimating group average hourly
earnings. Referring to the sampling scheme de­
scribed in the preceding example, assume that the
eight sample establishments in the under 25 size
group had 30 employees, each earning $1.50 an
hour, 20 earning $2 and 15 earning $3, and that the
12 establishments in the 25-49 size group and the 3
in the 50- and over-size group had employees with
earnings as shown. Making the calculations that
follow, estimated group average hourly earnings
are obtained.
WeightX employment™






Estimated group
average hourly

Weighted group
hourly earnings

Total weighted group hourly earnings _$2,015.00
^------- ~~
Total weighted group employment

I f the average is to be computed following the
second procedure (such as would be done in retail
trade studies), employment would be replaced by
weekly hours in the example and the remainder of
the calculations carried through. In this case, the
ratio now shown in the example would be replaced
by the ratio of total weekly earnings to total
weekly hours.

Analysis and Presentation
The Bureau generally issues a bulletin on the
results of each survey of employees’ earnings and
hours distributions. The report contains a
description and analysis of the survey results as
well as tabulations of data collected.
8 See chapter 2, ‘‘Employment, Hours, and Earnings” .
7 See Employee Earnings in R etail Trade, June 1962 (B L S
Bulletin 1380, 1963), p. 65.



In industry studies, tabulations are presented
for the entire industry and frequently for im­
portant segments of the industry. For example,
in a survey of manufacturing, in addition to data
for the entire industry division, data also might be
presented for most or all of the major groups
(2-digit SIC codes) for several groups (3-digit
SIC codes), and perhaps for some individual in­
dustries (4-digit SIC codes). In addition to na­
tionwide persentation, data for all or some of the
industry segments generally are shown separately
on a regional basis, and frequently on a
metropolitan-nonmetropolitan area basis.
Data obtained in area studies, which include
most industry divisions, are tabulated for all the
industries in the area combined, and then sepa­
rately for manufacturing and nonmanufacturing
industries. Where sufficient data are available,
tabulations are presented for selected industry
groups as well.
Earnings are tabulated to show the distribution
of employees by 5- and 10-cent average hourly
earnings intervals. Hours are tabulated to show
the distribution of employees by weekly hours of
work. The total number of employees, their aver­
age hourly earnings and average weekly hours, of
course, also are shown. Data also may be pre­
sented to show average weekly earnings and cross­
tabulations of average hourly earnings by weekly
hours of work.
The text accompanying the tabulations gives a
description of the area or industry studied and
provides a summary and analysis of the survey
results. Important relationships and differences
among areas and industries are highlighted and,
where data are available from earlier surveys,
period-to-period changes in earnings and hours are
discussed. In addition, a detailed description of

the scope and method of survey is included in the

Uses and Limitations

The design of these employee earnings and hours
distributions studies makes them particularly use­
ful in the analysis of Federal minimum wage and
maximum hours legislation—in analyzing the ef­
fects of legislation, in considering new legislation,
and in formulating wage and hours policy. The
information is used by the executive and legisla­
tive branches of government, organized labor,
business, academicians, etc. Special tabulations
are prepared for the U.S. Department of Labor’s
Wage and Hour and Public Contracts Divisions
for use in the Secretary of Labor’s annual report
to the Congress required under section 4(d) of the
Fair Labor Standards Act. In addition, the data
are used in collective bargaining, wage setting,
broad economic analysis of labor area and in­
dustry wage structures, comparisons of wage
levels in various parts of the country, and to show
trends in employment, earnings, and hours.
Employment estimates are subject to some error,
and in smaller groups this error may be relatively
large. Therefore, these estimates frequently are
provided to serve only as general guides to the
size of the labor force included in the survey and
as reference points from which to measure the
direction and general magnitude of employment
The average earnings reported are straight-time,
and any premium pay for overtime or late shift
work is not reflected. Similarly, differences in
prevailing supplementary compensation practices
(“fringe benefits” ) among establishments, indus­
tries, and areas are not considered.

Technical References

1. U .S . D e p artm en t of L a b o r, B u re au of L a b o r Statistics: E a rn in g s i n

W holesale T ra d e , J u n e 1958

(B u lle tin 1253, 1959).
2. ------- .

F a cto ry W o rk e rs ’ E a rn in g s , M a y 1958 (B u lle tin 1252, 1959).

3. ------- .

E m p loy e e E a rn in g s i n N o n m e tro p o lita n A reas o f the S ou th and N o r th C en tra l R e g io n s , J u n e

1962 (B u lle tin 1416, 1964).
4. ------- .

E m p loy e e E a rn in g s i n R e ta il T ra d e , J u n e 1962 (B u lle tin 1380, 1963).

E a c h of these bulletins contains a detailed description o f the m eth odology used in the su rv e y .

—A lvin Bauman
Office o f W a g e s a n d In d u stria l R elations

Chapter 16. Union Wage Scales
Background and Description of Survey
Annual studies of union scales of wages and
hours are conducted in four industries: Building
construction, local transit, local trucking, and
printing.1 Union scales, agreed upon through
collective bargaining between employers and trade
unions, are defined as (1) the basic (minimum)
wage rates (excluding holiday, vacation, or other
benefit payments regularly made or credited to
the worker each pay period) and (2) the maxi­
mum number of hours per week at straight-time
rates. Bates in excess of the negotiated minimum,
which may be paid for special qualifications or
other reasons, are excluded.
The use of union agreements or other union
records in studies of occupational wages is prac­
ticable in industries that are characterized by a
high degree of organization and in which (1)
defined craft groupings persist, as in building con­
struction or printing, or (2) key occupations can
be clearly delineated, as in local transit.
The Bureau’s annual union wage studies began
in 1907. Originally, information was obtained for
39 cities, but the number was gradually expanded
until in 1948, 82 cities were covered.2 That num­
ber was reduced to 77 in 1949 and to 52 in 1953.
The studies were again expanded, after the 1960
Census of Population, to the present coverage of
68 cities with 100,000 population or more. The
scope of the information for individual industries
has also been expanded. For example, 24 journey­
men crafts and nine helper and laborer classifica­
tions in the building trades are covered currently,
in place of the 13 journeymen and seven helper
and laborer classifications in the initial studies.
The study of union scales in the building trades
includes virtually all journeymen and helper and
laborer classifications. Indexes and other data are
shown for each important trade as well as for all
trades combined.3
The trucking study embraces motortruck drivers
and helpers engaged in local trucking. Over-theroad drivers and local city drivers paid on a mile­
age or commission basis are excluded. All data,
including indexes, are presented for the two classi­
fications indicated.

Union scales in the local-transit industry are*
limited to operating employees. Data are shown
separately for operators of surface cars and buses,
and elevated and subway lines, except that indexes
are shown only for the industry as a whole.
In the printing industry, 12 book and job trades
and 8 newspaper trades are studied and for the
newspaper trades, separate data are shown for day
and nightwork. Indexes and other data are pre­
sented separately, by type of printing, for each
trade and for all trades combined.

Data Sources and Collection Methods
The union wage studies are designed to include
all local unions in the covered industries in the
selected cities. Periodic checks are made with
central labor unions, district councils, and other
authoritative bodies to identify new local unions
which should be included in the studies.
Information is collected by mail from local
unions and when necessary from international
unions and regional union organizations. Per­
sonal visits are made to unions which do not re­
spond to the mail questionnaire. Before 1947, all
data relative to union wage studies were collected
directly from local union officials (generally the
secretaries or business agents) by Bureau repre­
sentatives and entered on forms designed specifi­
cally for this purpose.
To insure accurate reporting of data by mail,
quality control checks are conducted annually in
a number of cities. Checks are made in all cities
at least once every 5 years. In these checks, Bu­
reau representatives personally visit local union
officials to review and verify data previously re1 The coverage at various former periods also included barbers,
linemen, longshoremen, and workers engaged in breweries,
laundries, metal trades, millwork, restaurants, soft-drink pro­
duction, theaters, and bakeries.
2 In these studies, data relate to individual cities and con­
tiguous suburban areas, rather than to th e much broader
Standard Metropolitan Statistical Areas which are used in most
other Bureau surveys.
8 In addition to the annual studies in the building trades, a
quarterly survey of 7 m ajor construction trades is conducted in
100 cities. Estim ated average hourly scales for all trades com­
bined and for each surveyed trade are presented, together with
the estimated change during the quarter.


ported by mail questionnaire. Clarifying mis­
understandings, if any, of survey definitions and
requirements during such reviews enables the Bu­
reau to maintain its standards of reliable data.
Information requested relates to July 1 for all
industries. This date was adopted, after numer­
ous changes, because most new agreements in these
industries have been negotiated by that time each
year. In order to maintain year-to-year compara­
bility, scale and membership data for the previous
year are transcribed onto the forms before they
are sent out. Union officials are requested to check
the previous year’s data and revise any figures
which may have been incorrectly reported, and
to insert current data. Copies of union agree­
ments are also requested from union officials for
the purpose of (1) checking the data entered on
the schedules with the terms of the agreements,
and (2) building up the files of union agreements
maintained by the Bureau of Labor Statistics.4
The reporting form used for the building trades
survey is reproduced on pages 138 and 139.

Sampling and Estimating Procedures
The current series is designed to reflect union
wage scales in all cities of 100,000 population or
more, excluding Honolulu. All cities with 500,000
population or more are included, as are most cities
In the 250,000 to 500,000 group. The cities in the
100,000 to 250,000 group selected for study are dis­
tributed widely throughout the United States.
Data for some of the cities included in the study
are weighted to compensate for cities not surveyed.
To provide appropriate representation in the com­
bination of data, each geographic region was
considered separately when city weights were

An over-all average hourly rate is computed for
each of the industries included in the union wage
studies. In addition, averages are presented by
industry branch, trade, city, and region in build­
ing construction and printing; by city and region,
in local transit; and by region and city, in


Average union rates are calculated by weighting
each quotation for the current year by the reported
membership.6 These averages are levels designed
to provide comparisons among trades and cities
at a given time. They do not measure the trend
of union rates, the function served by the index

Chain indexes are calculated for each of the four
industries to portray the trend of union scales of
rates and weekly hours. In calculating these in­
dexes, the percentage change in aggregates is com­
puted from quotations for all identical classifica­
tions in the industry for 2 successive years. To
obtain the aggregates, the rates and hours for both
the previous and current years are weighted by the
membership in the particular classification for the
current year. The index for the current year is
computed by multiplying the index for the preced­
ing year by the ratio of the aggregate change. In
the 1964 study of building trades, the rate aggre­
gate for all quotations increased 3.7 percent over
the previous year. The July 1, 1964, index of
union hourly wage rates for all building trades
(126.2) is the result of multiplying the July 1,
1963, index (121.7) by the ratio of the aggregates
(103.7). This method of index calculation mini­
mizes the influence of year-to-year changes in
Indexes of union hourly wage rates and weekly
hours are computed for each classification as well
as for all classifications combined in the building
construction and printing industries. In the local
trucking and local transit industries, indexes are
provided only for all classifications combined.
Irregular hours of work for operating employees
in many of the covered cities prevent the computa­
tion of an index for union weekly hours in the
local transit industry.
The base period for the indexes of union wage
scales and weekly hours is the 1957-59 average.
The series for the building trades and printing
4 See chapter 20. “ Collective Bargaining Agreements.”
6 The cities in the sample were not selected on a probability
basis. Continuity of city data has taken precedence over
sampling procedure during periods of contraction and expansion
in the program.
6 Reported membership, a s used in this study, is defined a s
members working or immediately available for work.


BLS 1150.1

Budget Bureau N o. ^4~R 738.12.
■ A pproval ex p ires March 31, 1968.

(R e v . ’6$)

Union S c a le s of Wages and Hours in the Building Trades
(Annual Survey)
P A R T A . — G E N E R A L IN F O R M A T IO N ;
P le a s e attach a copy of your agreement in effect on July 1, 1965» and answer the follow ing questions?


When did your agreement go into effect?
When may it be reopened?
When does it expire?


Entirely by employer?

__________________________________________________ _______

What is the maximum number of hours that can be worked each day be­
fore overtime rate is effective?
_____________________________________ _____


D oes your agreement s p ec ifically provide for a paid vacation, financed
by employer paym ents—

In part of employer?

Amount of employer
contribution? .
U or %)








To a vacation fund L ]



D oes your agreement provide for employer payments to other funds,
such as hoUday, educational, promotional, unemployment benefits?
L 1Y es
|__ j No If Y E S , list below each type of contribution separately
and amount of employer payments (cents or percent).


What is the total membership of your union?

(B )

T o worker each pay period, as part of negotiated
( _ ] No

(C )

To worker each pay period, in addition to negotiated
s c a le

(D )

Other— (E x p la in )



_______ -____ _____________ _ _________________________

How many are journeymen? ___

_ _ _ _ _ _ _ _ _ _ _ _ _ _


How many are apprentices? _______________ ,_______________________________


If answ er is Y E S to any of above, indicate amount of employer pay­
______________________ per hour,
U or % )


Does your agreement provide for a health and insurance plan (life in­
surance, hospitalization, m edical, surgical, and other similar types
of health and w elfare program s) financed—

Between July 1, 1964,and July 1,1965, bow many apprentices completed
their appreriticeship? ( i f none enter z e r o . ) _ _ _____________________ _ _


Between July 1 ,1 9 6 4 ,and July 1, 1965, how many journeymen became
unavailable for work because of death, permanent d isab ility , or re­
tirement? (If none enter z e r o . ) _ __________ ____

Entirely by employer? l ]




In part by employer? Q I!
Amount of employer
( t or % )

How many are helpers and laborers?



per 4 shift





______________________ ,_______ ___




Does your agreement provide for a pension plan

P le a s e enter the data requested below.

Do not use columns marked “ C o d e .”

Employer payments to insurance, pension, vacation, or other funds should be excluded from the hourly scales in columns 3 and 4, and shown as cents or
percent of scale in the spaces provided above in Part A.
Payments s p e c ific a lly designated as being for or in lieu of a benefit should be excluded from hourly s c a le s reported in columns 3 and 4 below even though
they are made or credited to the worker each pay period,
Membership information w ill be kept in confidence and used only to compute average wage rates.

Schedule N o . __________

T it le


July 1, 1964
(3 )

July 1, 1965
(4 )

July 1, 1964

C ode



(6 )

July 1, 1965
( ? ) ________

C ode


July 1, 1964
______ (22________ |

Ju ly 1, 1965
(1 0 )


in each
July 1, 1965




Number < union
members vworking or
immediate ly a v a ila ble for tvork at
each rate

W eekly hours before
overtim e rate is
e ffe c t iv e


Wage s c a le for each
trade or occupation
in e ffe c t on

Trade or occu pation

C ode

P a g e _____________of


___________ _________________* Date

jm r)

P le a s e sign your name here
A ddress

(S t r e e t and numfcct)

(C ity and State)

, 1965

(Zip code)



industry date back to 1907, for local transit to
1929, and for local trucking to 1936. Although
data for the latter two industries were collected
for years prior to the dates of the index series,
indexes were not constructed because of inadequa­
cies in the available data.

Analysis and Presentation
The above mentioned averages and indexes to­
gether with other summary data are contained
in the bulletins published annually for the indi­
vidual studies. Included among the information
shown for individual trade classifications is the
proportion of union members having hourly scales
at different rate levels, as well as the proportion
of union members having, since the previous study,
scale increases of specified amounts in terms of
cents per hour and percent. The increase reg­
istered by the trade is shown also.
In addition, the union scales of wages and hours
in effect on the date of the survey, as reported by
union officials, for both the previous and current
years are published for each classification by city.
These furnish a direct comparison of union scales
between the 2 years for each of the industries
studied. The scales of wages are indicated as
hourly rates and the scales of hours as the weekly
hours of work before overtime rates are applicable.
For the building trades and local motortruck­
ing the current studies also present data on em­
ployer payments for insurance (health and wel­
fare) and pension payments; in addition employer
payments for vacation funds are shown for the
building trades. These payments are expressed
in terms of cents per hour or as percent of rate.
7 It should be noted that membership (used for weighting pur­
poses) relates only to active members in the city and contiguous
suburban areas. It does not reflect the total jurisdiction of local
unions, which may extend beyond these limits, nor do the data
reflect metropolitan area scales.

Uses and Limitations
The Bureau’s union-wage series provide a means
of determining intercity wage differences for com­
parable work, and the relationships between rates
applicable to workers in occupations requiring
varying degrees of skill. The data are used in
wage negotiations by both management and labor.
The scales of building-trades workers are espe­
cially important in estimating construction costs,
because labor expenditures constitute an important
element in the total cost of building construction.
The index series derived from these studies pro­
vide barometers of year-to-year changes in scales
of wages and hours in the industries covered.
Average union rates provide comparisons of
wage rates among industries, trades, and cities at a
given time. Unlike the indexes, they are not an
accurate measurement of year-to-year changes be­
cause of fluctuations in membership and other
factors. Membership figures for the various
trades or classifications do not remain constant
and changes may have a marked effect on average
scales. For example, if organizational drives in
cities having relatively lower scales of wages re­
sult in sharp increases in membership, the move­
ment of the scale levels for the affected trades as
a whole is naturally retarded. Conversely, in­
creases in membership in cities having high wage
scales accelerate the upward movement of
The union rates, it should be noted, are not
necessarily the actual rates paid to all workers, nor
are the union hours the hours actually worked.
Workers with above average experience and skill
may be employed at rates above the union scale,
especially during prosperous times when a tight
job market creates competitive bidding for the
better workmen. During periods of depressed
business activity, actual hours worked are often
less than hours specified in the union agreement.
— J o h n F. L a c i s k e y
Office o f W a g e s an d In d u s tria l R elatio n s

Chapter 17. Current Wage Developments
Since January 1948, the Bureau of Labor Sta­
tistics has issued a monthly report listing general
wage changes and changes in supplementary bene­
fits agreed to in selected collective bargaining situ­
ations, identifying the situations by company and
union name. The scope of the listing has varied
somewhat from time to time, but since 1953, it
has been limited to agreements affecting approxi­
mately 1,000 or more production and related
workers in manufacturing and selected nonmanu­
facturing industries.
The Current Wage Developments report was
initiated because of the rapid increase in wage
rates and prices in the early postwar period and
the interest in determining the extent to which set­
tlement patterns spread from industry to industry.
It also was prompted by the fact that reductions in
available resources made it necessary to discon­
tinue an index of wage rates that had been
initiated during World War II. Interest in the
listing was stimulated by the Korean emergency
when the Wage Stabilization Board needed data on
the extent to which wages and benefits were being
In 1949, and again in 1951 and 1952, statistical
summaries of wage changes were prepared to sup­
plement the listing, but regular preparation of a
statistical summary began in 1954. These quar­
terly statistical summaries show the distribution of
settlements and (since 1955) of workers by the
size of the general wage changes agreed to and the
frequency with which various types of supple­
mentary benefits were introduced or changed.
Beginning in 1959, another statistical summary
was instituted. It is limited to manufacturing,
1 The listing, as contrasted with these summaries, provides a
much more detailed account of negotiated wage and benefit
changes than can be presented in a tabular summary. When
available, information on changes for large groups of nonunion
workers, including professional, white-collar, and production em­
ployees, also is presented in the listing.
2 Only changes in benefits that represent changes in costs are
8 Prior to 1966, the construction service trades and finance
industries are also excluded.
4 General wage changes are defined as changes affecting at
least one-tenth of the workers a t any one time or all workers in
an occupation. Changes resulting from promotions, merit in­
creases, etc., are excluded.

but includes information on general wage changes
and changes in supplementary benefits for non­
union and small union situations, as well as for
large collective bargaining situations.1

Description of Series
The summary of major collective bargaining
situations hereafter is referred to as the “major”
series, and the summary that is based on changes in
wages and benefits in manufacturing firms of all
kinds is described as the “manufacturing” series.
The major series describes general wage changes
and changes in benefits2 in all collective bargain­
ing settlements involving 1,000 or more production
and related workers in manufacturing and 1,000
or more nonsupervisory workers in the nonmanu­
facturing sector, excluding Government.3
Supervisory or professional employees are ex­
cluded. Large units of technicians are included
even though they are part of a bargaining unit
that is predominantly professional.
Contracts covering multiplant firms are in­
cluded if the agreement as a whole covers 1,000 or
more workers even though each individual plant
employs fewer. Also included are contracts with
trade associations or with groups of firms that bar­
gain jointly with a union or unions, even though
the firms are not associated formally and each has
fewer than the minimum number of workers cov­
ered by the series. Situations in which two or
more unions, together representing more than
1,000 workers but individually accounting for
fewer, negotiate essentially identical contracts
with one firm or a group of firms, are tabulated as
one bargaining unit.
The summary for manufacturing as a whole
represents all establishments with four or more
employees that adjust wages by means of general
wage changes,4 regardless of whether the workers
are represented by a union.
Wage change data are presented in cents per
hour and, since 1959, as a percentage of average
straight-time hourly earnings, adjusted to exclude
premium pay for overtime work.



Two types of information are presented on wage
changes: (1) information on changes that go into
effect within 12 months of the date on which they
are decided or negotiated, thus reflecting the eco­
nomic climate at the time the changes are decided
upon; and (2) all changes effective during the
period being summarized whether or not they were
(a) decided upon during that period, (b) were de­
ferred—that is, decided upon earlier—or (c) re­
sulted from operation of cost-of-living clauses. In
distributions of workers by size of wage change,
all workers in an establishment or collective bar­
gaining situation are distributed according to the
average wage increase in the establishment or situ­
ation. The number of workers affected by changes
in supplementary benefits includes all production
and related workers in the situations where the
benefit is changed, whether or not all are affected
immediately. For example, if a fourth week of
vacation is added for workers with 20 years5service
in an establishment employing 1,000 workers, a
vacation change would be recorded for 1,000 work­
ers, even though only a relatively small propor­
tion would benefit from the change immediately.

Data Sources and Collection Methods
The statistical summary of the major series is
compiled from the summaries of collective bar­
gaining settlements presented in the monthly
Current Wage Developments listing which, in
turn, is derived primarily from secondary sources,
including general circulation newspapers and peri­
odicals, as well as union, management, and trade
publications. Other important sources of infor­
mation are the file of union contracts maintained
by the B L S and the U.S. Department of Labor’s
files of pension and health and welfare agree­
ments, maintained by the Office of Labor-Manage­
ment and Welfare-Pension Reports.5 At the end
of the year, the B L S contacts, almost entirely by
mail, either management or labor representatives
in any situation for which these other sources have
not yielded information on wage and benefit
changes during the year.
Information for nonunion and small unionized
firms is gathered semiannually by a questionnaire
mailed to participating establishments.6 The in­
formation on general wage changes and the types

of benefits that have been introduced, liberalized,
or made less liberal is supplemented by the contract
filed (unionized establishments) and from news­
paper clippings, purchased from a commercial
clipping service. At the end of the year, B L S field
representatives contact, primarily by telephone, a
sample of firms that have failed to respond to the
mail questionnaire or that have provided incom­
plete or unclear information.

Sampling and Estimating Procedures
As indicated earlier, all bargaining situations
with 1,000 or more workers in manufacturing and
nonmanufacturing industries are included in the
major series. It is believed that the current list
of nearly 2,000 such situations, built up since
Current Wage Developments was started in 1948,
is very nearly complete. After a bargaining situ­
ation is added to the universe, it is withdrawn only
if it ceases to be within the scope of the survey
(e.g., because of a change in business to one out­
side the scope of the survey, a change to nonunion
from union, or because of an apparently permanent
drop in employment to substantially below 1,000).
The sample for manufacturing is derived from
State unemployment insurance listings (UI)
which show reporting units with four or more
employees by location, number of employees, and
industry classification.7 The sample is a highly
stratified probability design, with sampling ratios
varying from 1 out of 150 establishments with 4 to
19 employees to all of those with 1,000 or more em­
ployees.8 The ratios are uniform for all indus­
tries. Since data are available from secondary
sources for all unionized situations with at least
1,000 production and related workers, data for all
establishments meeting this criteria also are in­
cluded in the summary for manufacturing. The
sample selected from the U I listings is compared
5 Information from contracts supplied on a confidential basis
is used only in the statistical summaries, not for the monthly
6 From 1959-60, when the series was initiated, through 1964,
the questionnaire obtained information each quarter. Beginning
in 1965, data were collected semiannually.
7 See appendix B, “ Industrial Classification.” For a more
detailed description of unemployment insurance data, see p. 15,
chapter 2.
8 In the case of a few companies with large numbers of estab­
lishments each with 1,000 or more workers, a sample of plants
is chosen.


with this list of establishments for which informa­
tion already is available; since data for these
sample members are obtained from secondary
sources, these establishments are not sent question­
naires. Approximately 4,000 establishments are
left for the questionnaire survey.
Although the sampling design yields a sample
in which large firms are relatively overrepresented,
this bias is overcome by the estimating procedure.
Each establishment in the sample is assigned a
weight which is the reciprocal of the sampling
ratio in the stratum from which it was selected.
An establishment selected from a stratum from
which 1 out of 4 establishments is chosen is as­
signed a weight of 4, so that it represents itself and
three other establishments. Information for each
establishment is multiplied by the weight assigned
to the establishment. Thus, all establishments,
regardless of size, are represented appropriately
in the final estimates.
An establishment in the subsample of nonre­
spondents followed up by a visit is weighted to
represent all nonrespondents in the stratum. It
is assigned a new weight—the product of the
original weight and the inverse of the subsampling
fraction. Thus, 1 out of 3 nonrespondent estab­
lishments subsampled from a group originally
sampled at the rate of 1 out of 2 would be assigned
a weight of 6. I f an establishment included in the
sample with certainty fails to respond, another
similar establishment would be weighted to rep­
resent it.
To the estimates derived from the weighting of
the sample questionnaire are added the data from
secondary sources—the numbers of major bargain­
ing situations.
The totals thus obtained are further adjusted to
reduce the hazards of sampling and to take account
of opening or closing of establishments between
compilation of the State unemployment insurance
listing from which the sample is chosen and the
date of the survey. Adjustments are made to em­
ployment levels for production workers in the 2digit Standard Industrial Classification manufac­
turing industry groups, as reported in the monthly
employment series of the Bureau of Labor Statis­
tics for the period covered by the summary of wage
changes.9 For example, if the estimate of produc-*
* See chapter 2.


tion worker employment in an industry group
derived from the sample is 100,000 but the Bu­
reau’s estimate of employment in that industry
group was 110,000 workers, each employment count
would be multiplied by
or LI. The result­
ing industry group estimates would be added to
provide the estimates for all manufacturing.
The major series for manufacturing and non­
manufacturing combined is not adjusted in this
fashion, since it is presumed to be all inclusive.
A new sample of nonunion and small unionized
plants in manuf acturing usually is selected every
2 or 3 years. Establishments with fewer than
four workers are omitted because in many States
they are not covered by these unemployment
compensation programs. After the initial contact,
establishments of any size that indicate that they
have a policy of adjusting wages on an individual
basis, rather than by means of general wage
changes, are omitted from further survey.

Presentation and Analysis
Press releases showing preliminary information
on general wage changes for the major collective
bargaining situations are issued approximately 6
weeks from the end of the quarter. Final sum­
maries of the major data include information on
changes in supplementary benefits, and are pub­
lished for the first 6 and 9 months of the year and
for the full year. The 6-month summary also in­
cludes separate tabulations of data relating to the
first 3 months. All of these summaries are pub­
lished as supplements to Current Wage Develop­
ments, with the full-year article usually also
published in the Monthly Labor Review . The 6and 9-month summaries are limited to analysis
of wage and benefit changes resulting from con­
tract settlements concluded during the year. The
yearend summaries present similar information,
but in greater detail, and also include data on total
effective wage changes—those negotiated in the
period plus deferred and cost-of-living changes
resulting from settlements in prior years but going
into effect in the current year. Information is
presented for manufacturing industries, for non­
manufacturing industries, and for both combined.



Summary releases of the manufacturing data
are published for the first half of each year and
for the full year. The article covering the full
year also appears in the Monthly Labor Review .
The 6-month summary stresses the results of settle­
ments or management decisions made during the
period, while the yearend analysis is broader and
more detailed. The examination of changes in
supplementary benefits is not as detailed as the
examination of major benefits, because of the need
to simplify the information requested by mail.
The respondent is asked only if each of the types
of benefits—such as holidays, etc.—has been
“established, improved, or decreased or discon­
tinued,” rather than to specify the precise types
of change.
Because it is based on data both for large and
small unionized and nonunionized establishments,
the manufacturing analysis can make many use­
ful comparisons of its components. For example,
comparisons can be made between the median
wage increases in unionized versus nonunionized
establishments, or the type and frequency of
changes in supplementary benefits.

Uses and Limitations
The data are used extensively by labor, manage­
ment, and the Federal Mediation and Conciliation
Service in collective bargaining; by private insti­
tutions and universities in studies of industries or
groups of industries; and by local and Federal
Government agencies interested in the current
economic picture to determine trends in wage and
benefit changes as well as for wage, income, and
gross national product forecasts.
Because estimates of the cost of supplementary
benefits usually are not available, the wage change
data must be presented in relation to average
hourly earnings, rather than to average hourly
labor costs. In 1965, a pilot study was undertaken
to find ways to determine such costs in major
Since the sample is relatively small, data are not
presented for individual industries.
— G eorge R u b e n
Office o f W a g e s and In d u s tria l R elatio n s

Chapter 18. Employer Expenditures for Supplementary
Compensation Practices
The measurement of employer expenditures for
selected supplementary compensation practices
and the composition of payroll hours was under­
taken by the Bureau of Labor Statistics (BLS)
to fill a large gap in the statistics of employee
compensation and hours paid for.
Prior to World War II, compensation for
American labor consisted mainly of wages and
salaries for time worked or units produced. Under
the New Deal, however, additional payments were
required under various social insurance programs,
and, later, during the war years, employers were
encouraged by the policies of the War Labor Board
to grant wage supplements instead of wage in­
creases, e.g., vacations, hospitalization. Shortly
after the war, the NLRB ruled that pension plans
were within the purview of collectively bargained
agreements.1 Expenditures for these and other
wage supplements began to comprise a substantial
portion of the total compensation of labor.
As early as 1875 the American Express Com­
pany instituted a private pension plan.2 In 1929,
a private study3 indicated that there were almost
400 such plans, and by 1962-63 there were about
33,410 pension plans and 161,750 private pension
or welfare plans in America.4
Direct wage supplements, such as paid vacations
and holidays, also have a relatively short history
for most workers. Paid vacations were fairly well
established for salaried workers by the middle of
the 19th century. Industrial workers, however,
first started to receive paid vacations around the
turn of the century, and not until after World
War I did the principle of paid vacations begin
to assume importance in the development of labor
policy; paid holidays generally were not found
in industry until World War II, although it had
been customary for salaried workers to receive
pay for time not worked on designated holidays.
By 1962, approximately 6 percent of the produc­
tion worker hours paid for in manufacturing in­
dustries were leave hours, almost all of which were
vacation and holiday hours.

The growing importance of the supplements in
the Nation’s wage and salary structure is illus­
trated by comparing the increase in expenditures
for production and related workers in manufac­
turing with the increase in average hourly earn­
ings. In 1962, manufacturing industry employers
had expenditures per plant hour for the selected
supplements studied5 that were on the average,
about 18-percent greater than in 1959, the previous
period studied. During the same period the aver­
age hourly earnings of production and related
workers rose by about 9 percent—a rate of increase
approximately one-half that of the increased
expenditures for the supplements.
The Bureau has for many years recognized the
necessity of studying outlays for supplementary
compensation. Early attempts, however, were
limited to exploratory work on methodology and
the availability of data.6 By 1959, many of the
technical and conceptual problems had been suf­
ficiently resolved to permit the initiation of a reg­
ular program.
The first survey in the program, 1959 expendi­
tures in manufacturing, was followed by a 1960
mining study, a 1961 finance, insurance, and real
estate survey, and in 1962, by another manufac­
turing industry study. The 1963 study of ex­
penditures for salaried (white-collar) workers,
which covered most nonagricultural industries in
the private sector, represented the first shift in pro­
gram emphasis from an industry to an economy­
wide orientation. Since then, the program has
1 Inland Steel vs. National Labor Relations Board, 170 Fed­
eral Reports, Second Series 247 (1948), 251 (1949).
2 This was the first recorded private pension plan in America.
8 Latimer, Murray Webb. Industrial Pension Systems in the
United States and Canada, Industrial Relations Counselors, Inc.,
New York, 1932.
4 Total active plans for which descriptions had been filed with
the U.S. Department of Labor by Ju ly 1, 1963, under the pro­
visions of the Welfare and Pension Plans Disclosure Act, as
amended (P.L. 85-836 as amended by P.L. 87-420).
8 The 1959 and 1962, studies related to employer expenditures
for production and related workers and included expenditures
for paid leav e; premium pay for overtime, weekend, holiday, and
shift w ork; Christmas, yearend, and other irregular bonuses;
legally required insurance program s; and private welfare plans.
6 Problems in Measurement of Expenditures on Selected Items
of Supplementary Employee Remuneration, Manufacturing
Establishments, 1953 (B L S Bulletin 1186, 1956) ; Wage Struc­
ture: B asic Iron and Steel, Jan uary 1951 (B L S Series 2, No. 81,
1951) ; Wage Structure: Electric and Gas Utilities, September
1957 (B L S Report No. 135,1958).




been redesigned to cover all employees in the pri­
vate nonfarm sector of the economy.

Description of Survey
Employer expenditure studies are limited to
selected compensation practices, and are analyzed
in terms of percent of payroll and cents per hour.
The study program is designed to biennially pro­
vide data for all employees, in each of two broad
occupational groupings, in the private nonfarm
economy and to provide data for specific manu­
facturing and nonmanufacturing industries in the
intervening years.
The composition of the hours-paid-for is de­
scribed by data on the number of hours spent at
the plant and the number of paid leave hours for
vacations, holidays, sickness, etc., as proportions
of total hours. Survey coverage extends to the 50
States and the District of Columbia and separate
data are provided for broad economic regions.
The data relate to cash disbursements of em­
ployers and the hours-paid-for during a calendar
year. Data for a lesser period of time do not com­
pletely reflect the outlays made by employers and
the hours-paid-for. Paid leave time, for example,
usually is spread unequally throughout the cal­
endar year; similarly, expenditures for many of
the legally required insurance programs stop after
a specified maximum amount is earned by each
worker during the year, resulting in wide varia­
tions between expenditures in the early part of the
year and in the latter part of the year.
The entire field of supplementary compensation
is changing in magnitude and breadth. Practices
differ by industry group and new supplementary
practices are being added. Each survey is de­
signed to capture the individuality of compensa­
tion practices in particular industries without im­
pairing comparability with others. This objective
is accomplished by retaining the same basic refer­
ence framework.

The expenditure practices studied are considered
to comprise the major elements of supplemental
compensation in American industry. They do not
include all practices which occasion outlays by em­

ployers that result from the employment of work­
ers. The supplementary practices studied fall into
two broad groups: payments made directly to the
workers; and employer expenditures made on be­
half of the workers, some of which expenditures
are obligatory.
The direct supplements include that part of
gross payroll which consists of payments for paid
leave time (vacations; holidays; sick leave; and
military, jury, witness, voting, and personal leave),
payments of premiums for overtime, weekend,
holiday, and shift work, nonproduction bonuses
and terminal payments.
The indirect supplements are those outlays not
included in the payroll for which the employer
irrevocably makes a payment under the provisions
of a legally required insurance program or private
welfare plan. Generally, the payments are to
funds, trustees, insurance companies, or Govern­
ment agencies which either make a payment to the
workers at a later date or provide full- or partialeconomic security against a future contingency
(e.g., unemployment, retirement, medical ex­
penses). The indirect supplements studied to
date include expenditures for social security, un­
employment compensation, workmen’s compensa­
tion, and other insurance for the benefit of the
workers required by law; life, accident, and health
insurance; pension and retirement plans; vacation
and holiday funds; severance and supplemental
unemployment benefit plans; and savings and
thrift plans.

Payroll Hours
The payroll hours studied are all hours for
which the workers receive pay. These hours con­
sist of plant or working hours, and vacation, holi­
day, sick and other hours of paid leave. Although
an hour normally is defined as 60 minutes of
elapsed time, a payroll hour does not necessarily
consist of 60 minutes. For example, hours worked
on a day that would otherwise have been a paid
holiday are paid for twice—once as a paid holiday,
and once as working hours. Therefore, an 8-hour
holiday worked for which 16 hours of payment
was made is counted as 16 hours—half of which
are holiday hours and half are plant-hours. Con­
versely, some hours of leave are paid for at less
than the regular rate and only the equivalent hours


are counted. Each overtime hour worked at
premium rate is counted as 1 plant-hour.
Establishment Policies
Although the data on company policies are used
primarily in the review and analysis of the expend­
iture and hours data collected in the survey, these
policies also have significance in their own right,
and often signal changes in supplementary prac­
tices before actual expenditures are incurred. In
addition, certain characteristics of American in­
dustry are measured on a national basis and the
relationships among these characteristics studied.
Examples are the actual distribution of workers
by amount of vacation and the degrees of

Data Sources and Collection Methods
The data are obtained from establishment
records. Generally, no single record is sufficient
and several record sources must be summarized
to arrive at annual totals. The data are entered
by the employer on preprinted forms in accord­
ance with detailed instructions.
Not all companies keep records in the detail re­
quested and approximations in these cases may be
accepted. In general, two types of approximations
are used. First, if the establishment records are
kept for a broader grouping of employees than are
being studied, the prorated share for the workers
included in the survey is computed on the basis
of employment, man-hours, or payroll, whichever
is most appropriate. Second, using collateral data,
estimates are made where records are not kept but
the practice is observed. For example, the ex­
penditures for holiday pay may be approximated
by multiplying the number of hours paid for holi­
day leave by average straight-time hourly earn­
ings. Errors occurring from the use of these
approximations would have to be in the same
direction in substantially all the cases (overstate­
ment or understatement of the actual values) to
have a material effect on the accuracy of the
Data are collected primarily by mail, although
personal visits are made to many of the large
employers and to a sample of the establishments



that have not responded to a second mailing of
the questionnaire. A questionnaire form used in
the expenditure study is reproduced on pages

Sampling Procedure
The surveys are conducted on the basis of a
highly stratified probability sample of establish­
ments selected by industry, location, and employ­
ment size. The samples generally are designed
to yield reliable data for an industry division at
the national level, in four broad economic regions,
and for major industry groups.
The lists of establishments from which the
samples are selected are those maintained by the
State agencies administering the employment in­
surance laws. These lists show the employment,
industry classification, and location of all estab­
lishments covered by those laws in each State.
Since some States do not cover establishments
with fewer than four employees under the
unemployment insurance (UI) law, the samples
exclude establishments in that size group. (See
method of estimation for treatment of the employ­
ment in such establishments.) Some establish­
ments in particular industries are exempted from
the U I laws even though they employ more than
four workers. The data used in sampling these
establishments are obtained from lists compiled
by regulatory Government agencies, trade asso­
ciations, and other sources.
Within each industry, the sample is selected to
yield the most accurate estimates possible with the
resources available—the principle of optimum al­
location. This is done by including in the sample
a greater proportion of large establishments than
of small. In general, an establishment’s chance of
selection is roughly proportionate to its employ­
ment size.
A subsample of establishments failing to reply
to the mail inquiries is selected to represent all
nonrespondents, following the same general plan
as is used in the original sample. Since the re­
sponse rate for those solicited by mail is about 50
percent, and a subsample of about a third is se­
lected, the final sample contains about two-thirds
of those originally solicited by mail. Establish­
ments in this subsample are visited personally,
instead of being solicited again by mail.



BLS 2868 A


Budget Bureau No. 44-R 1300
Approval expires 2-28-67


Your reply will be
held in confidence.

Employer Expenditures For Selected Compensation Practices, 1965

Location of unit for which data are requested
if different from address.


M ajor A ctivity

What w a s the prin cip al product produced or se rv ic e provided by this establishm ent during 1965?


Em ploym ent

F o r each em ployee group lis te d b elow , p le a s e enter the total number o f fu ll- and part-time em ployees on the e s t a b ­
lishm ent’ s payroll during the pay period including September 13, 1965. P L E A S E S E E P A G E 2 F O R D E F IN IT IO N S
O F P R O D U C T IO N A N D R E L A T E D W O R K ER S A N D N O N P R O D U C T IO N W O R K E R S. T o ta l employment is the sum
o f these two groups.

T o ta l employment


Production and related workers


Nonproduction w orkers:


P ro fe s s io n a ls , ex e cu tiv es, adm inistrators,
m anagers, and other sup erv iso rs (ex clu d in g
working su p e rv iso rs)--------------------------------------------


O u tside s a le sp e rso n s (i. e ., those w hose s a le s
a c tiv ities are primarily performed outside
the estab lish m e n t)------------------------------------------------


O ffic e c le r ic a l and other nonproduction workers*





P le a s e

a n s w e r a l l o f the q u e s t io n s w h ic h




f o llo w by p r o v id in g c a le n d a r y e a r in form atio n for P R O ­

W O R K E R S an d fo r N O N P R O D U C T I O N

W O R K E R S that w e r e e m p lo y e d

d u rin g 1965 by the e s t a b lis h m e n t id e n t if ie d on the f ir s t p a g e o f th is q u e s t io n n a ire .

P R O D U C T I O N .A N D R E L A T E D W O R K E R S a r e d e fin e d to in c lu d e a l l non s u p e r v is o r y w o rk e rs a n d
w o rk in g forem en e n g a g e d in b u ild in g , a lte r in g , d e m o lis h in g , e x c a v a t in g , m in in g, pu m pin g, f a b r i c a ­
tin g , p r o c e s s in g , a s s e m b lin g , in s p e c t in g , r e c e i v i n g , s to rin g , h a n d lin g , p a c k in g , w a r e h o u s in g , s h i p ­
p in g , tru c k in g , h a u lin g , m a in te n a n c e , r e p a ir,

ja n it o r ia l, w atch m en s e r v i c e s , p ro d u ct d e v e lo p m e n t,

a u x ilia r y p ro d u c tio n fo r the p la n t ’ s ow n u s e ( e . g . , p o w e r p la n t ), r e c o r d k e e p in g an d oth er s e r v i c e s
c l o s e ly a s s o c ia t e d w ith the a b o v e p ro d u c tio n o p e r a t io n s .

N O N P R O D U C T I O N W O R K E R S a re d e fin e d to in c lu d e a l l e m p lo y e e s in e x e c u t iv e , a d m in is t r a t iv e ,
an d m a n ag em en t p o s i t i o n s a b o v e the w o rk in g s u p e r v is o r le v e l .
P r o f e s s i o n a l , t e c h n ic a l, o f f i c e
c l e r i c a l , an d s a l e s e m p lo y e e s a re a l s o in c lu d e d in th is c a t e g o r y .

P r o p r ie t o r s , m em b e rs o f u n in c o rp o ra te d firm s, and u n p a id fa m ily w o rk e rs are not c o n s id e r e d to b e
e m p lo y e e s an d t h e re fo r e a re e x c lu d e d from th is s u r v e y .

I f y o u r r e c o r d s do n o t p r o v id e
fig u r e s re p o rte d

s e p a ra t e d a ta for P R O D U C T I O N


W O R K E R S p l e a s e p r o v id e c a r e fu lly c o n s id e r e d a p p ro x im a tio n s .

a re a p p r o x im a t e s , p le a s e

W O R K E R S an d
I f any o f the

l i s t on p a g e 8 the item s in v o lv e d and in d ic a t e h o w the

a p p ro x im a tio n s w e r e m ad e.

I f the o n ly fig u r e a v a i l a b l e c o m b in e s d a ta fo r s e v e r a l li n e s on t h is q u e s t io n n a ire , r e p o rt the
co m b in e d fig u r e an d b r a c k e t the li n e s in c lu d e d or o t h e r w is e in d ic a t e w h a t i s in c lu d e d in the
fig u r e r e p o rte d .

P L E A S E D O N O T L E A V E A N Y L I N E S B L A N K . I F N O E X P E N D I T U R E S O R M A N -H O U R S W E R E
I N V O L V E D D U R I N G 1965 F O R A G I V E N I T E M , E N T E R " 0 ” I N T H E A P P R O P R I A T E S P A C E .




P a ym en ts M ade D ire c tly to W o rk e rs a s P a rt of the 1965 P a y ro ll

In th is sectio n p le a s e report only those paym ents that com prised part o f the g r o s s p a y ro ll o f the establish m en t*
D o not report any paym ents to fun ds, tru stee s, in su ran ce co m pan ies, or governm ent a g e n c ie s h e re ; rep o rt th o se
paym ents in sectio n 4.

A. Gross payroll:




S e rv ic e Form s W-2 a s b e in g s u b je c t to F e d e r a l w ithholdin g
ta x e s ___________________________________________________________________

Production and
related workers



E nter the total o f w a g e s , s a la r ie s , and other paym ents m ade
during 1965 b efore any d e d u ctio n s. T h e amount reported
sh o u ld e q u al the sum o f w a g e s reported on Internal R e v e n u e

Employer payments to



B. Paid leave:
E nter the total paym ent m ade d irectly to your em p loyees for
each type o f le a v e . I f an em ployee re c e iv e d both pay in s te a d
o f time o ff and pay for w ork, in c lu d e only the paym ents made
in lie u o f time o ff in the fig u re reported here.


V a c a t i o n s ----------------------------------------------------------------



H o li d a y s --------------------------------------------------------------------



S ick l e a v e ---------------------------------------------------------------



C iv ic (m ilita ry , ju ry , w it n e s s , voting, e t c .) and
p e rs o n a l l e a v e --------------------------------------- -----------------


C. Pay for overtime hours and for extra work on
weekends and paid holidays:
R eport the total paym ent for d a ily and/br w ee k ly overtim e
hours and for extra w ork perform ed on w e e k e n d s, the 6th
and 7th day o f the w eek , on p a id h o lid a y s or during p aid
v ac atio n p e rio d s . E X C L U D E P A Y M E N T S , R E P O R T E D IN
O R V A C A T IO N S .



Pay m en ts at straight-tim e ra te s
R eport the total straight-tim e portion o f the paym ents
(to ta l payment m inus premium p a y ) here. (T h e p re ­
mium portion should be reported in Item 2 b e lo w .) —


Premium paym ents — -------------------------- -----------------------------






Employer payments to

D. Shift differentials:
R ep o rt here only the paym ents a b o v e the reg u lar d a y -s h ift r a t e s .
I f 8 hou rs o f pay is given for 7-1/2 hou rs o f work report the


Production and
related workers


total o f the 1/2-hour p a y m e n t s ------------ ----------------------------------------------

E. Nonproduction bonuses:
R eport the total amount p aid for nonproduction b o n u se s in ­
clu d in g lump-sum paym ents m ade under p ro fit-sh a rin g p la n s
and other irre gu la r or s e a s o n a l b o n u se s (e .g ., C h ristm as,
atten dan ce, merit, and s u g g e stio n b o n u se s or a w a rd s).
E x c lu d e paym ents for q u a lity or quantity o f work ------------------

F. Terminal payments:
Report the total o f a ll paym ents made by the establish m en t
to em p loy ees b e c a u s e o f temporary or permanent s e v e ra n c e
o f em ploym ent. (E x c lu d e here paym ents to fu n ds, and
paym ents to p e n s io n e rs m ade under the p ro v is io n s o f p a y a s -y o u -g o pen sio n p la n s . Such paym ents sh o uld be reported
in Item 4 -A , P riv a te W elfare P la n s.)


Em p lo yer Ex p en d itu re s in Addition to P ayro ll

A. Private welfare plans:
R epo rt net paym ents (a fte r deduction o f refun ds, re b a te s, and d iv id e n d s ) made during 1965 by the e s t a b lis h ­
ment to funds (in c lu d in g labor-m anagem ent fu n d s), tru stee s, in su ra n c e com panies, and paym ents m ade undier
the p ro v is io n s of s e lf-in s u r e d p la n s to em p loy ees or their b e n e fic ia r ie s . In c lu d e paym ents fo r current em ploy­
e e s , em p loy ees on la y o ff, retired em p lo y ee s and their depen den ts. D O N O T R E P O R T H E R E A N Y P A Y M E N T S
A L R E A D Y R E P O R T E D O N L I N E S 1-10. E x c lu d e em ployee co n tribu tio n s and a l l ad m in istrative c o s ts incurred
by the estab lish m en t.

A ls o e x c lu d e paym ents made by fun ds, tru ste e s, and in su ran ce c a rrie rs to your em ploy­

e e s or their b e n e fic ia r ie s .

Employer expenditures
during 1965 for
Production and
related workers

L i f e , a c cid e n t, and health in su ran ce



L i f e , a c c id e n ta l death and dism em berm ent,
s ic k n e s s and ac c id e n t, w a g e and sa la ry
continuance insurance> and death b e n e fits ------ ---------H o s p ita liz a tio n , s u rg ic a l, m ed ic al, m ajor
m ed ic al, d en tal, o p tic a l, and drug p la n s .
(E x c lu d e ex p en d itu res for in -p la n t m ed ical
care and v is it in g n u rses or p h y s i c i a n s . ) ------ -----------






A. Private welfare plans— Continued

P en sio n and retirem ent p la n s

Employer expenditures
during 1965 for
Production and
related workers

Inclu de direct paym ents to pensioners under a p a y -a s y ou -go pension plan; a ls o inclu de payments under
p rofit-sharin g plan s deferred until retirement. For
funded p la n s, inclu de paym ents for p a st and current
s e rv ice lia b ilit ie s __________________________________________





V aca tio n and holiday f u n d s --------------------------------------------


Severance or d is m is s a l pay and/or supplem ental
unemployment ben efit f u n d s --------------------------------------------



S av in g s and thrift p la n s




Other private w e lfa re p la n s ----------------------------------------------



B. Legally required insurance programs:
Report the total net lia b ility incurred during the four quarters o f 1965 under the p ro v isio n s o f State and F e d ­
eral la w s for each o f the program areas liste d below . Inclu de payments to government a g e n c ie s , insurance
com panies, and payments m ade directly to em ployees in addition to p a y ro ll. E xclu d e paym ents m ade by or w ith­
h e ld from em ployees.


Unemployment in su ran ce program s

N et lia b ility incurred
____ _____ during 1965 for
Production and
related workers

Report payments to State and F e d era l unemployment
com pensation funds for 1965 including charges for
extended benefit and other s p e c ia l unem ploy­
ment programs







Paym ents to F e d e ra l G o v e rn m e n t---------------------------



Paym ents to State governm ents--------------------------------


S o c ia l security and sim ilar le g a lly required retirement
incom e and protection p r o g r a m s --------------------------------------


O ccu pation al injury and illn e s s
Report payments made under the pro vision s o f work­
men’ s com pensation and rela ted State and
F e d e ra l l a w s ------------— ----------------- ----------------------------------


Other le g a lly required

insurance program s, inclu ding

State temporary d is a b ility insurance
program s (s p e c ify )





' 5. Hours Paid For and Scheduled W orkw eek During 1965

In parts A and B of this section, please report the total number of hours during 1965 for which the establishment
made a direct payment, reported in Item 3-A, to the employees. Do not report hours paid for by funds, trustees, or
insurance companies. In part C, please report the number of hours usually worked each week by the majority of
employees in each group. Before answering the questions please read the definitions, at the bottom of the page,
of nonexempt and exempt nonproduction workers.

Include all hours forked, all paid leave hours reported
in B, below, and hours equivalent to payments made
by the establishment directly to the workers
for other hours not worked but paid for-------------------

Production and
related workers

Nonproduction workers




A. Total hours paid for:


Number or hours paid for during 1965

B. Paid leave hours:
Report hours equivalent to the payments
made and not the time taken. For example:
3 hours paid for at 2/3ds the regular rate
should be reported a s 2 hours. Include
leave hours for which payment was made
in lieu of time off.

V acations---------------------------------------



H olidays-----------------------------------------



Sick l e a v e --------------------------------------



Civic (military, jury, witness, voting, etc.)
and personal leave-----------------------------


C. Regularly scheduled workweek:
How many hours per week were normally worked
by the majority of the employees in each of the
occupational g ro u p s--------------------------------

Number o f hour s norm ally work**d each week
by the :majority of emp) oyees

NONEXEMPT EMPLOYEES are those subject to the overtime provisions of the Fair Labor Standards Act. Estab­
lishments not subject to the act should consider all nonsupervisory, nonprofessional office and technical employees
a s being nonexempt.
EXEMPT EMPLOYEES are those not subject to the overtime provisions of the Fair Labor Standards Act. Estab­
lishments not subject to the act should consider all supervisors, above the working supervisor level, administrators,
executives, professionals, and all outside salespersons a s being exempt.



6. Establishm ent P ra ctice s and P o licie s

A. Paid vacations:
Report the number of employees who received vacation pay during 1965 directly from the establishment accord­
ing to the amount of pay received and not the time taken. If vacation pay was not a direct multiple of weekly
or hourly rates report the numbef"of weeks equivalent to the payment. For example: If the amount paid was
computed as a percentage of annual earnings report payments of about 2 percent as 1 week’ s vacation pay;
about 4 percent as 2 weeks’ vacation pay; etc.

Employee group


1 week’s

Number of employees receiving—
1 and
2 and
3 and
4 and
2 weeks’ 3 weeks’ 4 weeks’ 5 weeks’

5 weeks’
pay or

Production and related Workers





B. Paid holidays observed during 1965:
Enter number of days per employee. If more than one practice existed for an employee group, report that which
applied to the greatest number in the group.
N um ber o f h olidays per em p lo y e e

Full -days

H alf-days

Production and related workers.
Nonproduction workers_______

c. Welfare and pension plans (other than legally required):
1. Were any employees covered by a private plan, as defined in section 4,
that was paid for all or in part by the establishment (answer "Y E S ” if
there were such plans even though there were no employer expenditures
during 1965) that provided—

and related
w ork ers


(Check Yes or No for each group)

a. Life, accidental death, or sickness and accident insurance----------

Yes P | No P j Y e s |

b. Hospitalization and other health insurance------------------------------

Y es |

c. Pension and retirement plans----------------------------------------------

Y es □

1No 1 | Yes □
No □

| No 1 1
No Q

Yes □

No □

a. Life, accidental death, or sickness and accident insurance----------

Yes P | No P ] Yes P

No □

b. Hospitalization and other health insurance------------------------------

Yes □

No □

Y es □

No □

c. Pension and retirement p la n s----------------------------------------------

Y es □

No □

Yes □

No □

Y es □

No □

Y es P

No □

2. If there was a plan, did the workers contribute to the cost of the plan
(answer "N O ” if their contributions only purchased additional
benefits or dependents coverage)

D. Collective bargaining:
Did collectively bargained agreements cover a majority of the
non supervisory employees —----- —— -----— ---- -——--------- ---- -------




L ist the items here for which data were approximated and indicate the methods used in approximating. P lease also
include any other pertinent explanation of the data you have reported.

If any part of this report includes data for more than the unit described on the front of this form please indicate
the location, major product or service, and employment for the payroll period including September 13, 1965, for
each such unit included in the report; the line(s) on which the combined data were entered and the employee
groups involved.
Units for which data are reported

Major product
or service

Production and
related workeis

Name of authorizing official (please print or type);

Do you want a copy of the Bureau's report for this survey?

222-617*—-66----- 11

Combined data


Item or line




Y e $ n No |




Estimating Procedures
Data for each sample establishment are weighted
in accordance with the probability of selection of
that establishment. For instance, where 1 estab­
lishment out of 5 was selected in a size-industry
class, it is considered to be representing itself as
well as four other establishments, and accordingly
is given a weight of 5. Thus, if an establishment
had 1,000 hours of vacation leave, 50,000 hours
of paid hours of all classes, and had a weight of
5, it would contribute 5,000 vacation hours and
250,000 total hours to the final estimate.
The subsample of nonrespondents is weighted
to represent all other nonrespondents in the same
industry-size-location stratum. If, for instance,
the original weight of the sample units in a
stratum is 2, and 1 of 3 such nonrespondents is sub­
sampled, the unit would receive a weight of 6 in
the estimating procedure.
In the event that usable data cannot be ob­
tained from any unit visited in person, whether
among the followup of nonrespondents or among
large units often selected in the sample with cer­
tainty, its weight is assigned to units in the sample
with the most similar industry-size-location
All estimated totals derived from such weighting procedures are further adjusted by the level,
of total employment or other appropriate measure
for the survey year, based on data from the Bu­
reau’s monthly establishment employment sta­
tistics program, in each of four broad economic
regions. For instance, if the level of the aggre­
gates, as derived from the weighting procedures,
is 40,000,000 in an industry-region class and the
corresponding level as shown by the employment
statistics program is 44,000,000, the totals of the
survey items would be multiplied by 1.1. The
adjusted data represent all establishments, includ­
ing those with fewer than four employees, in the
industries studied.7
Some improvisation is necessary in the construc­
tion of such annual benchmark totals. The
monthly employment series provides data for only
one pay period each month, and the estimate of
annual totals is made by multiplying by the aver­
age number of weeks in a year (52.14).
Information from other sources, wherein a
detailed breakdown by State or region is shown, is

used as a basis for prorating the current employ­
ment (or hours) estimates into regional aggre­
gates. Such sources include the Census of Man­
ufactures and Comity Business Patterns (based
on Social Security establishment data) .8

Analysis and Presentation
The expenditure data are presented as a percent
of gross payroll, as a percent of straight-time
payroll, in cents per hour paid for, and in cents per
plant hour, for all establishments, and for estab­
lishments that actually had an expenditure during
the survey year. The payroll hours data also are
presented in ratio form for all establishments, and
show plant and leave hours as a percent of all
The expenditure ratios presented in the reports
are obtained as follows:
1. Expenditures as a percent of gross payroll=
Expenditure for the practice y^pp
Gross payroll
2. Expenditure as a percent of straight-time payroll=
_______ Expenditure for the practice
____ X100
Gross payroU minus premium payments for time
3. Expenditures in cents per hour paid fo r=
Expenditure for the practice
Total hours paid for including
hours of paid leave
4. Expenditures in cents per plant or working hour—
Expenditures for the practice
Total hours paid for minus leave
hours paid for

The hours ratios as indicated in the formulas
are derived by basically the same techniques as
Plant hours as a percent of total hours=
Total hours paid for minus leave hours paid for ^
T otal hours paid for


Leave hours as a percent of total hours=
Leave hours paid for ^ qq
T otal hours paid for

The expenditure ratios for all establishments
represent the expenditure for the supplement di­
vided by the payroll (or hours) for all establish­
ments—both those with and those without expend­
itures—whereas the ratios for establishments
7 See chapter 2, ‘‘Employment, Hours and Earnings” .
8 U.S. Department of Commerce, Bureau of the Census, County
Business Patterns (various years).

with expenditures for the practice relate the same
expenditure to the payroll (or hours) of only those
establishments that reported actual expenditures
for the supplement.
The distribution of workers by establishment
expenditures ratios is published as well as the aver­
age expenditure ratio for all workers. Expendi­
tures are analyzed in terms of establishment char­
acteristics such as size, wage level, and unioniza­

Uses and Limitations
Data from the surveys are used by employers
in comparing their expenditure and hours prac­
tices with the averages for their industry and with
those of other establishments having similar or
dissimilar characteristics (industry, size, location,
union status, and average earnings levels of work­
ers) . Labor and management use the data in col­
lective bargaining; and Government uses the
statistics in the formulation of public policy, in
producing estimates of industry output per manhour, and in making international comparisons.
They also are used in deriving estimates of the
amount and type of labor compensation and the
nature of the hours for which compensation is
received by workers.
The expenditure data can be compared with and
combined with average earnings data to obtain
estimates of the compensation for labor. (Some

of the supplements studied are integral parts of
payroll and caution, therefore should be exercised
in combining the ratios and earnings.) The esti­
mates thus derived do not, however, represent the
total expenditures for labor. Total labor cost is
a more encompassing concept and includes such
factors as the cost of recruiting and training
labor, the administrative expenses incurred in ad­
ministering benefit programs, and many other
expenditures resulting from the use of labor as a
factor of production.
The practices studied, as indicated earlier, com­
prise the major elements of supplementary com­
pensation in American industry. Some other
practices, which are excluded from the surveys,
may occasion important expenditures in particular
The expenditures and hours data are subject to
both sampling and reporting errors, the precise
magnitude and direction of which are not known.
Nevertheless, the errors resulting from sampling
generally are considered to fall within acceptable
confidence ranges; and reporting errors, to have a
material effect on the accuracy of the results,
would have to be in the same direction in substan­
tially all of the cases. The omission of establish­
ments with fewer than four employees from the
samples may result in some bias, but it is very
small since less than 1 percent of the workers in
most industries are employed by such establish­

Technical References

1. U.S. Department of Labor, Bureau of Labor Statistics.

E m ployer E xp en ditures fo r Selected S u p p le ­
m entary Com pensation P ractices fo r Production an d Related W orkers; Com position of P a y ro ll H o u rs:
M an u fa c tu rin g In d u strie s, 1962 (Bulletin 1428, 1965).


Each expenditure bulletin contains descriptive information on the detailed procedures and
techniques used in the study.
----- . E m ployer E xp en ditures fo r Selected Su pp lem en tary C om pensation P ractices fo r P roduction an d
R elated W orkers, M eatpack in g an d P rocessin g In d u strie s, 1962 (Bulletin 1413, 1964).
----- . E m ploy er E xp en ditures fo r Selected Su pplem en tary Rem uneration P ractices in F in a n c e , In su r ance , an d R e al E state In d u strie s , 1961 (Bulletin 1419, 1964).
------. E m ployer E xp en ditures fo r Selected Su pp lem en tary Rem uneration P ractices fo r P roduction
W orkers in M an u factu rin g In d u strie s, 1959 (Bulletin 1308, 1962).
----- . P roduction W orkers in M in in g In d u strie s, 1960 (Bulletin 1332, 1963).
----- . Problem s in M easurem ent o f E xpen ditures on Selected Item s of Su pp lem en tary Em ployee Re­
m uneration, M an u factu rin g Establish m en ts , 1958 (Bulletin 1186, 1956).
A study of the availability of records, willingness and ability of industry to provide data, the
quality of expenditure data, and other matters of methodology and definition.


— A rno ld S t r a sse r

Office of Wages and Industrial Relations

Chapter 19. Work Stoppages
Work stoppage statistics are compiled by the
Bureau of Labor Statistics to provide a quantita­
tive measure of the extent to which disputes be­
tween labor and management result in strikes or
lockouts and of the immediate economic disruption
resulting from such stoppages.1 When considered
along with general economic measures, these sta­
tistics also serve at times as a broad indicator of
the state of industrial unrest.
The first attempt by any Federal agency to com­
pile statistics on strikes was made in 1880,2* when
the Bureau of the Census sent questionnaires to
employers and workers involved in all disputes
during the year which were noted in the public
press. Information was received on 762 situations.
Some data were obtained on the causes of strikes
and their results, but not on the number of work­
ers involved or resultant man-days of idleness.
The next collection of strike statistics was under­
taken in 1887, when the Bureau of Labor, then in
the Department of the Interior, examined files of
newspapers, trade journals, and commercial peri­
odicals for references to strikes for all years from
1881 to 1886. Staff members visited the areas
where strikes were reported and obtained detailed
information about each strike from every available
person or source. Studies utilizing basically the
same procedures subsequently were made in 1894,
1901, and 1906. As a consequence of these efforts,
data were published for the 1881-1905 period on
the number of strikes and workers involved, with
breakdowns by industry and State; the number
of establishments involved; and the percentage of
strikes involving labor organizations.
No Federal agency collected national informa­
tion on stoppages occurring during the 1906-13
period. In 1914, the Bureau of Labor Statistics
attempted to compile a record of all strikes and
lockouts during the year, relying exclusively on
printed sources. In the following year, the
Bureau inaugurated a method for the collection of
strike and lockout material which has been fol­
lowed, with modifications, since that time.
Briefly stated, the procedure was to send question­
naires to the parties involved in work stoppages,

following receipt from the press and other sources
of notices relating to these situations.
Improvements in the program in 1927, in partic­
ular the procurement of data on the number of
workers involved in all stoppages and the compu­
tation of man-days of idleness, inaugurated the
modem series of monthly and annual strike data.8

Description of the Series
The present series on work stoppages covers all
strikes and lockouts known to the Bureau of Labor
Statistics and its cooperating agencies, and which
continue for 1 full day or shift or longer and
involve six workers or more. These limitations
on size and duration, somewhat arbitrary but of
long standing, are necessary for reasons of effi­
ciency, and, in part, because of the difficulty in­
volved in defining, identifying, and securing infor­
mation on strikes which last for but a few hours or
The Bureau defines a strike as a temporary
stoppage of work by a group of employees to ex­
press a grievance or enforce a demand. A lockout
is defined as a temporary withholding of work by
an employer (or a group of employers) to enforce
terms of employment upon a group of employees.
Since 1922, the Bureau has made no attempt to
distinguish between strikes and lockouts in its
statistics; both types are included in the term
“work stoppages” and, for the sake of convenience
in writing, in the term “strikes.”
While an employer-employee dispute is implicit
in these definitions, there are some inclusions in the
Bureau’s series which relate only indirectly to this
concept. For example, jurisdictional strikes and
1 Throughout this chapter, the terms “ work stoppage” and
“ strike” are used interchangeably; both terms, unless otherwise
noted, also include lockouts. The definitions, terms, and classi­
fications used by the Bureau in compiling work stoppage data
were adopted for statistical and research purposes and have
no legal significance.
8 On the State level, the Bureau of Labor Statistics of Mas­
sachusetts, issued a report in 1880 on strikes in that State since
1825. In 1881, the Bureau of Industrial Statistics of Pennsyl­
vania issued a report on strikes in that State since 1835.
* For additional information on the early history of the work
stoppage statistics program, see B L S Bulletin 651, Strikes in the
United States, 1880 to 1986 (1938).


rival union disputes between two or more unions
often have the employer on the sidelines. In a
sympathy strike, the issue of the stoppage does
not usually involve the immediate employer.
Moreover, protest strikes against the actions of
governmental agencies are not the result of a dis­
pute between an employer and his employees.
All stoppages, whether or not authorized by the
union, legal or illegal, are counted. On the other
hand, the Bureau’s series excludes strikes of Amer­
ican seamen or other workers in foreign ports and
strikes of foreign crews in American ports. Also
excluded are so-called slowdowns, where em­
ployees continue at work but at deliberately re­
duced production speed, and those instances in
which workers report an hour or two late each day
as a protest gesture or quit work several hours
before closing time to attend rallies or mass
The number of work stoppages occurring during
a given period provides a measure of the frequency
of disputes, while the severity and impact of such
actions are measured by the number of workers
involved, duration, and the resultant man-days of
idleness. The basic statistical unit in the Bureau’s
series is the individual strike or lockout. I f
groups of employees (regardless of their number
and type and location of employment) join in a
work stoppage for a common objective, their
action is classed as a single strike.
The figure used for the number of workers in­
volved in a strike or lockout is the maximum num­
ber actually made idle in the establishment or
establishments directly involved. No distinction
is made in arriving at this figure between the active
participants in the strike, the number of union
members or workers covered by an agreement, and
those sent home by the employer when a stoppage
by one group or department prevents plant
Man-days of idleness, like the number of work­
ers involved, are based on the idleness at the estab­
lishment or establishments directly involved.
Workers involved multiplied by workdays lost
equal total man-days idle. In instances where the
number of workers idle varies during the period
of the stoppage, appropriate adjustments are made
in this calculation. Allowance is made in these
computations for holidays and days not normally


Data Sources and Collection Methods
The task of collecting strike data has two basic
elements: (1) to learn of work stoppages when
and wherever they occur, and (2) to obtain the
necessary facts regarding each situation as quickly
as possible.
Information about the existence of stoppages is
currently obtained from various sources, includ­
ing: (1) Clippings from daily and weekly news­
papers throughout the country provided by com­
mercial clipping services; (2) notices received
directly from the Federal Mediation and Concilia­
tion Service; (3) a periodic compilation by the
local offices of the State employment security agen­
cies, provided through the Bureau of Employ­
ment Security of the U.S. Department of Labor;
(4) a number of other State agencies, such as
State mediation boards and labor departments;
(5) various employers and employer associations;
(6) international unions and their publications;
(7) firms under contract to the Atomic Energy
Commission; and (8) other Federal agencies and
Aside from the clippings from newspapers and
other publications, most of these sources have been
developed over a period of years. As a general
rule, expansion in the Bureau’s “lead” sources
brings an increase in the number of stoppages re­
ported, but has little effect on the total number of
workers and man-days of idleness, because the
added stoppages tend to be small.
Following the receipt of notices regarding the
existence of work stoppages, the Bureau mails
questionnaires to the parties involved so as to se­
cure direct information on each stoppage. Should
a reply not be received within 3 weeks, a second
questionnaire is mailed, and, in the case of con­
tinued nonresponse, a mailogram or telegram may
be sent, or an effort made to secure the necessary
data by telephone. In some instances of non­
response, field representatives of the Bureau se­
cure the necessary data; in others, cooperating
State agencies may be asked to contact the parties.
The types of information sought by the Bureau
through its questionnaire have changed over the
years, partly in response to changing needs. It
must be stressed that the primary function of these
reports is to compile statistics, not to keep records
on the strike activity of individual firms and



unions. The questionnaire currently used is shown
on page 161.4
Although strikes, by their very nature, are usu­
ally matters of public knowledge and of reporting
by newspapers and other publications, the Bureau
holds confidential the individual reports submitted
by employers and unions, as well as supplementary
data collected through State or Federal agencies.
The rules of confidentiality observed here are sim­
ilar to those followed in other Bureau surveys.

Estimating Procedures
Since the Bureau is able to obtain information
on virtually all work stoppages within the scope
of its definition, estimating is necessary only in
the preparation of its monthly reports on the level
of strike activity in the United States as a whole.
The availability of reasonably accurate data on
the larger stoppages at the time these estimates
are prepared—approximately4 weeks after the end
of the month of reference—assures approximate
conformity to the final statistics which are based
almost exclusively on the parties’ replies.
Monthly estimates are prepared on the number
of stoppages, the number of workers involved, and
man-days of idleness. As there is a lag between
the occurrence and reporting of a number of rela­
tively small strikes, the number of stoppages begin­
ning during a given month is estimated by increas­
ing the number of strikes on which leads have
been received by a percentage which is fixed for
each calendar month. An estimate of the total
number of stoppages in effect during the month
is obtained by supplementing the latter estimate
by a percentage of the stoppages which were in
effect during the prior month.
In estimating the number of workers involved
and total idleness, efforts are made to obtain as
much preliminary information as possible on the
size and duration of individual large stoppages—
those of at least 500 workers or 5,000 man-days
of idleness. To the known figures for these large
stoppages is added the product of the estimated
number of smaller strikes and the average num­
ber of workers (or man-days) that previous ex­
perience indicates for such stoppages.
In its preliminary reports, as well as in its final
reports, the Bureau relates the man-days of idle­

ness to the total estimated working time of all
workers. “Estimated working time” is computed
for this purpose by multiplying the average num­
ber of employed workers by the number of days
worked by most employees. This excludes Satur­
days when customarily not worked, Sundays, and
established holidays. Total employed workers, as
presently used, coincides with the Bureau’s figures
of nonagricultural employment, excluding govern­
ment workers. A somewhat different figure for
total employment was used prior to 1951; how­
ever, tests have shown that the percentage of total
man-days of idleness computed on the old and new
basis usually differed by less than one-tenth of
a point.

Analysis and Interpretation
The data presented in the parties’ reports are
analyzed and classified according to a number of
significant factors, briefly described here:
(1) Each strike is assigned an industrial class fication in accordance with the Standard Industrial
Glassification Manual prepared by the Bureau of
the Budget.56 In those cases in which a stoppage
affects workers in more than one industry, one of
two procedures may be followed. I f the stoppage
is small, the strike is classified in the industry in
which it was initiated; in large interindustry stop­
pages, a stoppage is recorded for each industry
affected, and the approximate numbers of workers
and idleness are allocated to each.
(2) The duration of each stoppage is taken as
the number of calendar days from its beginning to
end. For stoppages which begin at a definite time
and are terminated by a formal agreement, no
problem arises in the determination of duration.
However, there are stoppages which, for a variety
of reasons, are never formally settled. These
range from situations in which the workers grad­
ually return to their jobs without a settlement to
those in which the employer decides to go out of
business. In cases of the former variety, the de­
tails of each individual situation are studied before
4 A modified form of this questionnaire Is used in the case of
most jurisdictional disputes and those in coal mining.
In the case of prolonged strikes, a less detailed questionnaire
is sent to the parties periodically to determine the statu s of the
6 See appendix B.


BLS 817
(Rev. 1963)

Budget Bureau No. 44-R210.1
Approval expires March 31* 1969



y o u r report 'Will be held
in confidence
Please answer All questions


This request for information relates to:





I. Employer

Name and mailing address

& Establish­

Number of establishments directly involved or in which, workers observed picket lines.
If more than one establishment, use reverse side; if one, enter below;
a. Location____________________ ______________________________________
b. Industry__________________________ ________________________ _______________________________.
(Indicate major type of activity and also principal products or services, e.g., Mining - bituminous coal; Con­
struction- highways and streets; Manufacturing-wooden upholstered furniture; Wholesale trade- plumbing
supplies; Transportation - motor freight.)

J3. Union


involved in

Local No.


Settlement was ratified on
(if applicable)

Employees returned to
work on

4. Dates of

.Stoppage began, on

& Number

TOTAL Workers idled at least one full shift or day-------------------------------------------------------- -------- workers
—Include all worker# directly involved in the stoppage and workers made idle by lack of work in the
same establishments or by observance of picket lines. If exact figures are not available, please provide estimate;)



Settlement was reached on

) '

(Im p o r t a n t

Did the number idle change significantly during.the stoppage?
Yes □
No □
(If “ yes," 'please enter changes in number idle and dates of changes on reverse side of sheet.
8. Normal workweek prior to stoppage:...................................... ...................... ................................ .................
..... —
7* Contract status (check one)
Stoppage occurred
□ In negotiation of first contract or in obtaining union recognition*
□ In renegotiating contract terms (expiration or reopening)
□ During term of contract (change in contract terms not involved)
□ Other (specify)_____________________________________________________________________ ,
8. MAJOR issues in dispute (list in order of importance)


9. Did the agreement to return to work include a procedure for handling any unsettled major issues involved in the stoppage (e.g., by
submitting issues to arbitration)? Yes □
No □
If “yes”, indicate the issues and the procedure agreed upon.
10. Did a Federal,-State, or local government agency, or a private mediator, mediate in this dispute or assist in arranging the return to
work? (Check more than one if applicable). Federal □
State □
Local □
Private □
None □
Please identify assisting government agency, if any
.. ...............
.............. ...... .


(Signature and title of person making report)

(Company or organization)
Use reverse side for any
clarifying remarks.



a stoppage is terminated for statistical purposes;
in the latter instances, the stoppage is terminated
with the employer’s announcement of his decision
to discontinue operations. On occasion, if actual
settlement is reached later, the statistical record of
the stoppage is adjusted accordingly.
(3) Number of establishments involved. The
standard definition of establishment is used. (See
appendix B.) An establishment is a single work­
place, for example a factory, mine, or store. In
a widespread strike of intercity bus drivers, truckdrivers, or railroad workers, the establishment is
regarded as the terminal out of which the em­
ployees work; in a strike of seamen, the ship is the
establishment; and in a strike of dockworkers,
the individual dock or loading place is regarded
as the place of work.
(4) Geographical classification of stoppages
followed State and city boundary lines, through
1951. Beginning in 1952, the compilation of data
by Standard Metropolitan Statistical Areas super­
seded city boundary lines.6 In interarea stop­
pages, a stoppage is recorded in each area affected,
and workers and man-days of idleness are allocated
(5) The issues in dispute in most strikes are
many and varied, and do not always lend them­
selves readily to immediate and exact classification.
Stoppages are classified by major issue into the fol­
lowing broad groupings: (a) wages, hours, and
supplementary benefits; (b) union organization
and security; (c) job security; (d) plant adminis­
tration; and (e) inter- or intra-union matters.
Within each of these groups, there are further sub­
divisions into more specific categories.7
(6) Stoppages are classified by the contractual
relationship existing between the parties involved.
The following four situations apply: (a) negotia­
tion of the initial agreement; (b) renegotiation of
an agreement; (c) agreement in effect (new con­
tract terms not involved); and (d) no contractual
(7) The union involved is another major classi­
fication. For this purpose, the union is the orga­
nization whose contract was involved or which•
• See appendix C.
TWhen a major change in the classification of issues was
troduced in 1961, the Bureau included in its annual report
that year a method of linking the new classifications with
a Since 1949, these bulletins have been titled Analysis
Work Stoppages {year).


has taken active leadership in the stoppage. Dis­
putes involving more than one union are classified
as jurisdictional or rival union disputes or as in­
volving cooperating unions. I f unorganized
workers strike a separate classification is used.
For publication purposes, union information is
presented by major affiliation of the union, i.e.,
AFL-CIO, or nonaffiliation such as uIndepend­
ent,” “single firm,” or “no union.”
(8) The assistance of mediators , either govern­
mental or private, in the resolution of industrial
disputes is recorded.
(9) The manner in which stoppages are settled
involves classification into the following cate­
gories : (a) those ending with a formal settlement;
(b) those terminating without a formal settlement,
and in which work is resumed with either the old
or new workers; and (c) those concluded by the
employer’s decision to go out of business.
(10) A classification also is made of the manner
in which unsettled issues are to be resolved in those
situations where strikes are terminated with the
understanding that such matters will be handled
following the resumption of normal operations.

Publications in the area of work stoppages in­
clude monthly preliminary estimates, annual re­
ports, and special reports which are issued
Monthly preliminary estimates are issued ap­
proximately 40 days after the end of the month of
reference in the form of press releases. Such
reports presently are prepared for the first 10
months of each year, with a preliminary estimate
of total strike activity during the year being issued
a few days after the close of the calendar year.
Selected final tabulations of strike activity are
presented in a release which usually is issued in
May. This is followed by a summary article,
published generally in the June issue of the
Monthly Labor Review , on strike activity during
the preceding year. An annual bulletin,8 con­
taining detailed information on the characteristics
of work stoppages during the prior year, is pub­
lished each fall.
Special reports containing historical work stop­
page data by industry and area, or an analysis of

a particular aspect of strikes, are issued irregu­
larly. This latter category also may include
Bureau reports of a nonstatistical nature, includ­
ing chronologies of “national emergency” disputes
arising under the terms of the Taft-Hartley Act.

Uses and Limitations
The use of strike statistics as an indicator of
industrial unrest has been the traditional reason
for their compilation in the United States and in
other industrialized countries.9 Whether they
serve this elementary purpose today is open to
question.10 In any event, some qualifications must
be taken into consideration in this use. The wil­
lingness of workers to strike as a protest against
existing conditions may be encouraged or deterred
by outside influences, such as the employment situ­
ation, the state of the business cycle, and possible
political or public reaction. Within the plant, the
strength of the union or of employer opposition
may influence both the willingness of the workers
to start a strike or to extend its duration.
Whether as a measure of industrial unrest or
the state of labor-management relations, strike
statistics are necessary for Federal, State, and
municipal government agencies, particularly those
concerned with labor affairs. Unions, employers,
and employer associations use strike data to assess
their own experiences, and business and civic
organizations are concerned with their commun­
ity promotional possibilities. Schools, particu­
larly those teaching courses in industrial relations,
and industrial relations counselors also find strike
statistics useful. Finally, the press is interested,
since strikes and information about them are
Although it is virtually certain that the Bureau
is able to locate, and obtain information on, -the
larger work stoppages, some small strikes undoubt­
edly escape notice each year. While these omis­
sions do affect data on number of strikes, the
statistics on workers and man-days of idleness are

2 2 2 - 6 1 7 — 66 ------- 1 2


virtually complete. As has been noted, the addi­
tion of new sources of information has not materi­
ally changed these latter figures, but these new
sources have acted to lessen slightly the degree of
comparability in the number of strikes reported
from period to period. It follows that the nar­
rower the classification of strike data, the greater
is the chance of a significant omission. For exam­
ple, while the figures for individual States may be
taken as reasonably complete, the figures for a
specific industry group within a State may be
appreciably affected by the omission of one strike.
Secondary idleness is not measured; that is, the
figures do not cover those employees made idle in
other establishments or industries as a result of
material or service shortages resulting from a work
stoppage. At times, the idleness of employees
directly involved in a strike may be considerably
less than the idleness of other workers brought
about indirectly. No satisfactory method, how­
ever, has been found to measure or estimate such
indirect effects adequately.
The Bureau does not attempt to measure the
cost of strikes in terms of the amount of produc­
tion and wages lost. The calculation of cost in­
volves many complex and interrelated factors for
which information is not readily available, includ­
ing such matters as production schedules before
and after the stoppage, diversion of output or
services to other plants or employers, the flow of
raw materials, and the amount of overtime worked
before and after the strike.11 The problem is mag­
nified beyond statistical control if secondary costs
are to be accounted for.
• For a discussion of the methods used in compiling strike sta­
tistics in other countries, see Methods of Compiling Statistics
of Industrial Disputes (Geneva: International Labour Office,
Studies and Reports, Series N, No. 9, 1926). See also A. M.
Ross and P. T. Hartman, Changing Patterns of Industrial Con­
flict (New York: John Wiley & Sons, 1960).
D ata on work stoppages in other countries are published an­
nually in the Year Booh of Labour Statistics (Geneva: Interna­
tional Labour Office).
10 See article by Joseph W. Bloch, “ The Strike and Discontent,M
Monthly Labor Review, June 1963, pp. 645-651.
11A framework for the measurement of the cost of strikes is
developed by Neil W. Chamberlain and Jan e Metzger Schilling
in The Impact of Strikes: Their Social and Economic Costs
(New Y ork: Harper & Bros., 1954).


Technical References

1. International Labour Office, M ethods o f C om piling S tatistics o f In d u str ia l D isp u tes , Studies and
Reports, Series N (Statistics), No. 9, Geneva (1926).
Compares the methods used to compile statistics on industrial disputes, and outlines stand­
ards by which some degree of international comparability may be secured.
2. Peterson, Florence, “ Methods Used in Strike Statistics,” J o u r n a l o f the A m erican S ta tistic a l A sso cia­
tion , pp. 90-95, 1937.
Summarizes the definitions and methodology utilized by the Bureau of Labor Statistics in
its work stoppage statistics program.
3. U.S. Department of Labor, Bureau of Labor Statistics, A n a ly sis o f Work Sto p p age s {y ear).
The Bureau of Labor Statistics issues annually a detailed statistical bulletin analyzing work
stoppages in the United States.
4. ----- . S trik e s in the U nited State s , 1880 to 1986, Bulletin 651, 1938.
Contains a history of statistics on strikes and lockouts in the United States and major
statistical data available from the earliest recorded date through 1936.

— E dward D. O nanian
Office of Wages and Industrial Relations

Chapter 20. Collective Bargaining Agreements
Collective bargaining agreements and related
documents setting forth the provisions of em­
ployee-benefit plans provide a valuable source of
information on industry wage practices, supple­
mentary benefits, job and union security, the timing
of wage negotiations, the nature of plant opera­
tions and working conditions, and many of the dayto-day aspects of employer-employee and unionmanagement relationships.
The Bureau of Labor Statistics has utilized these
basic industrial relations materials for public and
Government informational purposes in three
major ways, by: (1) maintaining a file of current
agreements open to public inspection and inquiry;
(2) preparing reports which reproduce representa­
tive agreement provisions or the variety of pro­
visions relating to similar problems, or digests of
selected identified agreements or benefit plans; and
(3) by preparing studies measuring the prevalence
and characteristics of specific types of agreement
and benefit plan provisions or of other aspects of
collective bargaining such as multiemployer
The development of industrial relations prac­
tices that are now widely prevalent is reflected
in the Bureau’s studies over the years. The Bu­
reau’s interest in the collection and analysis of
union agreements dates back over 50 years.1 Sys­
tematic efforts to collect agreements began in 1912.
The first of a number of B L S bulletins devoted en­
tirely to the subject of collective bargaining agree­
ments appeared in 1925. A large number of reports
and bulletins, on a wide variety of industrial rela­
tions subjects, have since been published.
The Bureau’s responsibility in the field of agree­
ment collection and analysis received additional
sanction and guidance in the Labor Management
Relations (Taft-Hartley) Act, 1947, section 211,
which reads as follows:
Sec. 211. (a) For the guidance and information of in­
terested representatives of employers, employees, and the
general public, the Bureau of Labor Statistics of the De­
partment of Labor shall maintain a file of copies of all
available collective bargaining agreements and other
available agreements and actions thereunder settling or

adjusting labor disputes. Such file shall be open to in­
spection under appropriate conditions prescribed by the
Secretary of Labor, except that no specific information
submitted in confidence shall be disclosed.
(b) The Bureau of Labor Statistics in the Department
of Labor is authorized to furnish upon request of the
[Federal Mediation and Conciliation] Service, or em­
ployers, employees, or their representatives, all available
data and factual information which may aid in the settle­
ment of any labor dispute, except that no specific informa­
tion submitted in confidence shall be disclosed.

Concepts and Scope
Although the substance of collective bargaining
rests partly upon a foundation of unwritten
industry, company, and union practices, and upon
various legal requirements, the basic unit in agree­
ment collection and analysis is, of necessity, the
written agreement itself. The agreement may
cover a single plant, a number of plants of a multi­
plant company, or a number of companies, in some
cases over a thousand, bound together formally
or informally in an association for collective bar­
gaining purposes. It may express conditions of
employment in simple terms, leaving much of the
administrative details and other matters to the
day-to-day relationships between the parties, or it
may attempt to cover all details and, thus, leave
as little as possible to later bargaining or haggling.
Agreements vary in size from a few sheets to over
300 pages of a pocket-sized booklet, reflecting the
diversity of employment conditions among indus­
tries and companies and of the scope of the issues
over which bargaining takes place, as well as dif­
ferences in the degree of precision sought and the
language used.
Estimates of the number of agreements cur­
rently in effect range upwards of 140,000. The
number of workers covered by agreements is esti­
mated at about 18 million. The Bureau presently
maintains a file of approximately 5,000 current
1 A bulletin of the Department of Labor, Number 42, Septem­
ber 1902, included this n ote: “ It is the purpose of this Depart­
ment to publish from time to time important agreements between
large bodies of employers and employees with regard to wages,
hours of labor, etc. The Department would be pleased to receive
copies of such agreements whenever made.” (p. 1057)
Between 1888 and 1903, the Bureau of Labor (now the Bureau
of Labor Statistics) had independent status as a Department
of Labor, under the direction of a commissioner.




agreements covering about 9.8 million workers.2
All industries are represented in the file with the
exception of railroads and airlines. Since rail­
roads and airlines are required to submit copies
of agreements to the National Mediation Board,
the Bureau does not attempt to collect these
The Bureau’s quantitative analysis of agreement
provisions covers virtually the entire range of is­
sues dealt with in collective bargaining.3 The
basic assumption underlying such analysis is that
the variety of subjects can be defined, classified,
and counted.
In its general analysis of agreements, as distinct
from special industry studies, the Bureau is con­
cerned with these major objectives: (1) the pres­
entation of data by industry group and for
manufacturing and nonmanufacturing as a whole,
(2) the presentation of data by region or union
affiliation if the subject requires it, (3) a realistic
measure of representativeness in the agreements
studied, and (4) the study of practices which are
(as yet) relatively uncommon in collective bar­
gaining agreements.

Methods of Collection and Analysis
Collection of Agreements
The selection of agreements for the file is cur­
rently based on two guides; to enlarge to the full­
est the opportunities for public and governmental
use of the file,4 and to provide a diversified collec­
tion of agreements for special reports, which the
Bureau occasionally is called upon to prepare.
The extent to which these objectives are fulfilled is
obviously affected by the size of the file. A third
guide—that of constructing a file which truly rep­
resents all agreements and thus provides a firmer
basis for sound generalizations on all agreements—
has long been a goal of the Bureau.
The maintenance of a current file of agreements
is a continuous undertaking because of two factors:
(1) The typical agreement has a fixed duration,
after which it is no longer considered current
(unless notice of renewal without change has been
received), and (2) submission of agreements to the
Bureau is voluntary on the part of employers or
unions. To allow for the ratification and the
printing of new agreements, requests for copies

are mailed about 2 or 3 months after the expira­
tion date indicated in the previous agreement or
upon other notice of contract change. As in other
phases of the Bureau’s work, the voluntary coop­
eration of employers and unions is of utmost
importance. Any restrictions imposed by re­
spondents on the public use of agreements are
scrupulously observed by the Bureau.
To facilitate the use of the file in accordance
with the types of requests customarily made, each
agreement received is coded for a series of identi­
fying features, which include: union, location,
number of workers covered, industry, and effective
and expiration dates.

Agreement Analysis
The Bureau’s utilization of the agreements it
collects has moved through different stages over
the years, in pace with, or controlled by, the in­
creasing prevalence and maturity of collective bar­
gaining. During the early years, significant
agreements were reproduced in their entirety.
With the spread of collective bargaining, and the
increase in the size and representative character of
the Bureau’s file, attention was directed towards
reproducing and analyzing the variety of agree­
ment clauses relating to similar subjects, culled
from a large number of agreements. While illus­
trative clauses continue to be utilized in most of
the Bureau’s agreement studies, major emphasis
during recent years has been devoted to measuring
the prevalence and characteristics of particular
provisions and of types and levels of benefits.
The Bureau’s Bulletin 1425 series represents its
most comprehensive efforts to date. It is in this
kind of analysis that problems relating to tech­
niques of coding and analysis come to the fore.
In 1948 and 1949, when the Bureau’s file con­
sisted of more than 12,000 agreements, it was de­
cided that a sample of 3,000 agreements would be
feasible. The selection of specific agreements was
2 During the early postwar period, the number of agreements
on file exceeded 12,000.
* See Program Note by Joseph W. Bloch, “ Union Contracts—
A New Series of Studies,” Monthly Labor Review, October 1964,
pp. 1184-1185.
4 The agreements file is located in the Washington Office of
the Bureau’s Division of Industrial and Labor Relations. Agree­
ments submitted to the Bureau with a stipulated limitation on
public use are not available for inspection. Requests for in­
formation concerning specific agreements or agreement clauses
are accommodated, depending upon the nature of the request,
within the limits of staff resources.

based on a number of factors, including industry,
agreement coverage, location, union representa­
tion, and bargaining practices. Limited data
upon which to base a representative selection of
agreements were compensated for, at least in part,
by extensive experience with collective bargaining
practices on the part of the sample selectors.
During subsequent years, however, available
staff resources were not sufficient to deal with a
sample of this size. The feasibility of reconsti­
tuting a sample of 1,500 to 2,000 agreements, which
had become the maximum work load, and of assuring appropriate safeguards against deterioration,
was rejected as being beyond the resources of the
staff and the available data. The most advanta­
geous alternative, considering all things, was to
base the agreement studies on all agreements cov­
ering 1,000 or more workers and, thus, to avoid
sampling. The Bureau’s file already contained
almost all of these; the Bureau’s monthly report,
Current Wage Developments, was a ready source
of information on those that were not included.
The total number of workers covered by agree­
ments of this size (now between 1,700-1,800) is
about 7.5 million, representing a very substantial
worker coverage in agreement studies. The num­
ber of establishments covered is not known.5
A key analysis list containing all agreements
covering 1,000 or more workers, while perhaps not
the ideal coverage, has definite advantages: (1)
it achieves maximum worker coverage in the stud­
ies for a given investment of staff resources, (2) it
provides a simple, objective measure of the cov­
erage of the studies, (3) it permits the presentation
of various combinations and breakdowns of the
data without the necessity of complicated weight­
ing schemes and without the bias resulting from the
lack of proper weighting, (4) it is safeguarded
against obsolescence, since the Bureau is best able
to keep abreast of changes in agreements of this
size, and (5) it has a significant meaning to users
of these studies.
The distinguishing feature of agreement analy­
sis is that it deals mainly with legalistic language,
“ The distinction between size of agreement (employees cov­
ered) and size of establishment is important. A substantial
proportion of these agreements are association-negotiated and
cover a large number of small establishments. Two agreements,
for example, involving the United Mine Workers ( I n d . ) , cover
most of the anthracite and bituminous coal mines in the country.
Some association agreements in New York State cover more than
a thousand firms.


which requires interpretation, rather than with
numbers or other universal, sharply defined attri­
butes. The process of analysis consists of inter­
preting provisions, reducing them to numbers
(codes), aggregating the numbers, and converting
the aggregates back to types of provisions and
prevalence. Data are presented in terms of num­
ber of agreements and number of workers covered.
Since agreement provisions on the whole are no­
table more for their variety of expression and de­
tails than for their uniformity, the process of
analysis becomes a simplification process by which
some of the original content and variety is lost.
Under such circumstances, the planning of studies
acquires a special importance if significant differ­
ences are not to be buried.
The keystone of agreement analysis study is
obviously the interpretation of the agreement and
the assignment of the predetermined code number.
For some subjects, an agreement must be read in
its entirety; for others, only a portion. Long and
legalistic provisions must be reduced to their
essentials. Since the interpretation of agreement
provisions is often a troublesome matter for the
parties themselves (as reflected in the widespread
adoption of provisions for arbitration), misinter­
pretations undoubtedly occur. These are kept to
a minimum by a staff experienced in agreement
analysis and by continuous efforts to assure con­
sistency of interpretations.

Uses and Limitations
The studies of agreement provisions are of prac­
tical use to companies and unions engaged in col­
lective bargaining, to arbitrators and factfinding
boards, to administrators of company wage and
industrial relations programs, and to legislators
and Government officials. Persons not directly in­
volved in collective bargaining or in related ad­
ministrative functions (e.g., teachers and students
of labor problems, writers for newspapers and
trade and technical journals, and foreign observ­
ers) find value in the broader aspects of employeremployee relationships revealed in these studies.
The limitations of these studies of agreement
provisions and employee-benefit plans are deter­
mined, in large part, by their application. For
studies of paid holiday provisions or other supple-



mentary benefits, the fact that these studies cover
only the area of collective bargaining may con­
stitute a limitation on generalizations applying to
all workers but not necessarily on their uses in
collective bargaining or in wage and employee
administration. On the other hand, these studies
do not show locality practice, which may reduce
their usefulness for some collective bargaining
purposes but not for broad generalizations relating
to workers under collective bargaining.
Additional limitations of agreement provision
studies are inherent in the selection of agreements
for study—the exclusion of railroad and airline
agreements and of agreements covering fewer than

1,000 workers—and in the technique of analysis*
as indicated previously. There are also limita­
tions connected with the particular subjects
studied, which are pointed out in each study. A
fundamental limitation which must be borne in
mind in connection with certain studies (e.g.,
grievance procedure) is that they relate to written
policy rather than actual practice. Practices
which are not provided for in the agreements but
are, instead, matters of company policy going be­
yond the agreement, traditional industry policy,
informal acceptance by management and unions,
or arbitrators’ decisions, can be neither detected
nor measured in agreement analysis.

Technical References

1. Bloch, Joseph W. U nion Contracts — A New Series of Studies. M onthly L a b o r Review , October
1964, pp. 1184-1185.
2. U.S. Department of Labor, Bureau of Labor Statistics, M a jo r Collective B a rg a in in g Agreem ents
([BLS] Bulletin 1425 Series).
1425-1, 1964: Grievance Procedures
1425-2, 1965: Severance Pay and Layoff Benefit Plans
1425-3, 1965: Supplemental Unemployment Benefit Plans and Wage-Employment
1425-4, 1966: Deferred Wage Increase and Escalator Clauses
1425-5, 1966: Management Rights and Union-Management Cooperation
1425-6, 1966: Arbitration Procedures
3. ----- . BLS Bulletin 1353, M a jo r U n ion Contracts in the U nited S tate s , 1961 (1962).

—-Joseph W. B loch
Office of Wages and Industrial Relations

Chapter 21. Union Membership
The Bureau of Labor Statistics interest in union
membership as a significant social and economic
indicator is of long standing. The first B L S
publication listing membership figures for national
and international labor unions in the United States
was probably the Handbook of American TradeOnions, issued in 1926 (B L S Bulletin 420), fol­
lowed by a similar publication in 1929 (B LS Bul­
letin 506) and an extensive revision in 1936 (B LS
Bulletin 618). The Handbooks, however, devoted
little space to overall membership figures or trends;
their main emphasis was on the origins and early
history of particular unions, their government and
structure, trade jurisdiction, and types of beneficial
In 1939, on a modest scale, the Bureau began to
publish an annual trade union directory, but it was
not until 1948 (Bulletin 937) that the listing of
union officers, headquarters’ addresses, etc., was
supplemented with an entry for each union on
membership and number of local unions, resulting
in a compilation of total membership. In subse­
quent years the information sought from national
and international unions has expanded consider­
ably. Data on women members first appeared in
the 1953 Directory , and separate tabulations for
areas outside the U.S. (Canada, Puerto Bico, etc.),
in the 1955 edition. Since 1957, each Directory
has carried information on the number and pro­
portion of members who are white-collar workers
and on those in major industry groups. State
figures were introduced on a limited basis in 1959,
when AFL-CIO central bodies were asked to fur­
nish estimates on the number enrolled by Federa­
tion affiliates. In its 1965 Directory , the Bureau
shows State figures as reported by national and in­
ternational unions. Many of the items referred
to above have been refined since they were first
introduced, and the accumulated information now
1 For a membership survey of these unions, see Unaffiliated
Local and Single-Employer Unions in the United States, 1961
(B L S Bulletin 1348, 1962).
3 The names of all reporting unions appear in the Depart­
ment’s Register of Reporting Organisations, last issued in 1964.

permits analysis of trends in total membership and
several of its components.

Data Sources and Collection
For part of the Bureau’s biennial Directory of
National and International Labor Unions in the
United States , unions meeting the criteria noted
below are asked to report (see copy of question­
naire on pp. 170-172) the average number of duespaying members for the 2 most recent years,
including members outside the United States, the
proportion of white-collar and women members,
breakdowns by major industry and by State, and
other membership information. To be included
in the Directory , a union must be an affiliate of the
AFL-CIO or, in the case of unaffiliated unions, a
party to collective bargaining agreements with
different employers in more than one State.
Among the unions in the latter category are such
unaffiliated unions as the Teamsters and the Mine
Workers. In addition, the Directory accounts for
all unions of Federal Government employees that
have received “exclusive recognition” as specified
in Executive Order 10988. Thus, by definition,
the Bureau excludes from its Directory those un­
ions whose activities are confined to a single lo­
cality or to a single employer.1 Although the Bu­
reau sends questionnaires to all multilocal unions
who have submitted reports to the U.S. Depart­
ment of Labor’s Office of Labor-Management and
Welfare-Pension Beports, as required by the La­
bor-Management Beporting and Discloseure Act,23
it is possible that some small unaffiliated unions, in­
terstate in scope, escape attention. These inad­
vertent omissions do not affect membership totals
in any significant way.
When some unions are unable to furnish in­
formation for one or more of the questionnaire
items, estimates are derived from other sources on
file, notably union periodicals, convention proceed­
ings, financial statements, and collective bargain­
ing agreements. 1STo sampling procedures are
used; the data are based on the entire universe of
national and international unions, as defined.



BLS 2441

Budget Bureau No. 44-6503
Approval expires December 31,1965


Directory of National and International Labor Unions in the United States, 1965
1. Union name and address (please add zip code above if not shown)
2. Affiliation (check appropriate box)
None □
Other (specify) ----------------------------------------------------------------------------------------------------------------------------3. Telephone number (spell exchange in full) ___________________________________________________________
4. President (Mr., Mrs., Miss) -----------------------------------------------------------------------------------------------------------5. Secretary-Treasurer (Mr., Mrs., Miss) ----------------------------------------------------------------------------------------------6. Name and title of person in charge of organizing activities:
(Mr., Mrs., Miss) _______________________________________________________T itle____________________
7. Research Director (Mr., Mrs., Miss) ________________________________________________________________
Mailing address if different from headquarters:



(Zip code)

8. Education Director (Mr., Mrs., Miss) _______________________________________________________________
Mailing address if different from headquarters:



(Zip code)

9. Name and title of person in charge of social insurance (health, insurance, pension, etc.) activities:
(Mr., Mrs., Miss) _______________________________________________________T itle____________________
Mailing address if different from headquarters:





How often published

Number of locals in operation as of the end of 1964:

(Zip code)

___________ T itle__________


1. Frequency of conventions.
2. Next convention_______
3, Name of official publication^)



10. Name and title of person in charge of legal activities:
(Mr., Mrs., Miss) ______________________________
Mailing address if different from headquarters:

(Zip code)




(Mr., Mrs., Miss)




1. Indicate annual average dues-paying membership count for 1963 and 1964. If complete returns for 1964 are
not yet available, use 9- or 10-month average.
2. Indicate the formal policy of the union, as expressed in the constitution or other regulations, concerning the dues
Involved in work stoppages
Armed Forces
Other groups (specify)

Full dues

Less than
full dues


No formal





3. If members in any of the above categories pay less than fu ll dues. is the per capita tax to the national or inter­
national union for those members—
(check one)
Paid in full
Waived ^
Paid in part
No fixed policy
1. Approximate percentage of membership who are women............................................. . ........... ..... (If
none, enter zero)
2. Approximate percentage of membership who are “white collar” workers (include professional,
technical, sales, and office workers)______________________________ ____ _______________ ___________
(If none, enter zero)
3. For any area outside the United States, please indicate the number of dues-paying members and the number of
local unions in existence as of the end of 1964 or any other appropriate current period:

A pproxim ate number o f union members

Canada........_............................... .......... ..........................................................
Puerto Rico......................... .................... ..........................................................
Canal Zone________________________ _____________________________
Other (specify)_____________________ _____ _________________________

Number of local unions


4. Industry composition of union membership. Indicate the approximate percentage of all union members working
in establishments in each of the following industry groups:
Food, beverages, and tobacco--------------------------------- - .............................................. ......... ....................... %
Clothing, textiles, and leather products________ ___________ _______________________ ___________ %
Furniture, lumber, wood products, and paper____________ ____ __________ _________ ___________ %
Printing and publishing_________________ ____ ______ ______ ____ ____ ____ _____ ___________ %
Petroleum, chemicals, and rubber________________________________________________ ___________ %
Stone, clay, and glass--------------------------------------------------------------------------------------- ----------------- %
Metals, machinery, and equipment except transportation equipment--------------------------- ----------------- %
Transportation equipment (automobiles, aircraft, shipbuilding)________________________ ___________ %
Manufacturing (<classification not available)------------------------------------------------------------- ----------------- %
N onmanufacturing:
Mining and quarrying (include crude petroleum and natural gas production)_____________ ___________ %
Contract construction (building and special trade)__________________________________ ___________ %
Transportation (include railroads , air, bus, truck and water transportation , and allied services) ----------------- %
Telephone and telegraph----------------------------------------------------------------------------------- ----------------- %
Public utilities (<electric, gas, and water)___________________________________________ ___________ %
Trade ( wholesale and retail) _______________ ________ - ......... ...........................................- ____________%
Finance and insurance---------------------------------------------— ............................................... ....................... %
Service industries (include hotels, laundries and other personal services, repair services, motion
pictures, amusements and related services, hospitals, educational institutions, nonprofit membership organizations) ___________________________________________________________ ___________ %
Agriculture and fishing.............. ................................................................................................ .......................%
Nonmanufacturing (<classification not available)...............................................— .................... .......................%
Federal_____________________ _________________ ____ ________ ____________ ____ ___________ %
State and local------------------------ --------------------------------------------- ------------ ----------- ------------------%
Total........... .............. ............................................................................................ ...........



V. 5. State distribution of union membership. Indicate the approximate number or percentage of members in each of
the 50 States. Estimates are for general analysis purposes only and will not be shown for individual unions.
Nebraska___________________ —
Nevada________ ___ ___________
Alaska________________ _
New Hampshire ________________
Arizona_____ ____________
New Jersey.___________________
Arkansas__ __________*__
New Mexico____________________
California______________ _
New York_____ _______________
North Carolina________________ _
North Dakota______________ ___
Florida......... .................................
Oklahoma___ . . . ______________ _
Georgia_____ — -------------Oregon_________________________
Hawaii_____ ____________
Pennsylvania___ ______________ Id ah o .__________________
Rhode Island______ ____ _______
Illinois__________ _______
South Carolina_________________
Indiana____________ ____
South Dakota__________________
Iowa__________________ _
Tennessee____ ___ _____________
Kansas________ ________
Texas__________________ _____ _
Utah_______ ____ _____________
Louisiana___________ ____
Vermont___________ ___________
Virginia..__________ __ __ ______
Maryland-Distriet of Columbia.
Washington_____________ ______
Massachusetts_______ ____..
West Virginia___________________
Michigan________ _______
Wisconsin- - ................ ....... .......... ..
Minnesota____________ __
Wyoming_________ _____ ______
Mississippi________ ______
Number or percentage not accounted
Missouri______ ________ ...
for_________ _______ ______
1. Approximate number of basic collective bargaining agreements with employers (do not
include various supplements, pension, health, or insurance agreements as separate docu­
United States only__ ________ _______________________ ____________ ___ ________ _ agreements
2. (a) Approximate number of different employers covered by collective bargaining agree­
ments----------------------------------------------------------------- -------------------------------- -------------- employers
United States only________________________ ___________________________ _________ employers
(b) If more than one employer, are the employers located in at least two States?_________ Yes Q No Q
3. Approximate number of workers covered by all collective bargaining agreements (in­
clude nonmembers in bargaining units)_ -_____________ __________ ___________ _ _________ workers
United States only..__________________ __________ ___ ___________ ___ _________ workers
1. Approximate number of full-time paid employees (executive, professional, representa­
tives, organizers, clerical, etc.) on the payroll of the national or international union in
the United States (omit those on the payroll of local unions, intermediate or other
bodies, etc.)________ _________ ___________ _______________________ __________ _____ . . employees
May we have your comments regarding the present Directory and proposals for changes in future editions?
Name of person reporting




The data for each union and summaries are pre­
sented in the Directory and in articles in the
Monthly Labor Review . Wherever appropriate,
separate data are presented for AFL-CIO affiliates
and unaffiliated unions and cover such question­
naire items as total membership for the last 2 years,
members in and outside the United States, size of
unions, women and white-collar members, and
industry distribution. Gains and losses over the
past decade are analyzed, and major unions are
ranked according to size at particular points in
time. Of considerable interest to users of the
Directory are the changes of union membership
related to changes in the total labor force and to
employees in nonagricultural establishments.
Data by industry and on women and white-collar
members permit rough approximations on the ex­
tent of union penetration as compared with earlier
In addition to statistical summaries, the Direc­
tory also contains appendixes which list, for each
union, membership in areas outside the United
States, number of women, proportion of whitecollar members, and proportion of members in ma­
jor industry groups.

Uses and Limitations
The Bureau’s membership figures are published
on a regular basis, and serve as the principal indi­
cator of gains and losses for particular unions and
for the labor movement as a whole, particularly as
a measure of inroads or declines in industry and
occupational groups. Comparisons of labor rela­
tions policies and their economic effects can be
made between organized and unorganized sectors.
They are used by agencies of the Federal Govern­
ment, State and local governments, by manage­
ment personnel, union officials, students of the
labor movement and economic affairs, and the gen­
eral public. The Directory also is distributed
widely abroad, notably to international organiza­
tions and labor unions.
Difficulties in measuring union membership
arise from (1) the variety of concepts and prac­
tices among unions as to the definition and report­


ing of membership, (2) the availability at union
headquarters of the various data requested, and
(3) the willingness of the unions to make those
data available to outsiders.
In an attempt to achieve uniform reporting
practices, the Bureau asks unions to report on the
annual average number of dues-paying members.
Although a worker when joining a union assumes
an obligation to pay dues, uniform reporting
practices do not result from applying this criterion
alone. Unions define eligibility for membership
in a variety of ways, with payment of dues being
only one of several criteria. Some unions set less
than full dues requirements or waive the payment
for workers who are unemployed or on strike.
Such exonerated workers, however, usually remain
members in good standing, with the same rights as
full dues-paying members. Similar qualifications
may apply to members who are apprentices, retired,
or in the Armed Forces.3
In an attempt to determine union practices in
reporting membership, the Bureau has repeatedly
requested unions to indicate whether they include
or exclude from membership reports five specified
groups: the unemployed; those involved on work
stoppages; those in the Armed Forces; appren­
tices; and the retired. Moreover, unions were
asked to furnish an estimated or actual figure on
the number of members in “excluded” categories.
Ideally, if all unions could furnish such data, it
would be possible to compute the total number of
workers who are, at least in some way, still at­
tached to unions. The responses, however, have
3 In its 1963 Directory, the Bureau took a closer look at the
relationship between dues submitted to the international union,
the so-called per capita tax, and reported membership totals for
particular unions. Per capita tax receipts were divided by the
per capita tax rate. Briefly, the findings indicated that while a
number of unions use a “ per capita” receipt figure in their
reports to the Bureau, such an approach was inappropriate in
the case of other unions for reasons which the unions explained
in detail. Some unions include a large number of seasonal em­
ployees whose dues payments are limited to several months during
the year. Other unions questioned the use of a computed member­
ship figure where for large groups of workers (sick, unemployed,
those promoted out of the bargaining unit, etc.) only a partial
per capita tax is paid, at times less than one-tenth of the amount
required of other members. A few unions indicated categories
of members completely dues exempt (e.g., life members, 50-year
members, etc.). Financial obligations also frequently are waived
for workers recently organized and for local unions in economic
difficulties. In addition, several unions set dues on a sliding scale
proportionate to the income of members, a method which rules
out computations of this sort. Thus, on examination it became
clear that computing membership by dividing the tax rate into
tax receipts could not be used as a uniform yardstick applicable
for all unions.



fallen far short of this goal. Only a small number
of unions reported, in whole or in part, the prac­
tices they followed. An even smaller number of
unions were able to furnish figures on the number
of workers involved.4
Thus, there is no uniform answer to the decep­
tively simple question: who is a union member?
The answer varies from union to union, as deter­
mined by its own policies and practices.
While financial statements may be of some help
in arriving at membership approximations, they
cannot be used to obtain data on various compo­
nents of union membership, such as the proportion
4 At various times, suggestions have been made on the use of
alternative sources for membership information, notably the use
of “ voting strength” of international unions a t AFL-CIO con­
ventions, based upon average paid membership to the Federation.
This, however, is not a reliable measure. Some unions, for rea­
sons of prestige, tend to overpay, while others, as a matter of
convenience, pay an arbitrary amount. Moreover, this method
could not be used for unions outside the AFD-CIO.

of white-collar and women members, those under
contract in particular industries, etc. For such
data, the Bureau must rely entirely on the coop­
eration of national unions. These, however, are
not always able to furnish the information for the
simple reason that it is not compiled at union head­
quarters. Many national unions are decentralized
organizations and do not, as a rule, seek reports
on these items from their local unions. Also, some
unions in possession of such information may be
reluctant to make it public.
Data presently compiled are submitted by offi­
cials of national and international unions. Per­
haps more accurate data, and certainly more de­
tailed data, could be obtained by contacting local
unions, as is done in Canada, but the task of
soliciting responses from more than 70,000 organi­
zations is beyond the present resources of the

Technical References
N um ber

1. Troy, Leo. D istribution o f U nion M em bership A m ong the S ta te s , 1939 an d 1953. New York,
National Bureau of Economic Research, 1957.
A study measuring union growth by State and region, analyzing geographic and industrial
shifts in membership. Includes a discussion of sources and methods of measurement.
2. U.S. Department of Labor, Bureau of Labor Statistics. “Membership of Labor Unions, 18971950.” Handbook of L ab o r S ta tistic s , 1950 ed., Bulletin 1016, 1951, pp. 137-139.
3. ----- . D irectories o f L ab o r U n ion s in the U nited S ta te s; Bulletins 937 (1948); 980 (1950); and
1127 (1953).
4. ----- . D irectories o f N a tio n a l an d In tern atio n al L ab o r U n ion s in the U nited S ta te s; Bulletins
1185 (1955); 1222 (1957); 1267 (1960); 1320 (1962); and 1395 (1964). In various articles in
M onthly L ab o r Review. (Excerpted)
In addition to membership data, the D irectories also include a brief description of the methods
used, and a copy of the questionnaire.
5. Paschell, William, “Limitations of Union Membership Data,” M onthly L ab o r Review , November
1955, pp. 1265-1269.
6. Wolman, Leo. Ebb an d Flow in Trade U n ion ism , New York, National Bureau of Economic Re­
search, 1936.
Of invaluable aid to students in the field because of its extensive discussion of measurement
problems. Appendix tables present figures for the period 1900 to 1934 including data on the
extent of organization by industry.

— H akry

P. C ohany

Office of Wages and Industrial Relations

Productivity and Technological Developments
Chapter 22. Output Per Man-Hour Measures: Private Economy
Background and Description
of Measures
In order to fill a growing need for information
concerning the relationships between productivity,
prices, wages, employment, and growth of the
economy, the Bureau of Labor Statistics annually
publishes indexes of output per man-hour and re­
lated data for broad sectors of the economy. Meas­
ures have been developed for the total private econ­
omy and the agricultural and nonagricultura]
sectors for the period from 1909 to the present
time. Comparable measures have been developed
for the manufacturing and nonmanufacturing
segments of the nonagricultural sector for the post­
war period. These measures, first published in
1959,1 represent the culmination of a long series
of developments in productivity measurement in
the Bureau.2
The Bureau of Labor Statistics output per manhour measures for the total private economy and
major sectors refer specifically to the ratio between
constant-dollar gross national product (GNP)
originating in the private economy, or individual
sector, and the corresponding hours of all persons
employed. An index of output per man-hour
shows the change in the ratio of output to labor
input (man-hours).
Two sets of output per man-hour estimates have
been developed. One set of estimates is based on
labor force data from a survey of households,
conducted each month by the Bureau of the Census
for the Bureau of Labor Statistics. The other
1 Trends in Output Per Man-Hour in the Private Economy,
1909-1958, B L S Bulletin No. 1249, 1959.
2 For a detailed description of the historical development of
the productivity measurement program of the BLS, see chapter
8 A detailed description of the methods and procedures of
estimating gross national product in current and constant dol­
lars is given in the 1954 National Income Supplement to the
Survey of Current Business, U.S. Department of Commerce.
Further information on GNP estim ates for major industry sec­
tors is presented in the October 1962 issue of Survey of Current

set is based primarily on a monthly B L S survey
of establishment payroll records.
In concept, the output per man-hour estimates
based on labor force data are defined as output
per hour worked , while those based on establish­
ment data are defined essentially as output per
hour paid . Theoretically, the difference between
the two measures is equal to paid vacation time
and other paid leave. Actual differences between
the two sets of estimates can result from statistical
variation and differences in statistical method­
ology and reporting, as well as differences in

Data Sources and Estimating
Output. The output concept used in preparing
output per man-hour estimates is that of real gross
national product originating in the private econ­
omy or individual sectors.
Gross national product is the market value of
final goods and services produced in the economy
within a certain time period. It includes pur­
chases of goods and services by consumers, gross
private domestic investment, net foreign invest­
ment, and purchases by Government. GNP also
is equal to income received by labor and property
for services rendered in the current production of
goods and services, plus capital consumption allow­
ances, indirect business taxes, and several other
minor items.
Gross national product in current dollars can­
not be used directly as the output measure since
it would reflect price changes in addition to
changes in physical volume. The Office of Busi­
ness Economics, U.S. Department of Commerce,
prepares estimates of constant-dollar GNP for
the total private economy and major sectors.
These estimates are used in developing output per
man-hour measures.3



Labor Input (Man-Hours ). As mentioned previ­
ously, there are two sets of estimates of labor in­
put for measuring output per man-hour. One set
is based on labor force data derived from a survey
of households; the other set is based on the B L S
survey of establishments. Both sets of data are
published monthly in Employment and Earnings
and Monthly Report on the Labor Force . Annual
summaries also are published for both series.
Labor Force Series Man-Hours . The labor force
estimates provide information on the total number
of persons engaged in production, including wage
and salary workers, self-employed, and unpaid
family workers. Since the labor force estimates
do not provide detailed industry information, the
output per man-hour estimates based on this series
are prepared only for the total private economy
and the broad sectors of agriculture and total
The data on employment and hours are obtained
from a monthly survey of the noninstitutional
civilian population of the United States. The sur­
vey is conducted by the Bureau of the Census, but
the labor force statistics are processed and pub­
lished by the Bureau of Labor Statistics. The in­
formation is collected by trained interviewers from
a scientifically selected sample of about 35,000
households. The data collected relate to the activ­
ity of the calendar week including the 12th of the
month. For methodology, see chapter 1.
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. The estimates of total
hours worked are derived by multiplying average
hours worked times the number of persons at work.
Several adjustments to the basic labor force data
are necessary to adapt them for productivity meas­
urement. Since labor force man-hours apply to
the whole economy, it is necessary to subtract the
man-hours of general Government in order to ob­
tain an estimate for the private economy. An ad­
justment is made also to eliminate the effect of
holidays which occur during the week of the sur­
vey. This adjustment is necessary because the data
for the survey week are taken as representative
of the whole month. Several other minor adjust­
ments are made to provide better comparability
of the data over long periods of time.

Establishment Series Man-Hours . In contrast to

the broad classes of labor force data, the B L S
establishment series provide considerable detail
on the industrial composition of wage and salary
employment. Their coverage, however, is more
limited than the labor force estimates, in that they
exclude agricultural employment, household work­
ers, self-employed, and unpaid family workers.
Estimates of average weekly hours of production
workers and nonsupervisory workers are provided
for most of the individual industries and sectors;
in the establishment survey, but there are signifi­
cant gaps, particularly in the finance and services
industry categories. Average weekly hours data
are not available for nonproduction workers. The
establishment data are supplemented from other
sources in order to fill the information gaps and
provide estimates of man-hours for the total priv­
ate economy.
The establishment hours and employment esti­
mates are based largely on the monthly B L S sur­
vey of establishment payroll records. Since this
survey does not cover total employment in the
private economy, and because there are some gaps
in the hours information, it is necessary to use
some supplementary data to derive man-hours esti­
mates for the total private economy.
The B L S establishment survey is based on a
sample stratified by size of establishment. Within
each industry, the sample is an optimum alloca­
tion design with the probability of selection of an
establishment proportionate to average employ­
ment per establishment within each of the strata.
The data generally refer to employees who received
pay during the pay period including the 12th of
the month. For methodology, see chapter 2.
The establishment man-hours estimates are
based on an hours paid concept. These estimates
include paid holidays, vacations, sick leave (except
that paid for under insurance plans) and other
time off paid for by the employer, in addition to
actual hours worked. Total man-hours figures are
derived for each industry and then aggregated to
the total private economy or sector level. Industry
man hours are estimated by multiplying employ­
ment times average weekly hours. While estimates
of the number of persons on establishment payrolls
are available from the establishment survey, aver­
age weekly hours are not reported for some in­
dustries. These hours are estimated or derived
from other sources, usually the labor force data.

Man-hours information reported in the establish­
ment survey cover only production and related
workers, construction workers, and nonsupervisory
employees. For all industries except manufac­
turing, average weekly hours of supervisory and
nonproduction workers are assumed to be the same
as those reported for production workers and non­
supervisory workers. For manufacturing nonpro­
duction workers, average hours estimates are de­
rived from other sources.4
Since the B L S establishment survey covers
employees of nonfarm establishments only, em­
ployment and hours data for such groups as pro­
prietors and self-employed persons, unpaid family
workers, farm workers, and domestic workers in
households must be obtained from other sources.
Some of these data are taken from the employment
series published by the Office of Business Eco­
nomics, U.S Department of Commerce, but most
are obtained from the labor force survey.

Analysis and Presentation
Indexes of output per man-hour show changes
in the ratio of output to labor input (man-hours).
Although an output per man-hour index relates
output to man-hours, the index should not be in­
terpreted as representing only the contribution of
labor to production. Bather, it reflects the com­
bined influence of many things such as changes in
technology, capital investment, rate of plant utili­
zation, managerial efficiency, and scale of opera­
tions, as well as skill and effort of the work force.
Indexes of output per man-hour also are influ­
enced by shifts in the relative importance of
components with different levels of output per
man-hour. For example, an increase in the rela­
tive importance of a sector with a comparatively
high level of output per man-hour will cause a
rise in output per man-hour for the total private
economy even though output per man-hour in each
individual sector remains unchanged. Shifts cani
occur at all levels of the economy, including occu­
pations, products, plants, industries, or sectors.
It is necessary to consider both long- and short­
term trends in output per man-hour and related
variables, in addition to annual changes. Trend
measures, however, may be affected by such factors
as the period selected for measurement and the


type of statistical description used (e.g. least
squares or compound interest averages). There­
fore, some element of judgment is present in any
analysis of trend.
Indexes of output per man-hour and related
data are published periodically in B L S releases.
The historical indexes and detailed procedures for
developing output per man-hour indexes are pre­
sented in Trends in Output per Man-Hour in the
Private Economy , 1909-1958, B L S Bulletin 1249,
1960. Some of the current indexes also are pub­
lished in the statistical supplement to the Monthly
Labor Review , the Statistical Abstract of the
United States, the Economic Report of the Presi­
dent, and the Manpower Report of the President .

Uses and Limitations
The indexes of output per man-hour are de­
signed for use in economic analysis and policy
planning, both public and private. They have
applications in such areas as economic growth,
manpower requirements, wages, prices, living
standards, foreign trade, and measuring effects of
technological changes.
One example of the uses of output per man-hour
indexes and related data is the study of relation­
ships between productivity, wages, prices, profits,
and costs of production. Within the framework of
national income and product accounting, gross
national product represents the market value of
all final goods and services produced in the econ­
omy, or the sum of all costs of production, both
measured for a specific time period. These costs
of production can be separated into labor and
nonlabor costs (or payments). An increase in
labor costs per unit of output reflects an increase
in average hourly compensation in excess of the
4 For manufacturing nonproduction workers, average hours are
estimated as follow s: Estim ates of vacation time, holidays, paid
sick leave, and personal time off are subtracted from an estimate
of scheduled annual hours paid nonproduction workers. Sched­
uled annual hours are derived by extrapolating the 1959 level
of scheduled weekly hours with data from B L S Community Wage
Surveys, and then multiplying by the number of workweeks in
each year. The 1959 level of scheduled weekly hours for non­
production workers was calculated from data collected by BL S
in connection with the study, Employer Expenditures for
Selected Supplementary Remuneration Practices for Production
Workers in Manufacturing Industries, 1959, B L S Bulletin 1308,
1962. The estimates of vacation time, holidays, sick leave,
and personal time off are derived primarily from data from
Community Wage Surveys and Social Security Administration



gain in output per man-hour. A decrease in unit
labor costs, on the other hand, shows that output
per man-hour increased more than average hourly
compensation. It is in this sense that output per
man-hour is a crucial element in the wage-price
relationship. It indicates the extent to which in­
creases in compensation can take place without an
increase in prices or a reduction in the share of
nonlabor payments or costs.
The Bureau of Labor Statistics prepares indexes
of labor and nonlabor payments per unit of out­
put and related measures corresponding to the
estalishment series output per man-hour measures
described in an earlier part of this section.5 Labor
payments, as used in these measures, are equal to
total compensation of all employees in the private
nonfarm economy. Nonlabor payments represent
the difference between compensation of employees
and gross product originating in the private nonfarm economy. The agricultural sector is ex­
cluded from these measures because proprietors’
income accounts for a large portion of agricultural
income and it is difficult to distribute proprietors’
income between labor and nonlabor income.
6 See pp. 176-177.

The labor and nonlabor payments indexes are
published periodically in B L S releases and in the
statistical supplement to the Monthly Labor
Review and Statistical Abstract of the United
States. Analytical notes concerning these meas­
ures appear in the B L S release Productivity ,
Earnings , Costs , and Prices in the Private Nonagricultural Sector of the Economy , 19^7-56 (R e­
vised ), dated May 29, 1957. Subsequent releases
coiitain a description of those measures and com­
ments concerning their interpretation.
Certain characteristics of the output per man­
hour measures should be recognized when apply­
ing them to specific situations. First, the indexes
reflect not only changes in the various component
industries but also changes in the relative import­
ance of these industries. Second, the indexes are
after-the-fact representations of economic events;
they do not show causal relationships between the
variables. Third, the measures of productivity,
output, and input, are limited both by the under­
lying concepts and the data available for estima­
tion. Therefore, the meaning of index changes
should be interpreted with caution, keeping in
mind the conceptual and practical limitations as
well as the possibilities of statistical error.


Technical References
Num ber

1. Fabricant, Solomon. “ Which Productivity? Perspective on a Current Question.” M onthly
L ab or Review , June 1962, pp. 609-613.
A discussion of various concepts and measures of productivity and the uses of different types
of measures.
2. Kendrick, John W. Productivity T rends in the U nited States (National Bureau of Economic Re­
search, Number 71, General Series), Princeton, N.J., Princeton University Press, 1961.
A presentation of historical measures of output, input, and productivity for the U.S. economy
and industry groups, including descriptions of concepts and methods of measurement. Also
includes discussion of implications of productivity change for economic growth, prices, incomes,
and resource allocation.
3. National Bureau of Economic Research, Conference on Research in Income and Wealth. Output ,
In p u t , an d Productivity M easurem ent. (Studies in Income and Wealth, Volume Twenty-five),
Princeton, N.J. Princeton University Press, 1961.
A collection of papers and comments devoted to an appraisal of the measurement of output,
input, and productivity.
4. U.S. Department of Labor, Bureau of Labor Statistics. T rends in Output per M an -H o u r in the
Private Econom y , 1 9 0 9 -1 9 5 8 (BLS Bulletin 1249, 1959).
A presentation and analysis of output per man-hour indexes and trends for the total private
U.S. economy and major sectors. Includes a description of methods and sources for developing
output per man-hour measures.
5. ----- . Productivity: A B ib lio graph y , (BLS Bulletin 1226, 1958).
A collection of nearly 900 references concerning productivity and productivity measurement.
Each reference includes a brief annotation giving the gist of the subjects covered.
6. ----- . Productivity , E a r n in g s , Costs an d P ric es in the P rivate N on agricu ltu ral Sector o f the Econom y,
1 9 4 7 -6 6 , (Revised). May 1957.
A description of the concepts and methods used in developing the BLS labor and nonlabor
payments series, including a discussion of the limitations and qualifications of the indexes.

—J erom e A. M a r k


V ir g il F . K

e t t e r lin g

Office of Productivity, Technology and Growth

Chapter 23. Output Per Man-Hour Measures: Industries
Studies of output per man-hour for individual
industries have long been a part of the B L S pro­
gram. The first Commissioner of Labor, Carroll
D. Wright, conducted a study of 60 manufacturing
industries in 1898. The findings in his report on
“Hand and Machine Labor,” provided striking
evidence of the savings in labor resulting from
mechanization in the last half of the 19th century.
Commissioner Wright’s study was prompted by
concern on the part of Congress that human labor
was being displaced by machinery. The impact of
productivity advance upon employment remained
an important focus of the B L S program through­
out the 1920’s and 1930’s. It was also during this
period after World War I that the Bureau began
the preparation and publication of industry
indexes of output per man-hour, based on available
production data from the periodic Census of Manu­
factures and employment statistics collected by
the BLS.
In 1940, Congress authorized the Bureau of
Labor Statistics to undertake continuing studies
of productivity and technological changes. The
Bureau extended earlier indexes of output per
man-hour developed by the National Research
Project of the Works Projects Administration,
and published measures for selected industries.
This work was reduced in volume during World
War II, owing to the lack of meaningful produc­
tion and man-hour data for many manufacturing
The advent of World War I I also caused
a change in program emphasis, from problems of
unemployment to concern with the most efficient
utilization of scarce manpower. The B L S under­
took a number of studies of labor requirements for
defense industries, such as synthetic rubber and
shipbuilding. After the war, the industry studies
program resumed on a regular basis, and was sup­
plemented by a number of industry studies based
on the direct collection of data from employers.
Budget restrictions after 1952 have prevented the
continuation of direct collection of data. Con­

sequently, the preparation of industry measures is
limited to those industries where readily available
data can be utilized to construct measures.
In recent years, public interest in productivity
has grown, and there has been greater recognition
that increases in output per man-hour are impor­
tant indicators of economic progress and the means
to higher levels of income, rather than merely a
threat to job opportunities.
The Industry Studies program covers a variety
of manufacturing and nonmanufacturing indus­
tries. For these industries, indexes of output per
man-hour, output per employee, unit labor require­
ments and the related data on output, employment,
and man-hours are prepared and published on an
annual basis. The indexes are generally available
for most years from 1947 to the most recent year
for which data are available, and for many indus­
tries also for 1939.

Industry i n d e x e s of output per man-hour
measure changes in the relationship between the
physical volume of output of an industry and the
man-hours expended in that output. Although,
traditionally, output per man-hour has been the
measure most frequently used, discussion of physi­
cal output per man-hour often is simplified if
conducted in terms of its reciprocal: man-hour
requirements per unit of output (unit man-hours).
Therefore, this form of index is used in the fol­
lowing description.
For an industry producing a single uniform
output, the unit man-hours index is simply the
ratio of the man-hours expended to produce a
unit of output over two periods of time. This
ratio may be expressed as follows:

Where I u represents the unit man-hour index, I p
represents the output per man-hour index, and h
and lo denote unit man-hours expended in the
current and base periods, respectively.

For an industry producing a number of pro­
ducts—the more typical case—the unit man- hours
index is the ratio for two periods of the total
hours required for the output of a given composite
of products. Indexes of such industries vary with
the composite of products specified and can take
many forms. Letting q0 and qi represent base
period and current period quantities of a given
product, respectively, two of these forms are:
a. Using a current period composite

2 QoU
U~ ? ( L o l o

An index constructed according to (a) compares
the man-hours expended in the production of the
current composite with man-hours which would
have been required to produce the current com­
posite in the base period. An index constructed
according to (b) compares the man-hours required
in both periods to produce the base period com­
posite. These indexes thus eliminate the effects of
variations over time in the relative importance of
products on unit man-hours.
In either form, an index of unit man-hours also
can be viewed as the quotient of an index of man­
hours and an index of output, i.e.,
Man-hours ^Output index__ Unit man-hours
* (Lasp eyres)
index (Paasche)
/ _


hit .

2/ loQi

Man-hours Outp ut index _ Unit man-hours index
* (Paasche)


u~ 2 l 0q0





Z ifa


'2 i ljq 0

2 l 0q0

The man-hours index measures the change in
aggregate man-hours between the base and current
periods. The man-hours data are the total hours
expended by employees in establishments classified
in the industry, in producing the base period and
current period composites.
As can be seen in the formulas, the appropriate
output index is one which compares the quantities
of the various products in the current and the base
periods, each weighted by the man-hours expended

per unit produced in a given period. A current
period weighted unit man-hours index employs a
base period weighted output index divided into the
man-hours index. Conversely, a base period
weighted unit man-hours index is consistent with
an output index which utilizes current period

Methods and Sources
Output Per Man-Hour

Z f r lo

b. Using a base period composite



The Bureau of Labor Statistics computes an in­
dex of output per man-hour by dividing an output
index by an index of aggregate man-hours. Meas­
ures are prepared separately relating output to
(a) all employee man-hours, (b) production work­
er man-hours, and (c) nonproduction worker man­
hours. (The standard definitions of production
workers and nonproduction workers are used.)
Three corresponding measures also are computed
relating output to the number of employees. Re­
ciprocal measures of unit man-hours also are

B L S industry output indexes are primarily
based on the physical output of the products of
the industry combined with fixed period weights.
However, the availability of quantity data on
physical output varies among industries, and, for
manufacturing and mining industries, may vary
depending on whether the data are for a year when
a Census was conducted or for a noncensus year.
For manufacturing and mining industries, quan­
tity data on physical output are usually most com­
prehensive 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 output
data (generally, in less detail than for Census
years) or if such data are not available, value of
output adjusted for price change (i.e., the value
of output in constant dollars). The annual series
subsequently are adjusted to the benchmark levels
for the census years.



Weights. The mathematical form of the output
index implies use of man-hour weights, and such
weights are used whenever possible. In most in­
dustries, however, unit man-hour information is
not available for individual products. Conse­
quently, the B L S uses substitute weights which
are assumed proportional to unit man-hour
weights. Generally, when unit man-hour weights
are not available, unit value weights are used as
substitute weights.
For those industries where unit man-hour
weights are used, the unit man-hours data are ob­
tained either from special surveys or are derived
from data published for specialized establishments
in the Census of Manufactures. Where unit man­
hours are derived from Census data, the following
procedure is used: Man-hours per dollar of ship­
ments are derived for each 5-digit product group
from statistics for establishments specializing in
that group. Unit values of the products are com­
puted from the “wherever made” data, i.e., the data
for all establishments—regardless of industry clas­
sification—producing the product. The multipli­
cation of each unit value by the man-hours per
dollar of shipments yields an estimate of man­
hours per unit. This procedure is used only when
the “specialization” and “coverage” ratios of the
industry are high and specialization data for all or
most of the product groups are available.1
For industries where unit value weights are used,
the weights are computed from census or survey
data on the quantity and value of shipments of the
primary products of the industry. The introduc­
tion of these substitute weights results in an in­
dustry output per man-hour index which reflects
shifts in value per man-hour of the various prod­
ucts in the industry. Thus, a change can occur in
the index without any change in the output per
man-hour for any product of the industry.
The extent to which error or bias may be intro­
duced by the use of unit value weights is not
1 The “ specialization ratio” is the value of shipments of
primary products of plants in the industry as a percent of total
shipments of all products (primary plus secondary) made by these
same establishments. The “ coverage ratio” is the value of ship­
ments of the primary products made by plants classified in the
industry as a percent of the total shipments of the industry’s
primary products made by all producers, both in and out of the
specified industry.
3 See Irving H. Siegel, “Further Notes on the Difference Between
Index Number Form ulas,” Journal of the American Statistical
Association, December 1941, pp. 519-5(24.

known. The index is equivalent to one weighted
with unit man-hours if the unit man-hours and
unit values of the products are proportional or
if there is no correlation between the relative
change in quantity and value per man-hour.23
There is evidence that unit values are fairly reli­
able approximations for individual products where
wages constitute a large proportion of total value
of output. The error generated in the output index
by an error in the weights is generally consider­
ably smaller than the error in the weights
Most published industry indexes have used 1947
weights for 1947-58,1958 weights for 1959-63, and
1963 weights for years after 1963. The Bureau
policy is to revise the weights as more current data
become available from the periodic censuses.
For manufacturing and
mining industries, indexes reflecting changes in
output between census years are constructed.
For industries where unit man-hour weights are
used, the index is based on the weighted aggre­
gates derived from the physical output of primary
products “wherever made,” and the corresponding
weights. No adjustment is made for changes in
specialization and coverage ratios since unit man­
hour weights are used only where the specializa­
tion and coverage ratios are high.
For industries where unit value weights are
used, the benchmark index is constructed by the
following procedure: A physical quantity output
index is derived from quantity data for the pri­
mary products of the industry, whether made in
the industry or elsewhere. An index of the total
dollar value of primary products also is computed.
This value index is divided by the physical output
index to obtain a price index for the primary prod­
ucts. This primary product price index is then
used to convert an index of the value of industry
output, i.e., of primary and secondary products, to
constant dollar estimates by dividing the value of
output index by the price index. The resulting
index is comparable conceptually with the physical
output index with unit value weights. The use
of the primary product price index is based on the
assumption that the price movements of all prod­
ucts (primary and secondary) produced by the
Benchmark Indexes.

industry are the same as the price movements for
all primary products wherever made.3
For the mining industries, benchmark indexes
are computed from unweighted tonnage data as
reported in the Census of Mineral Industries.
Annual Indexes. Annual output indexes are con­

structed by the following described procedures.
For manufacturing and mining industries, the
annual indexes are adjusted, if necessary, to the
levels of the benchmark indexes previously
described. The adjustment factors for 2 census
years are used to determine the adjustment factors
for the intervening years by linear interpolation.
1. jPhysical output. Most annual output in­
dexes are based on physical quantities of products
combined with fixed-period unit man-hour or unit
value weights. The basic quantity data are gen­
erally primary products of an industry classified
into product groups; the greatest amount of detail
available is used. The quantity data relate to pri­
mary products “wherever made” and in some cases
to shipments of the products.
The Bureau’s annual measures of production are
constructed from data on physical quantities of
products comprising a high percentage of the total
value of an industry’s output. Coverage varies
between 60 and 100 percent. Complete coverage
generally is obtained in mining and other welldefined industries with a relatively homogeneous
2. Deflated value. When adequate annual
physical output data are not available, indexes are
derived from data on the value of industry out­
put, adjusted for price change. Since the adjust­
ment for price change is most often downward,
the indexes usually are called “deflated value”
indexes. Such indexes are conceptually equivalent
to indexes which use data on physical quantities of
products combined with unit value weights. To
derive this index, data on the industry’s value of
output are divided by an industry price index. An
index of these deflated values shows the change in
the real value of output between the past and cur­
rent periods.4
Often data on value of production are not avail­
able and data on value of shipments must be used.
In this case, data on value of shipments for each
year are divided by an industry price index rep­
resenting the average annual price for the year.


Beginning- and end-of-year inventories are
divided by a price index relating to December of
each year. The estimated value of shipments in
constant dollars is then adjusted by the net change
in finished inventories, also in constant dollars, to
yield an estimate of the constant dollar value of
Industry output indexes are prepared
from basic data published by various public and
private agencies, using the greatest amount of
detail available.
The Bureau of the Census, U.S. Department of
Commerce, is the major source of output statistics
for the manufacturing industries. The Bureau of
Mines, U.S. Department of the Interior, compiles
most of the information for mining, cement, coke,
and metals industries. Other important Govern­
ment sources include the U.S. Department of Agri­
culture, the Fish and Wildlife Service, U.S. De­
partment of the Interior, the Interstate Commerce
Commission, and the Internal Revenue Service.
Important sources of trade association data in­
clude the Textile Economics Bureau, Inc., National
Association of Hosiery Manufacturers, Inc., Na­
tional Oanners Association, Rubber Manufacturers
Association, and the American Iron and Steel
For deflated value series, industry price indexes
are derived from wholesale price indexes published
by the Bureau of Labor Statistics.


An index of man-hours is computed by dividing
the aggregate man-hours for each year by the baseperiod aggregate. Man-hours are treated as
8 See 1954 Census of Manufacturess volume IV, “ Indexes of
Production,” chapter 3.
It is currently planned to modify the benchmark procedure to
include the actual price movements of secondary products.
* For example:
Value index + Price index (Paasche) —Output index (Laspeyres)







2 p ,g *

where pt and p* represent prices of products in the industry in the current
and base periods, respectively. This index requires quantities of all items
produced in each year. These data are not available for the particular indus­
tries where this measure is used, and quantity data are usually available for
the base year only, so that the deflated value indexes employed usually take
the following form:
Value Index -5- Price index (Laspeyres) ■■ Output index (Paasche)








homogeneous and additive with no distinction
made between hours of different groups of em­
ployees. Data on changes in qualitative aspects
of man-hours, such as skill, efficiency, health, ex­
perience, age, and sex of persons comprising the
aggregate, are not used and generally not available.
Man-hours indexes are developed for all em­
ployees, production workers, and nonproduction
Sources . Industry employment and man-hours

indexes are developed from basic data compiled by
the Bureau of Labor Statistics or the Bureau of
the Census. For most private nonagricultural in­
dustries (including manufacturing), the Bureau of
Labor Statistics publishes employment and aver­
age weekly hours data for production or nonsupervisory workers and employment data for all em­
ployees. For manufacturing industries, the Bu­
reau of the Census publishes employment and ag­
gregate man-hours data for production workers
and employment data for all employees (including
nonproduction workers). The Bureau of the
Census provides data in greater industry detail
within manufacturing than BLS.
The two sources differ in their definition of
man-hours. The Census data include all hours at
the plant, worked or paid for, and exclude paid
time for vacations, holidays, or sick leave, when
the employee is not at the plant. Overtime and
other premium pay hours are included on the basis
of actual time at the plant. In contrast, the Bu­
reau of Labor Statistics data include time for paid
vacations, holidays, and sick leave, as well as plant
man-hours. Differences in the data from the two
sources for the same industry, however, also stem
from sampling and reporting differences.
Whenever employment and hours data are avail­
able from both the Bureau of the Census and the
Bureau of Labor Statistics, the labor input data
which are used are those consistent with the data
on output. Thus, when output data from the
Bureau of the Census are used, employment and
hours data from the same source usually are
While both the
Bureau of the Census and the B L S provide data
on production worker man-hours, neither source
provides annual data by industry on nonproduc­
tion worker nor all-employee man-hours. There­

Nonproduction Worker Hours.

fore, these measures are estimated.
The estimates of aggregate nonproduction
worker man-hours for the manufacturing indus­
tries are derived from published employment data,
and estimates of average annual hours worked or
paid per nonproduction worker.
The estimates of average annual hours worked
are calculated by multiplying the number of work­
weeks in the year times the scheduled weekly hours.
This produces an estimate of average annual hours
paid. Estimated hours for vacations, holidays,
disability, and personal time off are subtracted
from average annual hours paid, to obtain an esti­
mate for average annual hours worked.
Vacation and holiday trends are based on data
from various B L S surveys. Estimates of disa­
bility time are based on studies of the Department
of Health, Education, and Welfare, and data from
B L S surveys. Personal time off has been esti­
mated as a constant from references in relevant
All employee man-hours estimates for manufac­
turing industries are derived by summing the
aggregate man-hours for production workers, and
the estimated aggregate man-hours for nonproduc­
tion workers.

Comparability of Output and Man-Hours Data
Man-hours data are based on total man-hours of
establishments classified in an industry, whether
the man-hours are applied to production of pri­
mary or secondary products. Annual physical
output data, on the other hand, usually include
only primary products of an industry. In addi­
tion, they are usually reported on a “wherever
made” basis. Thus, there can be some discrepancy
in the coverage of output and man-hours measures.
This is not a serious problem unless there is con­
siderable variation from year to year in the pro­
portion of primary products to total products of
an industry, or if there is change in the proportion
of primary products which are made in other in­
dustries. The comparability of the man-hours
and output data is indicated by the specialization
and coverage ratios which the Bureau of the
Census publishes. All industries in the B L S
5 For consistency with output indexes, B L S annual employ­
ment and man-hours data for the mining industries are adjusted
to benchmark levels indicated by the periodic Census of Mineral


industry measurement program have high spe­
cialization and coverage ratios.
In selecting industries for the measurement pro­
gram, attention is also given to changes in the
degree of vertical integration. Man-hours relate
to all operations performed by establishments of
an industry, while output usually is measured in
terms of the final product. I f establishments
undertake additional operations, such as the manu­
facture of components which had previously been
purchased from suppliers, man-hours will increase
but there will be no corresponding increase in final
output. Thus, output per man-hour indexes would
be biased. In developing industry indexes, the
B L S examines data such as the ratio of cost of
materials to value of shipments for any indication
of a change in the degree of vertical integration.

B L S indexes are published annually in the re­
lease, Indexes of Output Per Man-Hour , Selected
Industries. Since data for the separate industries
become available at different times, interim reports
are issued to avoid delaying publication of avail­
able information until data for all industries are
available. As new industry indexes are developed,
special reports are issued which describe in some
detail the methods and data used in the construc­
tion of the indexes, as well as background informa­
tion about the industry.
Indexes of output per man-hour also are pub­
lished in the Statistical Abstract of the United
States , and some indexes for earlier years are pub­
lished in H istorical Statistics of the United States .

Uses and Limitations
Industry measures of output per man-hour are
particularly useful for studying changes in man­
power utilization, projecting future manpower re­
quirements, analyzing trends in labor costs, com­
paring productivity progress among countries,
examining the effects of technological improve­
ments on employment and unemployment, and
analyzing related economic and industrial activi­
ties. Such analysis usually requires that indexes


of output per man-hour be used in conjunction
with other industry data. For example, to study
technological effects, related data on production
and employment are useful; to study trends in
labor costs, data on earnings and other labor ex­
penditures are necessary.
Although the measures relate output to one in­
put—labor time—they do not measure the specific
contribution of labor, capital, or any other factor
of production. Rather, they reflect the joint effect
of a number of interrelated influences such as
changes in technology, capital investment per
worker, utilization of capacity, layout and flow
of material, skill and effort of the work force,
managerial skill, and labor-management relations.
Also, indexes which relate output to one group
of employees represent the total output of the
industry resulting from all employees and do not
represent the specific contribution of that group
of employees.
These industry measures of output per man-hour
are subject to certain qualifications. First, exist­
ing techniques cannot fully take into account
changes in the quality of goods and services pro­
duced. Second, although efforts have been made
to maintain consistency of coverage between the
output and labor input estimates, some statistical
differences may remain. Third, changes in the de­
gree of plant integration and specialization often
are not reflected adequately in the production sta­
tistics. This may result in overstatement of pro­
ductivity gains in some years, understatement in
others. Fourth, indexes involving nonproduction
worker man-hours are subject to a wider margin
of error than are the indexes using production
worker man-hours because of the technique for
estimating average man-hours of nonproduction
workers. Errors in estimating man-hours for non­
production workers, however, have a relatively in­
significant effect on the estimates of man-hours for
all employees. Fifth, year-to-year changes in out­
put per man-hour are irregular, and, therefore, not
necessarily indicative of basic changes in long-term
trends. Conversely, long-term trends are not neces­
sarily applicable to any one year or to any period
in the future. Because of these and other statistical
limitations, these indexes cannot be considered pre­
cise measures; instead they should be interpreted
as general indicators of movements of output per



Technical References


1. Dunlop, John T. and Diatchenko, V. L ab or Productivity (New York, San Francisco, Toronto,
London: McGraw-Hill Book Company, 1964).
An international collection of papers presented by participants of the conference on labor
productivity held at Cadenabbia, Lake Como, Italy, in 1961. Papers cover concepts and
measurements of productivity; international comparisons of productivity; wages and productivity;
and technical, managerial, and organizational factors affecting productivity.
2. European Productivity Agency, Organization for European Economic Co-operation, Productivity
M easurem ent , Volume I (Concepts) Prepared by G. Deurinck. Paris, August 1955, Project
No. 235.
The following listed essays are included: The Concept of Productivity and Its Corollaries;
Alternative Productivity Concepts; Aspects of Productivity Measurement and Meaning; The
Role of Official Statistics in Measuring Productivity; Productivity, Efficiency and Wages;
and Indices of Industrial Efficiency.
3. Evans, Duane and Siegel, Irving H. “The Meaning of Productivity Indexes,” J o u r n a l o f the
A m erican Sta tistic a l A sso ciation , March 1942, pp. 103-111.
The nature, use, and limitations of productivity indexes.
4. Fabricant, Solomon. “Of Productivity Statistics: An Admonition,” Review o f E con om ics a n d
S ta tistic s , November 1949, pp. 309-311.
The deficiencies and limitations in productivity measurement.
5. Greenberg, Leon. D a ta fo r M easurem ent o f In d u str ia l Productivity in the U nited S tate s , U.S.
Department of Labor, Bureau of Labor Statistics, 1965.
A review of the status of available statistical data and determination of what gaps exist in
order that improvements can be made.
6. International Labour Office. M ethods o f L ab ou r Productivity S ta tistic s , Geneva: Report prepared
for the 7th International Conference of Labor Statisticians, Geneva, September 1949. Studies
and Reports 1951, New Series No. 18.
Reviews the methods of measuring and comparing productivity, principally in the United
States. Includes concepts, definitions, and factors influencing productivity, and discusses the
problem of comparability of production and man-hour data.
7. Kendrick, John W. Productivity T rends in the United States (National Bureau of Economic Re­
search, Number 71, General Series) Princeton, N.J., Princeton University Press, 1961.
A presentation of historical measures of output, input, and productivity for the U.S. economy
and industry groups, including descriptions of concepts and methods of measurement. Also
includes discussion of implications of productivity change for economic growth, prices, incomes,
and resource allocation.
8. Mark, Jerome A. “Industry Indexes of Output Per Man-Hour,” M onthly L ab o r Review , November
1962, pp. 1269-1273.
A description of the methodology used in the construction of Bureau of Labor Statistics
indexes of output per man-hour. Covers methods and sources, construction of production and
man-hour indexes, and limitations.
9. National Bureau of Economic Research, Conference on Research in Income and Wealth. Output,
In p u t, a n d Productivity M easurem ent, Volume 25 of Stu d ies in Incom e an d Wealth, Princeton,
N.J., 1961.
A collection of papers and comments devoted to an appraisal of the measurement of output,
input, and productivity.
10. Siegel, Irving H. Concepts an d M easurem ent o f Production an d Productivity, Working Paper of
the National Conference on Productivity, 1952.
The rationale and techniques of measurement of changes in the physical volume of production
and the level of productivity. Includes an extensive bibliography on production and productivity
11. United Nations. In dex N um bers o f In d u str ia l Production, Series F, No. 1 of Studies in Methods,
New York: 1961.
Deals with the more technical aspects of compiling production indexes.
12. Productivity: A B ib liograph y (BLS Bulletin 1226, 1958; and BLS Bulletin 1514, 1966).
Two collections of annotated references concerning productivity and productivity measure­

— L loyd


P r o c h n o w , F r it z K a f k a , a n d J o s e p h


D ragonette

Office of Productivity, Technology and Growth

Chapter 24. Technological Change
Studies of technological changes and their labor
implications have been undertaken by the B L S
over the years for a variety of purposes. During
the 1930’s, public interest focused on the unem­
ployed and reports were prepared on technological
change and displacement of workers in various in­
dustries. During World War II, emerging tech­
nologies were studied for purposes of improving
manpower utilization.
Beginning in the mid-1950’s, nationwide atten­
tion was focused on the implications of new devel­
opments classified under the general term “automa­
tion.” The B L S made a series of studies on a
plant basis, in the insurance, petroleum refining,
bakery, air lines, and electronics industries, to ex­
plore the manpower implications of various
changes. Later, broader studies were undertaken,
including a survey of manpower impacts of
changeover to electronic computers in 20 large
companies 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 investiga­
tion of future changes. Research now underway
pinpoints technologies which will become increas­
ingly important over the next decade in key indus­
tries and attempts to provide advance information
about their manpower implications.

Description of Studies

logical change. The framework of such studies is
a single plant or office. A typical case study covers
such topics as: management’s objectives and re­
sults of introducing technological change at the
plant or office; extent of displacement and reas­
signment of employees; practices regarding trans­
ferring, retraining, and selecting employees for
new occupations; characteristics of employees
whose jobs were eliminated and who were assigned
to new positions; and implications of automation
for older employees. Both qualitative and quanti­
tative data are presented.
The case study approach has also been used to in­
vestigate in detail special aspects of adjustments to
technological change. Among the topics covered
by case studies are: the performance of older work­
ers in industry retraining programs, as shown
in four plants in different industries; manpower
planning procedures followed in connection with
a series of technological changes at a public utility;
the adjustments to electronic data processing in a
government agency; experience under an adjust­
ment program in tho railroad industry; post-lay­
off work experience of displaced workers in plants
in different industries and regions; and job rede­
sign for older workers at different plants. In
these studies, information about single plants is
presented, as illustrative, rather than as representa­
tive cases.

Reviews of Technological Trends in Major

Case Studies

To provide a broad overview of significant
trends in the economy, the Bureau prepares a sum­
mary report, applying to key industries on new
types of machinery, processes and products which
are believed likely to have an important effect over
the next 5 to 10 years. The industries covered com­
prise a cross section of the economy and include
those where the pace may be slow as well as those
where change is rapid. The first version of this
report, entitled Technological Trends in 36 M ajor
American Industries, was issued by the President’s
Advisory Committee on Labor-Management

B L S case studies provide detailed information
on various aspects of adjustments made to techno­

1 For discussion of various research methods used in studying
automation, see technical references 1 and 2.

The Bureau’s research program on technological
change involves a variety of reports and studies of
different degrees of detail and approach. The cur­
rent program thus provides detailed case studies
of changes within a single plant or office; summary
reports surveying trends in major industries; de­
tailed industry studies; and studies of major tech­
nological innovations, such as computers, that
affect workers in different industries.1





Policy in 1964. A revised edition covering 40 in­
dustries was published in 1966.2
The emphasis of the report is on technological
developments within each industry in an early
stage of the innovation’s commercial use—i.e., the
period after introduction on the market but before
widespread adoption. Inventions and discoveries
still in the “drawing board” stage are considered
unlikely to have as much impact over the next
decade as those already tested and are generally
not discussed.
The report briefly describes recent technological
developments, indicating insofar as practicable
some economic advantages of various types of new
equipment, processes or products; their importance
in terms of the man-hours engaged in the opera­
tions affected; estimated extent of use currently
and in 5 to 10 years, and some factors affecting
adoption such as the volume of investment and ex­
penditures for research and development. The
advantages described include not only labor sav­
ings per unit, but also quality improvements, fuel
and material economies, greater accuracy, new
markets, etc.
In assessing the employment implications of
technological changes, account is taken of the
possible rate of growth in output per man-hour
and in the industry’s total output. Appraisal also
is made of the changes in occupational structure
and of some issues and examples of adjustment that
are taking place.

Outlook Reports on Industry Technological
Intensive studies are made of selected major in­
dustries where far reaching changes, on a large
scale, are taking place, such as coal, electric power,
transportation, and textiles. These studies involve
detailed analysis of the economic implications of
major technological developments within individ­
ual industries. Factors analyzed include invest­
ment trends and factors affecting the prospects for
the diffusion of recent technological advances, such
as the structure of the industry. Estimates are
developed of the displacement of present by new
methods over the next 10 years. Unit labor re­
quirements under new and old technologies are
compared, wherever possible. Since the focus of
2See technical references 10 and 11.
3See technical references 8, 10, and 11.

the study is on the industry as a whole, data on
recent industry trends in output per man-hour,
production, and employment are examined in re­
lation to long-term trends and projections of future
trends are developed.

Technological Innovation Studies
Some technical innovations have applicability
in many industries. Among these are such devel­
opments as computers, numerical control of
machine tools, materials handling equipment, and
control instruments.3 Because of their far reach­
ing impact, special studies have been made of the
nature, status, prospects for adoption, and impli­
cations for unit labor requirements, occupational
change, training needs, and problems of industrial
relations. In analyzing their impact in different
industries, differences as well as similarities are

Data Sources and Collection Methods
A variety of data sources and collection methods
are utilized in making studies of technological
change and its impact.

Personal Interviews
In making case studies, analysts personally con­
duct intensive interviews with plant managers, per­
sonnel directors, and other officials who have direct
knowledge of changes at their plant. Union offi­
cials at the plant, and in some cases, individual
workers are interviewed. The analyst uses a
checklist of questions in conducting informal in­
terviews in order to elicit the maximum amount of
data. Plants and offices included in these studies
are selected on the basis of having recently made
a major change in their equipment, products, or
methods of production.
Personal interviews also are utilized to help de­
termine industry trends. Informal interviews are
conducted with engineers, scientists, economists,
and other experts in companies which produce and
use new technology, unions, trade associations,
government agencies, universities, etc., who have
specialized knowledge of particular technological
developments of industry trends. One objective


in these cases is to obtain their expert judgment
about the nature, pace of introduction, and pos­
sible impact of developments with which few
plants have had any experience. The emphasis
in these interviews is on the technological change
rather than on experiences in adjusting.

Trade and Technical Publications
Important sources of information concerning
technical trends are trade journals, technical maga­
zines and books, conference proceedings, govern­
ment hearings, and company reports. Annual re­
ports of leading corporations and company house
organs often contain useful information on current
technical development in some industries. In mak­
ing studies of industries, these publications are
reviewed to obtain information about the status
and prospects of important developments and to
ascertain which companies and plants merit
more intensive field visiting. Reports and publi­
cations of firms that produce particular types of
equipment often are found useful in studies of
industries that use such equipment.

Statistical Data Sources
Quantitative information about the status of spe­
cific technological developments is fragmentary
and scarce. The Bureau makes use of available
data from many public and private sources. These
sources include, for example: Bureau of the Bud­
get: annual inventory of computers in the Fed­
eral Government; Business and Defense Services
Administration: annual survey of numerically
controlled machine tools; McGraw-Hill Company:
survey of computers in industry; American
Bankers’ Association: survey of banking automa­
tion ; American Machinist M agazine: inventory of
metalworking machinery; and Bureau of Mines:
annual reports on mining equipment.
Statistical information on industrywide trends
are useful in analyzing the economic implications
of technological change. Among the important
sources used in preparing studies are the Bureau’s
indexes of output per man-hour and related series
on production, employment and hours; the Bureau
of Census data on expenditures on plant and equip-*
* See technical reference 5.
BSee technical reference 4.
222—617— 66----- 14


ment; and the National Science Foundation’s
estimates of research and development.

Mail Surveys
Because of the complexity of the subject, rela­
tively little use has been made of mail surveys in
studying the impact of technological change. This
technique, however, is useful in obtaining informa­
tion on a broad scale not otherwise available, to
supplement detailed information collected through
case studies. Thus, to determine the extent and
manpower impact of computer use in the insurance
industry, the Bureau conducted a two-part survey
of 400 companies employing most of the employees
in the industry.4 In the first stage, a brief, 1-page
questionnaire covering data on employment and
extent of computer use was sent to a group con­
sisting of all large offices, and a sample of small
companies. From information received from this
screening questionnaire a number of companies
were selected to whom a more detailed question­
naire was sent, asking for information about com­
puter uses, number of employees engaged in elec­
tronic data processing, planned applications, etc.
The mail survey technique also was used in fol­
lowup surveys of workers who have been laid off
as a result of technological and economic change.5
The names and addresses of such workers were
obtained from plant personnel records, selecting
only a sample in cases where a large number of
workers had been displaced. The questionnaire
was mailed at least 6 months after the layoff oc­
curred, in order to allow a period of time for some
adjustment to take place. Information collected
from the mail survey dealt with the personal char­
acteristics of displaced workers such as age, sex,
occupation, level of education and skill, post-layoff
work experience, such as labor force status at time
of survey, type of job held after layoff, source of
jobs, etc.

Plant Records
In making detailed studies of the impact of tech­
nological change on individual workers within a
plant, analysts sometimes can obtain from em­
ployers’ files, data on such aspects as the age, sex,
and related personal characteristics of employees
whose jobs are eliminated and the jobs in the plant
held by each individual affected before and after



the change; similar data are collected on individ­
uals who are selected for the positions created in
connection with automated equipment.6

Expert Review
In preparing forecasts of future technological
trends, a critical step is the review of preliminary
reports with outstanding experts in each industry.7
Drafts of industry reports are mailed to company
executives, union research directors, trade associa­
tion officials, technical journal editors, and univer­
sity and government specialists for their assess­
ment of the validity and adequacy of projected
trends. Over 450 persons were contacted in this
way in the preparation of a report on technological
trends in major industries. Some experts are vis­
ited personally to review draft statements in
detail. Through this means, reports on techno­
logical prospects are designed to reflect, as much
as possible, the authoritative views of a number of
persons who have expert, first hand knowledge of
each industry.

Analysis and Interpretation
For a better understanding of research results
in this field, it is important to keep in mind the
meaning of certain key ideas and concepts. Some
of the key problems of interpretation and analysis
in this type of research are therefore set forth,

Definition of Technological Change
Technological change is defined broadly in the
B L S studies, as encompassing significant changes
in processes and equipment, products and services
produced, and materials, fuels, and energy used.
The term “automation,” which is sometimes popu­
larly used as a synonym for “technological
change,” designates, strictly speaking, a particular
type of current development. It has been vari­
ously defined, for example, as “automatic opera­
tion,” “the mechanization of sensory, control and
thought processes,” and “a concern with produc­
tion processes as a system.”
6See technical references 1 and 6 for use of plant records.
7 See technical references 10 and 11 for further detail.
8See technical reference 8 for further discussion.

While B L S studies have been concerned with
developments in “automation,” particularly in an­
ticipating long-term trends in the future, they are
not the only technological changes taking place
that affect labor requirements and industrial rela­
tions. For example, new ways of generating
power, piggybacking in transportation, use of
synthetic materials in manufacturing, mechanized
methods of materials handling, and faster steel­
making processes are important technological
developments, not usually covered by technical
definitions of “automation,” but having significant
manpower implications.

Impact on Productivity
Since one of the principal consequences of
technological change, so far as manpower utiliza­
tion is concerned, is an increase in productivity—
i.e., output per man-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
In determining the impact of a specific tech­
nology, B L S studies try to indicate the reduction
in unit labor requirements that the new processes
are designed to achieve. In some cases, estimates
of labor savings are derived on the basis of com­
parisons 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
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 operation,
but would have little impact on finishing, and
assembling, and may even require additional labor
in engineering and maintenance work.8 The im­
pact on plant productivity, therefore, would be
considerably less than the effect on productivity
of any department or operation directly affected.


Impact on Employment
In assessing the impact of technological change
on employment, it is necessary to consider the im­
plications of plant manpower policies and the
effects of economic changes, with which technical
changes interact. Analysis of the impact of tech­
nological change purely in terms of machinery is
At the plant level, for example, the substitution
of machinery for labor may substantially reduce
job opportunities in operations directly affected.
I f 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 from the plant.
Moreover, the employment impact of technical
change is also interrelated with the effects of the
business cycle. Thus, workers whose jobs are
eliminated by technical changes may not be dis­
placed from a plant until a decline in demand
results in layoffs—a long time after the change has
been made in some cases. In the subsequent re­
covery, however, they may not be hired back be­
cause their jobs no longer exist.
Since many changes exert their effects on em­
ployment through the competitive market, the em­
ployment 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,
merger, and consumer taste, it is very difficult to
isolate the expanding and displacing effects of
technological change.

Impact on Occupations
Two aspects of occupational change resulting
from technological change are examined. Changes
in job structure—the distribution of the plant or
office work force by function or broad skill group­
ing—are studied to determine the extent of up­


grading 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 perform these duties as indicated by job descrip­
tions and observation. The content of newly
created jobs, such as programmer, also is studied
and the qualifications required and personal
characteristics of individuals selected for these
new positions are described, so far as possible.9

Adjustments to Technological Change
Technological change has important implica­
tions for personnel management and collective bar­
gaining 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 promotion, the setting of
wage rates, and selection of persons for new jobs.
Often the adjustment proceeds according to rules
established in advance through collective bargain­
ing. 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.10 The limitations
of these measures as well as their advantages are
important matters studied.

Uses and Limitations
B L S studies of technological change are pre­
pared as part of the U.S. Department of Labor’s
program for carrying out the objectives and re­
sponsibilities of the Manpower Development and
Training Act. Under this act, the Secretary of
Labor is required “to establish techniques and
methods for detecting in advance the potential
manpower impact of automation, technological
progress, and other changes in the structure of
production.” As part of such an early warning
system, B L S studies and reports of technological
change are useful to managers, union leaders, edu-*
* See technical references 1, 6, and 8.
30 See technical references 1 and 6 for further discussion.



cators, economists, government officials, and others
in planning policies to cushion the impact of
change. The study of emerging technological
trends and possible implications, moreover, pro­
vides a basis for more valid projections of pro­
ductivity and economic growth. They also are use­
ful in pinpointing manpower problems and
determining the most productive direction of
future research to obtain possible solutions.
Some limitations of the Bureau’s studies of
technological change must be kept in mind in as­
sessing their appropriateness for particular uses.
In general, it is important to recognize that judg­
ments about the future direction and pace
of technological change and its implications are
necessarily complex and difficult. The rate of in­
troduction of new technology depends not only on
technical advantages but also on many economic
factors, such as the volume of investment, market

prospects, and the availability of trained workers,
all of which are subject to significant variations.
Moreover, since the period of introduction gen­
erally spans a number of years, new developments
are constantly appearing so that assessments of
the outlook must be reappraised from time to
time in the light of new information.
Finally, studies of the impact of technological
change deal primarily with changes within in­
dividual industries. But these changes often
involve changes in the type and amount of goods
and services purchased from other industries and
could therefore have important implications for
production and employment in industries supply­
ing inputs. The accumulation of information on
interindustry relationships, through the Bureau’s
economic growth studies, will provide a quantita­
tive basis for analyzing this aspect of technological

Technical References





(BLS Bulletin 1276, 1960).
A study of manpower implications of the installation of electronic data processing in 20 offices
in private industry, with special reference to older workers. Discusses problems of measuring
impact of electronic data processing on employment and occupational requirements.
A u tom ation : A D iscu ssio n o f Research M ethods, Labour and Automation Bulletin No. 1, Geneva,
International Labour Office, 1964.
Fourteen papers on problems of research methodology in studying manpower impact of auto­
mation at the plant and industry level. Includes paper by Leon Greenberg and Edgar Weinberg,
“Nationwide Studies in the United States,” describing Bureau of Labor Statistics research
program on technological change.
A utom ation an d Technological Change (edited by John T. Dunlop). The American Assembly,
Columbia University, Englewood, N.J., Prentice-Hall, 1962.
Case studies of psychological impact are summarized in Chapter 3, by Floyd C. Mann. Chap­
ter 7, “Employment,” by Ewan Clague and Leon Greenberg, discusses problems of measuring
employment impact.
C ase Stu d ie s o f D isplaced W orkers (BLS Bulletin 1408, 1964).
Studies of post-layoff experiences of nearly 3,000 workers formerly employed at five different
plants. Discusses use of mail questionnaire in followup of displaced workers.
Im p ac t o f Office A utom ation in the In su ran ce In du stry (BLS Bulletin 1468, 1966).
Mail survey of over 400 insurance companies. Covers extent, pace, and employment
implications of electronic data processing.
Im pact o f Oifice A utom ation in the In tern al Revenue Service (BLS Bulletin 1364, 1963).
Detailed case study of large-scale changeover to electronic data processing. Illustrates
use of internal personnel records in analyzing effects of office automation.
In d u stria l R etrain in g P ro gram s fo r Technological Change (BLS Bulletin 1368, 1963).
Four case studies of experience of older and younger workers in industrial retraining pro­
grams, based on plant records. Discusses some problems of measuring comparative performance.
Outlook fo r N u m erical Control o f M achine Tools (BLS Bulletin 1437, 1965).
A study of manpower implications of a key technical innovation affecting metalworking
industries. Discusses problem of generalizing about change in productivity as result of specific
A djustm ent to the Introduction o f Office A utom ation


Technical References—Continued

9. Edgar Weinberg, S o c ia l Im p licatio n s o f Office A utom ation. Paper presented to the Fifteenth
Annual Conference of the Association for Computing Machinery, August 1960.
Analyzes techniques and findings of case studies and some implications for personnel manage­
10. Technological T ren ds in 86 M a jo r A m erican In d u strie s . A BLS study prepared for the President’s
Advisory Committee on Labor-Management Policy, 1964.
11. Technological T rends in M a jo r A m erican In d u strie s (BLS Bulletin 1474, 1966).
Description of outlook for major technological developments based on a variety of data
sources and expert review.

— E dgar W einberg
Office of Productivity, Technology and Growth

Chapter 25. Construction Labor Requirements
New construction is a major component of the
Nation’s output of goods and services and a major
source of employment. The jobs it creates occur
not only at the site of employment, but also in the
many manufacturing, trade, transportation, and
service industries which furnish the materials and
services required in the construction process.
Because of this far-reaching employment impact,
the creation of new construction projects often is
regarded as a means of counteracting cyclical
To assist in assessing the extent of the impact
of construction expenditures on employment, a
series of labor and material requirements studies
for different types of construction was started in
1959. The program was established as a result of
Congressional legislation, requiring the Bureau
of Labor Statistics to provide estimates of the
amount of total employment generated directly
and indirectly by various kinds of construction
per dollar of expenditure.
Earlier Bureau efforts to relate employment
and volume of construction included the “Labor
Required for New Construction” series, with
reference mainly to on-site man-hours. This series,
started in the early thirties, appeared intermit­
tently through the years, but was not based on
actual up-to-date surveys and was finally halted
in the mid-fifties because the factors employed
were found to be obsolete. There was also a series
of Public Works Administration studies pub­
lished in 1940, covering Federal public works
projects constructed in the mid-thirties; and
a few individual studies of specific types of
The present studies include the major types of
building construction (schools, hospitals, public
and private housing, etc.) and also heavy con­
struction (highways, dams, etc.). However, only
one or two selected types of construction are sur­
veyed in any given year. Selected types of con­
struction are resurveyed periodically. These re­
surveys may, in addition to providing current
information on labor requirements, contribute in­
formation useful in preparing construction cost


indexes and estimates of changes in productivity
of on-site construction labor.

Description of Survey
The surveys are designed primarily to deter­
mine the number of man-hours represented by a
fixed dollar volume ($1,000) of contract construc­
tion. Man-hours, as defined by the surveys, in­
clude both on-site construction employment and
the off-site employment required to produce and
deliver materials used in the construction. Data
for on-site labor include total man-hours for the
supervisory, engineering, clerical, and custodial
employees, as well as those for workers in each
construction trade at the site of construction.
Data for off-site labor include employment in the
off-site (e.g., office and warehousing) activities of
construction contractors; in building materials
and equipment manufacture and distribution; and
finally, employment in all the other industries
which are affected directly or indirectly by the
production and distribution of building materials
from raw materials to the final manufactured
product. Man-hours are also estimated for the
employment created by overhead expenditures of
Certain types of employment are not covered
by the survey. For example, no estimate is made
of the employment used in the planning and de­
signing of the projects studied. Also excluded
are estimates of employment required in govern­
ment and public utility agencies which might be
affected by the construction being studied. Em­
ployment created by the respending of wages and
profits of the workers and their employers—the
multiplier effect—are not considered within the
scope of the studies.

Data Sources and Collection Methods
The surveys include construction which is nonfederally assisted and/or that which may be
totally or partially financed through Federal
funds. Although the type of construction labor


and material requirements data sought are similar
for both federally and nonfederally aided proj­
ects, the sources for the data are different, particu­
larly for the on-site man-hour information.
For the construction of nonfederally aided proj­
ects—those financed entirely by various levels of
local government, private individuals or voluntary
groups—on-site man-hour data, by occupation, are
obtained by field representatives from local au­
thorities, architects, contractors, and other direct
participants in the projects. These individuals
provide the desired man-hour information from
payroll records, daily work force reports, or sum­
mary time reports.
For those projects financed wholly or in part
by Federal funds, on-site employment information
generally is obtained from contractors’ payrolls
submitted to the government under regulations of
the Davis-Bacon Act or other Federal legislation
covering federally aided construction.
These payrolls furnish the data for estimating
on-site man-hour requirements, as well as data on
wages for all hourly rated workers on the projects.
Data for on-site salaried employees, not accounted
for on the payrolls, are obtained by the field agents
from the contractors.
Access to these payrolls also makes possible the
collection and presentation of additional detail for
the projects studied. This includes information
on wage relationships, timing of construction op­
erations, and requirements by type of contractor.
Information on material and equipment costs
for the projects studied—i.e., the items which are
used for estimating off-site employment—is col­
lected by the field representatives from the indi­
vidual contractors and subcontractors engaged in
the construction of the sample projects.

Sampling and Estimating Procedures
Sampling procedures vary with the type of con­
struction being studied. The “universe” of proj­
ects for a specified study generally represents all
of the projects known to have been completed dur­
ing a selected 1-year period.
The projects are then stratified into cells having
similar characteristics which may affect man-hour
requirements. Factors considered important in
cell stratification include (a) regional location,
(b) metropolitan or nonmetropolitan locality, (c)


types of structure (when known), (d) purpose of
structure (if variable)1 and (e) amount of total
construction contract (i.e., size of project).
One or more projects are selected at random
from each cell and assigned weights to give total
representation of the cell. Substitutions, when
necessary, are made from the same cells.
Data from the sample projects are combined in
aggregate form to represent the total dollar vol­
ume of each cell.
As indicated previously, the off-site employment
estimates are derived from the materials and equip­
ment cost information obtained from the contrac­
tors and subcontractors cooperating in the studies.
The contractors provide a list of the value of each
type of material used in the construction of sample
projects. These material listings are classified into
categories consistent with the 4-digit Standard
Industrial Classification code as used in Census of
Manufactures product groups. For each of these
product groups, average amounts of material (in
dollars) required for each $1,000 of contract con­
struction cost are determined. The value of ma­
terials is reduced by a ratio representing the differ­
ence between valuation by the purchaser and valua­
tion by the producer. (This ratio is based on
valuation data provided by the U.S. Department
of Commerce, Office of Business Economics.)
This latter step is required because all data re­
ported by contractors are in purchasers’ value, and
reduction to producers’ value is necessary to obtain
figures consistent with Census data which are used
to calculate manufacturing employment.
Primary employment in manufacturing is con­
sidered to be that required to produce the construc­
tion bill of materials in their final stage of fabrica­
tion. In this stage, man-hours are developed by
multiplying average producers’ value of each con­
struction material by the ratio of manufacturing
man-hours per $1,000 of production. (This ratio is
established by using the Census of M anufactures .)
Primary man-hours in the trade, transportation,
and service industries are estimated from the dif­
ference between producers’ and purchasers’ value
for each construction material. The value differ­
ences are alloted to trade and transportation, and
primary man-hours for each component of trans­
portation and trade are then estimated from labor
factors provided by BLS.
1 For example, elementary or secondary schools.



Secondary employment is defined as the em­
ployment in all industries involved in production
and transportation of building materials and
equipment from basic extraction to, but not in­
cluding, the final manufacturing stage. The 1958
Interindustry Study is used to obtain these esti­
mates.2 For each group of materials, the inter­
industry study provides inf ormation on the amount
of secondary products required from each of its
78 industry sectors. The product data are con­
verted to man-hours by use of output per man-hour
ratios for each of the sectors. Adjustments for
price and productivity are made to provide esti­
mates for years subsequent to 1958, consistent with
the year of construction and bill of materials.
For each off-site stage (primary and secondary),
a man-hour figure per $1,000 for the construction
being studied is obtained. When these man-hours,
plus the builders5 off-site employment, are com­
bined with the direct or on-site man-hours, the
total employment effect, within the definition used
by the studies, is determined.
An exact study of the off-site employment of
each construction contractor is not attempted,
since it is almost impossible to relate accurately
such employment to the projects studied. Build­
ers5 off-site employment is occupied not only with
the sample projects studied, but also with other
current or future projects. The estimate of con­
tractors5off-site man-hours for each $1,000 of con­
struction contract is based on the difference be­
tween construction worker employment and total
employment in the construction industry, with
adjustments for on-site supervisory and adminis­
trative employment.

data, the statistical tabulations are supplemented
with an analysis of the various factors which ap­
parently affected the man-hour requirements for
the specific types of construction studied. The
bulletins contain information on the characteris­
tics of the sample projects and the man-hours per
square foot and per $1,000 of contract cost, with
an analysis of the variations in requirements and
costs arising from differences in design, type of
structure, and regional or metropolitan versus
nonmetropolitan influences.
The reports include information and analysis
o f: on-site man-hour requirements by occupation;
the employment share of the general and special
trades contractors engaged in the construction
work; direct on-site wage cost; the distribution
of employment by periods of construction time;
and the cost of major materials.
When feasible, man-hour and material compari­
sons with earlier periods and other types of
construction are included in the bulletins.

Uses and Limitations

The construction labor and materials require­
ments studies are published in Bureau bulletins.
Summary articles, based on the survey find­
ings, appear in the Monthly Labor Review. The
bulletins and articles highlight the total man­
hours generated per dollar volume of construction
expenditures, with subtotals for on-site and off-site
man-hours. In presenting the labor and materials

The results of the labor requirements surveys
are used by other offices of the Bureau, other Bu­
reaus of the Department of Labor, other govern­
mental agencies, congressional committees, and
industry research and trade organizations, to
assist in gauging the impact of planned expendi­
tures for construction on employment, and the
economy, generally. Of special interest to mar­
ket research analysts and companies manufactur­
ing materials for use in construction, are the
materials listings per $1,000 of construction
While the overall estimates of employment are
believed to be reasonably accurate, the detailed
data would have a wider margin of sampling error
and may be subject to other limitations. Man­
hour and material requirements are affected by
a number of factors such as location, size of proj­
ect, type of structure, architectural design, avail­
ability of certain materials or equipment, labor
skills, and local building codes and customs. The
effects of these separate factors cannot be isolated.

a See M. R. Goldman, M. L. Marimont, and B. N. Vaccara,
“The Interindustry Structure of the United States.” Survey of
Current Business, November 1964, pp. 10-29.

—J ames F. W alker
Office of Productivity, Technology and Growth

Analysis and Presentation

Industrial Safety
Chapter 26. Frequency and Severity of Work Injuries
To the individual, work injuries can mean loss
of work time, suffering, bodily limitation, or even
death. To his family, they can mean income loss
and anguish. To industry, they mean loss of time
and product, disruption of work and organization,
compensation expense, and the effort and expense
of hiring and training replacements. All these
results of work injuries illustrate the need for the
greatest possible safety precautions in the work
place, the goal of the pervasive safety movement
of the times. Work-injury statistics serve the
safety movement by indicating the extent and
nature of the injury problem, arousing interest in
it, and providing direction for overcoming it.
Work injuries have always been associated with
work, but they have been and are rare in the ex­
perience of individuals. So it was the collecting
of many workers in one place as well as the greater
hazards from power tools that made the factory
the first place where these injuries became a matter
of general public concern. While Massachusetts
began factory inspection as early as 1867, interest
was limited until nearly the turn of the century
and then centered on compensation for injury
rather than prevention of accidents.
In 1893, the Bureau issued its first report relat­
ing to work injuries, a study of European work­
men’s compensation procedures. A series of spe­
cial reports on workmen’s compensation problems
followed and to these were added studies of the
hazards associated with particular industrial
By 1910, interest both in compensation and in
the prevention of accidents and injuries had in­
creased greatly. In the next 10 years, all but eight
of the States were to adopt workmen’s compensa­
tion legislation. Industrywide safety programs
were organized in iron and steel production
and railroad transportation. In 1910 the Bu­
reau inaugurated a continuing series of annual in­
jury rate compilations for the iron and steel
industry. Other industries have been added as it

became possible to compile statistics for them. By
1926,30 industries were covered; by 1952, over 200.
At present, over 650 industries and industry groups
in manufacturing and nonmanufacturing are in­
cluded in the Bureau’s publication. For most of
these, both injury-frequency and injury-severity
rates are given, together with estimates of average
days of disability per injury and the percent of
disabling injuries resulting in death, permanent
impairment, and temporary-total disability.
These statistics have been used mainly by the
occupational safety movement, including plant
personnel responsible for safe workplaces and the
reduction of injury loss, their associations, suppiers, teachers, and advisors. Annual estimates of
the total number of work injuries and deaths and
of days of disability, made in connection with the
compilation of injury rates, have indicated the
general magnitude of loss and have contributed
to continued interest in and support of safety ef­
forts. These are described under “Uses” in this
chapter. The rates available on an increasingly
detailed industry basis, have provided those di­
rectly responsible for plant safety with an indica­
tion of the areas needing increased attention.
Also, the rates enable those people directly con­
cerned to measure their success in comparison
with others in the same industry having similar
problems and hazards. Attention to these measurments was intensified as the result of a war-time
Bureau statistics available when the United
States entered World War II indicated that in the
immediately preceding defense period work in­
juries had significantly increased in several indus­
tries important to the war effort. Conservation of
available manpower as well as the continuing rea­
sons mentioned above demanded rapidly available
indicators of the number of injuries in those indus­
tries. In 1943, the Bureau initiated publication of
a monthly series of rates for 53 selected manufac­
turing industries. Following the war, this series
was continued with data available for all months,
but with quarterly collection and publication. At



present, the quarterly reports cover 140 manufac­
turing industries and industry groups.
Data from which the Bureau compiles its rates
are provided in voluntary reports from establish­
ments in the industries covered. To reduce the
burden of reporting establishments, agreements
have been made with 13 States, which also com­
pile similar statistics, to collect the information
jointly and obtain from a single report the data
needed for both State and national statistics.

Description of Survey
This chapter is concerned with the annual se­
ries of statistics for manufacturing and non­
manufacturing industries and with the monthly
series (collected and published quarterly) for
manufacturing only. The annual series include
frequency rates, severity rates, average days of
disability per case, and percent of disabling in­
juries resulting in death, permanent impairment,
and temporary-total disability, while the quarterly
manufacturing release is limited to frequency
The history of these surveys has been initimately connected with efforts to achieve standards
and measures which would permit meaningful
comparisons of injury experience among establish­
ments and industries. The statistics are impor­
tant in providing a standard of comparison of the
experience of individual establishments and the
standards and rates are used as plant or company
indicators. These uses, as distinct from the eco­
nomic and social evaluations of most large-scale
statistical surveys, have resulted in a great deal
of contact and cooperation with experts from out­
side the discipline of statistics.

Efforts to standardize the method of compiling
work-injury statistics were initiated by the Bu­
reau of Labor Statistics in 1911. In 1914, the
Bureau called a formal conference of labor and
workmen’s compensation officials and others inter­
ested in this subject. The work of the conference
was continued in later years by the International
1 Standardization
tin 276, 1920).

of Industrial Accident Statistics

(BLS Bulle­

Association of Industrial Accident Boards and
Commissions (IA IA B C ), culminating in the pub­
lication of the first standardized procedures in
1920.1 In 1926, a sectional committee of the
American Engineering Standards Committee,,
later the American Standards Association, under­
took a revision of these procedures. This work
led to the publication in 1937 of the first American
Standard Method of Compiling Industrial In ­
jury Rates (Z16.1 ). The 1954 revision of this

standard was reaffirmed in 1959 and is reviewed
continuously by a sectional committee of the
American Standards Association. A second
standard, The American Recommended Practice
for Compiling Industrial Accident Causes (Z 1 6 .2 )
developed under the American Standards Asso­
ciation procedures, was published in 1941 and re­
vised in 1962. These two standards constitute the
basis for the concepts utilized in the compilation
of all injury and accident statistics by the Bureau
of Labor Statistics.

Measures of Injury Frequency and Severity
Injury measures included in the two American
Standards are used in these annual and quarterly
injury statistics programs and in other injury sta­
tistics compiled by the Bureau, as the following
Injury-Frequency R ates . Injury-frequency rates

are the primary measures of the incidence of
work injuries. They indicate the relative level of
injury occurrence prevailing in different establish­
ments, operations, or industries during a specified
period of time, and provide a means of determin­
ing trends in injury occurrence or of progress in
accident prevention.
The standard injury-frequency rate is defined as
the average number of disabling work injuries for
each million employee-hours worked. The lack
of comparability inherent in simple injury totals,
arising from variations in employment and oper­
ating time, is thus overcome by expressing the
injuries in terms of a standard unit of exposure.
A disabling injury is defined as any injury in­
curred in the course of and arising out of employ­
ment, which (1) results in death or in any degree
of permanent physical impairment, or (2) renders
the injured person unable to perform any regu-


larly established job, which is open and available
to him, during the entire time interval correspond­
ing to the hours of his regular shift on any one or
more days after the day of injury (including Sun­
days, holidays, and days on which the plant is
shut down). Under this definition, the report­
ability of an injury for injury-statistics purposes
is in no way related to the eligibility of the in­
jured person for workmen’s compensation pay­
ments. In case of doubt as to whether or not an
injured person is able to work, the attending phy­
sician’s decision is accepted for statistical purposes.
Injury-Severity Measures. The severity of a tem­

porary injury is measured by the number of days
during which the injured person is unable to work.
For death and permanent impairment, the Amer­
ican Standard provides a table of economic time
charges. These time charges, based upon an aver­
age working-life expectancy of 20 years for the
entire working population represent the part of
working ability lost as the result of specified im­
pairments, expressed in terms of unproductive
days. For example, death, representing the com­
plete loss of all future production by the injured
person, is assigned a time charge of 6,000 man-days
(i.e., 20 years of 300 days each). The complete
loss or loss of use of an arm is estimated as result­
ing in an average reduction of 75 percent in work­
ing efficiency. By applying this percentage to the
20-year working life expectancy, the time charge
for this type of injury is established as 4,500 mandays.
The standard injury-severity rate commonly is
used as a comparison measure indicating the rela­
tive level of economic loss resulting from work
injuries. It weights each disabling injury with its
established time charge or days of disability, and
expresses the aggregate in terms of the average
number of days charged for each million employeehours worked.
The average severity is computed by adding the
actual days lost for all temporary disabilities and
the time charges for all deaths and permanent
impairments and dividing the total by the number
of disabling injuries. This measure constitutes
the basis for direct evaluation of the severity of
injuries in different industries, establishments, or
a See appendix B.



, ____ , ____

The formulas for these injury measures are:

Number of disabling injuries X
1 000,000
requency rate—rpotai number of employee-hours ’
Total of days lost or charged X
1 ,000,000_______
Severity rate= ____________
"Total number of employee-hours , and
number of days lost or charged
Average severity—TotalNumber
of disabling injuries

Industries included in the annual survey cover
nearly the whole range of manufacturing and nonmanufacturing activity. Domestic service, agri­
culture, and rail and air transportation are com­
pletely excluded and insufficient sample coverage,
particularly in nonmanufacturing, makes it im­
possible to present data for some industries. Cur­
rently, rates are presented for 490 manufacturing
and 190 nonmanufacturing categories. However,
not all these rates are obtained from the B L S sur­
vey. Rates for mining and a few closely related
processing industries are obtained from a survey
of the Bureau of Mines and those for the Federal
government from reports to the Bureau of Em­
ployees’ Compensation. The monthly data com­
piled from the quarterly reporting program are
limited to manufacturing and cover 140 categories.
Reports are classified by industry on the basis of
the Standard Industrial Classification ( S I C ).2
Beginning with 1958 figures for the annual survey
and 1963 figures for the quarterly survey, defini­
tions used are from the 1957 edition of the SIC
Manual. Classification for mining and Federal
government, however, are those of the collecting
agencies and are not strictly comparable to the
SIC Manual definitions.
Rates compiled from B L S reports cover all
classes of employees (production, operating, and
related workers; construction workers; sales, serv­
ice, delivery, technical, professional, office, admin­
istrative, and clerical workers; and all other per­
sonnel) . Self-employed persons are not included.
Rates from Bureau of Mines reports cover workers
engaged in production, development, maintenance,
and repair work, and supervisory and technical
personnel at the operations, but exclude office per­
sonnel and employees in stores or affiliated opera­



tions not directly connected with mining or
refining operations. Working proprietors are in­
cluded for the Bureau of Mines rates. The Federal
government data, based on injuries reported to the
Bureau of Employees’ Compensation, cover all
Federal civilian employees. Areas covered by
the rates include the 50 States and the District of

Data Sources
Publication of statistics by industry requires
that the individuals whose experience is summar­
ized be in the same industry. The establishment,
rather than the larger company or corporation or
the smaller department or division, is the best re­
porting unit for work injury surveys. Nearly all
the reports included in the compilations of B L S
work injury statistics are for establishments.
Three types of information are needed to con­
struct these work injury statistics: number of hours
of exposure to hazards of work injury; number
and characteristics of injuries and disabilities;
and data necessary for classification by industry.
Two basic forms, both titled Work Injuries, are
used to collect this information, form B L S 1418
in the annual survey and form B L S 1417 in the
quarterly survey. Each of these forms provides,
first, for the entry of the total number of employee
hours worked in the period covered and the total
number of employees. The latter figure helps to
avoid some errors due to misunderstanding, and
also permits classifying the establishment by size.
As a further check there is an inquiry about
changes in the level of operation. Second, each
form asks for a count of injuries in each of several
classes of severity (deaths, permanent impair­
ments, and temporary-total disabilities). Because
of difficulty in obtaining information to classify
hernias as required by the American Standard
Method of Recording and Measuring Work Injury
Experience , the number of hernias is requested

separately. Eliminated from the count are firstaid and medical treatment cases which are not
disabling under the Standard. A separate line
is provided for entering them. To provide for
calculation of the severity rate, which is not in­
cluded in the quarterly reports, the annual report
form requests detailed information on the nature
and extent of injuries. (See the copy of form

B L S 1418 on pp. 201 and 202.) Third, each of the
forms provides for entry of information on the
industrial activity of the establishment, its opera­
tions, and its products or services.
Collection of the reports is by mail. The person
preparing the report for the larger establishments
is ordinarily the safety officer. For smaller estab­
lishments, which may not have a full-time safety
officer, reports usually are prepared in the person­
nel or accounting departments. Information such
as is needed for these reports also is required for
other purposes and very few establishments are
unable to report because the information is not
In a number of cases where a State agency,
another Federal agency, or an association com­
piles similar statistics, arrangements have been
made to have one report meet the needs of two or
more statistical programs. These arrangements
are discussed elsewhere in this chapter.

Collection Methods
Annual Survey
This survey is conducted by mail. Report
forms are mailed to a sample of establishments in
each of the industries covered. Response is en­
tirely voluntary. Refusals to report are rare;
most of the establishments asked to furnish the
information are willing to do so because of the use­
fulness of the resulting statistics to the safety
movement and as a standard of comparison for
their own experience. The Bureau pledges con­
fidentiality for the individual reports. The infor­
mation is limited to statistical use and individual
reports are never released except within the limits
of the cooperative arrangements where one report
is used in compiling statistics in two or more agen­
cies. Special care is taken to base published rates
on enough reports to prevent the disclosure of
either establishment or company information,
even by deduction on the part of a reporting
Report forms are mailed to each establishment
at the end of the year of reference. Second re­
quests are mailed about February 15 to all estab­
lishments which have not reported by that date.
About 90 percent of the reports are received by



B .L .S . 1418


B u d g e t B u r e a u N o . 4 4-R 002.15.
A p p r o v a l e x p ir e s N o v . 30, 1968.


W A S H IN G T O N , D .C . 202 1 2

Return THIS copy
to above address
Please change mailing address if
incorrect— Include Postal ZIP Code

Please complete this report whether
or not there were any disabling inju­
ries. See separate instruction sheet.

I* EX PO SUR E D A T A , 1965. (See instructions I.)
(Please complete this section even though there were no disabling injuries.)
1. Average number of employees
(Enter average of year; include
all classes of employees) .............................................


III. IN J U R Y SUM M ARY, 1965. (See instructions III.)
Report all disabling (or “lost-time”) injuries arising out of
employment; include occupational diseases.
DO NOT count any case in more than one section.
If no disabling injuries during year, enter “0” on line 11.

by all em­

Total hours worked
ployees during entire year . . . _______________________

Injuries resulting in—

A verage


(A ll deaths resu ltin g from 1 9 6 5 w ork In­
ju rie s. regard less o f tim e between in­
ju ry a n d d e a t h . )
(D e s c r ib e o v e r } . . .

of weeks

8. Permanent-total impairments

Peak operations .....................................................................
Normal operations.................................................... Slack operations ................................. -.......................
Total number of weeks operated during year...............
Comment -.....-......... ........................................... ........


(D escribe over) . . . . . . . . . . . . . . .

9. Permanent-partial impair­

(A lso include unrepaired in guin al her­
n ias.)
(D escribe over) . . . . . . . . .

(A ll w ork in juries, not reported above,
which disabled worker fo r 1 full calen­
d ar day or more a fte r the day of injury.)

Principal type of activity of this establishment (i.e., manu­
facturing, wholesale, retail, construction, public utility,


Disabilities of 1 , 2, or 3 days
e a c h .............................................. . .
( b ) Disabilities of 4 or more days
each (except inguinal her­
nias) .....................................

(Total days
lost, including
S aturday s and
Sun days)


5. Enter in order of importance the principal
p e r c e n t o f to ta l
products manufactured, lines of trade, a n n u a l s a le s v a lu e
specific services, or other activities.
or rece,pts*1965
( a ) ................................................................................
( b ) _..............................................................................................................................%

(c) Hernias, inguinal, repaired . .
Disabilities of unknown du­
ration (Describe o v e r ) ...............

(d )

(c) ............................................................... %
(d ) ..........................................................................-%

(e ) Sum of items 10(a), (6),

(c), and (d).........................




10. Temporary disabilities

(See instructions II.)

I f manufacturing, please indicate:
(a) Principal
materials used ( e .g ., A lu m in u m c a s t in g s , m a ­
ch in e d p a r t s , a sse m b le d p a r t s ; ra y o n s t a p le , w ool y a r n , co tto n

T im e c h a rg e s
(se e o v er)

7. Deaths

3. Operations: I f hours averaged less than 1,500 or more
than 2,500 per employee, please supply the following data:
Hours per
week per

N um ber
o f c a se s

11. Grand total— All disabling in­
(Sum o f items 7 , 8, 9 , and 1 0 (e ). I f no
disabling in juries, enter " 0 ” ) . . . . . .

f a b r ic s , e tc .— in th e fo r m b r o u g h t in to p l a n t ; d o n o t l i s t m a ­
t e r ia ls p ro d u c e d in t h is p la n t.)

12. Medical treatment cases

(In ju rie s requiring only first-aid or medi­
cal treatm ent. I f records o f these cases
a re not readily available, enter “ N .A ."
DO N O T include in count o f disabling
in juries a b o v e .) ...............................................

General types of operations performed (e.g., foundry,
machine shop, assembly; spinning, weaving, sewing,
etc.) ...............................................................................


F re q u e n c y


13. Injury rates


(fo r office use only) ...........................................

S e v e rity



B.Lt.S. 1418

Describe below all deaths, permanent impairments, and cases of unknown duration. If there is any question as to how or whether
a case should be reported, please describe fully.
7. Deaths (In clud e all d e ath s re su ltin g from work in juries w hich occurred du rin g re po rt year, re gard le ss of the tim e betw een in ju r y a n d d e ath .

( *)

Date of

Date of




Case No.

See in stru c tio n s I I I —Ite m 7.)

(D escrib e circum stan ces, an d in d icate in w h at w ay d e ath w a s w ork-connected.)




Scratched finger while loading truck; infection developed; died of blood poisoning..

6, 000


8 and 9. Permanent impairments (In clu d e a m p u tatio n s,
u n rep aired in guin al hern ias.

Case No.

perm anent loss of u se of m em bers or p a rts of m em b ers of the bo d y , p e rm an en t im pairm en t of fun ction s, an d
See in structio n s I I I —Ite m s 8 an d 9.)

Nature of injury

Part of body injured

(A m p u tatio n , d islocation , fracture, stra in ,
silico sis, etc.)

Example— Amputation....,..............................
Example— Crushed bones and torn ligaments__ _

(A rm , h an d , finger—sp ecify d ig it a n d
jo in t—leg, eye, etc.)

Actual days
of disability



Middle joint, index finger..........


loss of use of —if known
body part

1....... .....

10(d) Cases of unknown duration (F o r cases still open, giv e to ta l d a y s lo st to d a te an d
o r ex te n t o f p rob ab le pe rm an e n t im p airm en t.

doctor’s b e st e stim a te o f ad d itio n a l d a y s before em ployee will be able to w ork,
F o r oth er cases, give a s m an y d e tails a s p o ssib le.!_________________________________________________________________ _

Case No.

Nature of injury

Part of body injured

(A m p u tatio n , dislo catio n , fractu re, stra in ,
silico sis, etc.)

(A rm , h an d , finger—sp ecify d ig it an d
jo in t—leg, eye, etc.)




S tr a in

. ________


B ack


Actual days
of disability
to date
of report

of future

( I f ad d itio n a l sp ace is ne ed e d , a tta c h an o th er sh eet.)



•U.S. GOVERNMENT PRINTING OFFICE; 1965— ,"885-713 #28 89 5 * 7 1 3



Report prepared by.... .........................................

Do not use


April and acceptance of reports is terminated
about June 30.
In early 1966, cooperative arrangements were
in effect with 11 State agencies—Labor Depart­
ments or Industrial Commissions. These agencies
are in Alabama, Connecticut, Florida, Indiana,
Maine, New Jersey, New York, Pennsylvania,
South Carolina, Virginia, and Wisconsin. B L S
provides these State agencies with report forms
and postage and with technical assistance. The
State agencies return to B L S completed report
forms for establishments in their States.
Data for the automobile industry are obtained
through the Automobile Manufacturers’ Associa­
tion, which cooperates in securing reports from
its membership. For gas utilities and pipelines,
by a special arrangement of B LS, the Bureau of
Mines, and the American Gas Association, re­
port forms sponsored by the two Federal agencies
are used by the Association to collect data meeting
the statistical needs of all three organizations. As
was just indicated, rates for mining industries
and related processing industries are obtained
from the Bureau of Mines and for Federal gov­
ernment agencies from the Bureau of Employees’

Quarterly Survey
The monthly rates obtained in the quarterly sur­
vey also are based on voluntary reports collected
by mail. Again, individual reports are held in
Keport forms are mailed to cooperating estab­
lishments at the end of each quarter. After 4
weeks, second requests are sent to those who have
not responded by that time and seven weeks from
the first mailing the tabulations are closed. The
four quarterly reports of establishments cooperat­
ing in this survey are combined at the end of the
year and the summary of the twelve monthly en­
tries is used in the annual survey.
The Iowa Bureau of Labor and the Michigan
Department of Labor have cooperative arrange­
ments and participate in the quarterly program
in the way described above for the 11 State agen­
cies participating in the annual program. In all
other areas, the Bureau of Labor Statistics con­
tacts the reporting establishments directly.
In both the annual and quarterly surveys, the
completed reports are edited in the B L S office,


that is, they are examined for completeness, in­
ternal consistency, and reasonableness. When a
questionable entry is found, the person making
the report is asked by letter to review the entry
and either confirm or correct it. Special notes
have been prepared for the most common difficul­
ties, furnishing reporting establishments with a
more detailed explanation of what is wanted than
can be included in the instructions accompanying
the original request.

Collecting injury-rate information from all
establishments in the United States would be pro­
hibitively expensive and would place an unwar­
ranted burden on the business community. There­
fore, a sample of establishments is asked to report.
The total size of the sample is determined by the
need of a fairly large number of hours of exposure
to hazards for each industry which is to have
rates published. Injuries are statistically rare
events and rates based on small numbers of hours
of exposure are liable to extreme fluctuations. The
operating rule of the Bureau is that an industry
rate cannot be published which is based on less than
5 million man-hours (1 year for 2,600 full-time
The samples for both the annual and quarterly
surveys are made up of continuing panels of re­
spondents. The sampling plan for the annual sur­
vey calls for the inclusion of all establishments
with over 500 employees and of a sample from
smaller size classes. As reports are lost to the sur­
vey, an attempt is made to replace them with re­
ports from another establishment of the same size
class. Some of the cooperating State agencies re­
quire reports from all establishments in the State.
In these cases a sample of the reports to the State
is used in the national compiliation. Reports for
about 115,000 establishments are included.
The sample for the quarterly survey is limited to
15,000 establishments and in 1965 was somewhat
below that figure. The survey is limited to manu­
facturing, where large establishments are more
common. Also, to obtain a sufficiently large man­
hour base to publish monthly rates, it is necessary
to include many large establishments. As a result
the proportion of large establishments is greater
than in the anual survey and the sample of small



establishments smaller. As larger establishments
characteristically have lower frequency rates,
adjustments are made in the estimating process to
correct this deficiency.

Estimating Procedures
In editing, all the reports are reviewed and those
with questionable entries returned for confirma­
tion or correction. Annual survey reports go
through another process at the same time. The
severity rate which is included in the statistics
published in the annual survey requires determina­
tion of time lost or charged for each inj ury. When
the employee has returned to work, the time act­
ually disabled is used as reported. For cases of
unknown duration the number of days is estimated
based on days of disability at the time of the report,
the doctor’s estimate of future disability as re­
ported, and arbitrary but reasonable number of
days, based on the Bureau’s past experience with
reports on similar injuries. For deaths and
permanent impairments, time charges are assigned
as defined in the American Standard Method of
Experience .




For individual industries, the rates are calcu­
lated as simple averages. The rates published for
industry groups and divisions are calculated by a
weighting procedure which gives each industry a
weight equivalent to its estimated total
The computed rates in the quarterly survey are
characteristically low because of the large pro­
portion of large establishments in the sample.
They are adjusted to levels determined in the more
comprehensive and more adequately balanced
annual survey for the same industries. Prelimi­
nary adjustments are made on the basis of rela­
tionships in the previous year’s surveys and final
adjustments are made when the annual survey
rates for the year of reference become available.
These adjustments correct the level of the monthly
and quarterly rates, but preserve the month-tomonth fluctuations and short-term indications.

Analysis and Interpretation
Mail surveys cannot collect information on the
underlying or immediate causes of injuries. Uses

of these statistics fall into two broad classes; they
are used as general indicators of the level and trend
of injury experience to indicate the success of gen­
eral safety work and areas needing greater efforts,
and as basis of comparison for individual estab­
lishments. Therefore, analysis is limited to com­
parisons of broad trends and identification of in­
dustries with high or low and growing or shrink­
ing rates. Meaningful analysis of individual in­
dustry rates is impossible currently because of their
large number, but the industry coverage of each
rate is carefully specified in terms of the Standard
Industrial Classification (SIC) code numbers, to
permit comparisons with individual establishment

Presentation and Uses
The annual survey provides injury-frequency
rates for 490 manufacturing and 190 nonmanufac­
turing categories. The other measures published
for a majority, but not all, of these industries are:
injury severity rates; average days of disability
per case for all disabling injuries and separately
for permanent partial and temporary total dis­
abilities; and percent of disabling injuries result­
ing in death, permanent impairment, and tempo­
rary total disability. For the years 1959-62, a
period of revision of the series on the basis of the
1957 edition of the SIC Manual, average days of
disability per case are available only for all dis­
abling injuries and the distribution by severity
class is not available.
The annual data are published in a press re­
lease available late in the year following the year
of reference and in the Monthly Labor Review .
Injury-frequency rates only are available from
the quarterly survey. They are available for 140
industry categories. The rates are published for
each month, accompanied by a rate for the entire
quarter. Cumulative rates for six and nine
months and annual rates are published during the
year. These rates are published in a press re­
lease available about 2y2 months after the quarter
of reference and in the Monthly Labor Review .
The annual work injury rates constitute the
basic measures of work-injury occurrence in the
United States. A potentially rich source of similar
information lies in Workmens’ Compensation

statistics. This latter source does not meet cur­
rent needs, however, because of omission of in­
juries when disability is of short duration,
variation in coverage under State laws, differ­
ences in interest in and resources for the develop­
ment of statistics, and lack of coordination in
techniques and procedures for producing statistics.
Other statistics are available for some sectors of
the economy, but these, too, are incapable of com­
bination to form national indicators.
The B L S annual survey, therefore, presents the
only comprehensive data available. The sectors
it does not cover are sometimes covered by similar
statistics (as in mining, transportation, and Fed­
eral government) and sometimes lacking in indi­
cators (as in agriculture and domestic service).
Rates from the survey permit comparison of the
incidence of injuries and of the resulting losses
in the various industries and industry groups
covered. In year-to-year comparisons, they indi­
cate the basic trends in injury occurrence and
severity. They point out the industries in which
accident prevention needs to be intensified and
provide measures of the success or failure of
industrywide safety programs. Most important,
they provide a norm or basis of comparison against
which management can evaluate the disabling
injury experience of individual establishments.
The monthly rates from the quarterly survey
provide measures of seasonal variations in injury
occurrence. They also give early indications of
trends of injury frequency. The data on which
these rates are based are used also in compiling
the annual rates. Their greatest importance in the
safety movement, however, is that their existence
helps to maintain interest in safety records in
the periods between the basic annual surveys and
provide current norms for comparison with the
experience of individual establishments.

Annual Estimates of Volume of Injuries
Annual and quarterly survey results also serve
the important internal function of supporting the
estimates of the total volume of work injuries
which the Bureau publishes annually. These esti­
mates cover all employees in the United States ex­
cept domestic service workers. The estimates
cover the total number of disabling injuries and
the number of deaths for all workers and for em­


ployees only. Separate estimates are provided for
agriculture; mining; contract construction; manu­
facturing; transportation and public utilities;
trade; and finance, service, government, and mis­
cellaneous industries. Estimates of permanent
impairments and of temporary-total disabilities
are provided for contract construction, manufac­
turing, and trade.
These annual estimates of work-injury volume
and of the resulting manpower losses are prepared
in cooperation with the National Safety Council.
They represent the combined judgment of the
technical staffs of the two organizations based
upon a pooling of all data available to either group.
In the absence of a centralized system of report­
ing work injuries in the United States, the accu­
mulation of data providing national totals must
be based upon the assembly of many bits of data
drawn from a wide variety of sources. These
basic data frequently overlap or omit entirely
certain segments of employment. Additional
problems are introduced by a lack of uniformity
in the reporting and compilation procedures of
the organizations from which the basic data are
The State workmen’s compensation agencies and
certain Federal agencies constitute the primary
sources of the data on which the estimates are
based. In the Federal service, work-injury data
for particular segments of the economy are regu­
larly compiled by the Bureau of Labor Statistics,
the Bureau of Mines, the Interstate Commerce
Commission, the Office of Vital Statistics, and the
Bureau of Employees’ Compensation. The De­
partment of Agriculture and the Coast Guard pro­
vide intermittent data for operations under their
jurisdiction. The Interstate Commerce Commis­
sion, the Coast Guard, and the Bureau of Employ­
ees’ Compensation have compulsory reporting
requirements and obtain practically complete re­
porting in accordance with their respective regu­
lations. The other Federal agencies operate on a
voluntary reporting basis yielding sample cover­
age. Reporting requirements of the State com­
pensation agencies vary widely, but reporting is
compulsory and reasonably complete within the
respective regulations.
Data drawn from the National Safety Council
surveys and from surveys conducted by various
trade associations, such as the American Petroleum



Institute and the Portland Cement Association
supplement the data from the public agencies.
(Some association data are used in the compilation
of frequency and severity rates to avoid having the
Bureau make duplicate requests for information
already available.)
The estimating procedure requires reconciliation
of the various available data with standard report­
ing definitions, evaluation of the coverage in each
segment of the economy, and direct expansion of
the adjusted data to the total estimated employ­
ment in each area of industrial activity.
The estimates constitute an overall evaluation of
the magnitude of the occupational injury problem
in the United States. They indicate the aggregate
social and economic losses resulting from work
injuries and emphasize the national interest in
advancing accident-prevention activities.
The estimates for mining and quarrying, manu­
facturing, and rail transportation are based upon
very comprehensive data and are considered as
having a high degree of accuracy. The estimates
for construction, public utilities, miscellaneous
transportation, trade, and for finance, service, gov­
ernment, and miscellaneous industries are based
upon less comprehensive data, but are considered
reasonably accurate. The estimates for agricul­
ture are based upon fragmentary data and may
reflect a comparatively high degree of error. Tests
have indicated that underreporting is prevalent in
respect to agricultural injuries. The estimating
error, therefore, is probably that of underestimat­
ing rather than of overestimating.

The Bureau’s injury-rate surveys provide infor­
mation on a broad national basis. Because of
coverage limitations, however, the data cannot be
presented in State or local breakdowns. Supple­
mentary State details are available in a few States
from similar surveys conducted by the State labor
From an accident-prevention standpoint, it is
recognized that the most useful injury and accident
information is that relating to particular processes
and operations. Operating requirements, how­
ever, restrict the presentation of quarterly and

annual injury-rate data to the standard industrial
classifications. Reporting establishments are clas­
sified according to the money value of their leading
products or services, rather than in terms of proc­
esses or operations.
Because of resource limitations, only a small
proportion of the nonrespondents in the mail sur­
veys can be called or visited to determine the possi­
ble bias due to nonreporting. Although there is
no “contest” incentive element involved in the
Bureau’s surveys, there is a possibility that estab­
lishments with unusually high injury rates may be
reluctant to report their experience. Some estab­
lishments with no disabling injuries in the report
period may assume their reports are not wanted,
even though a note on the report form states that
the data are wanted.
No studies have been conducted to determine the
net effect of nonresponse on the published rates.
However, both the sampling plan and patterns of
response tend toward higher representation of
large than of small establishments in the sample.
Large establishments tend to have lower injury
frequency and severity rates than smaller ones.
Therefore it is thought that the published rates
are more likely to be minimums than maximums.
The estimates of total work-injury volume are
measures of the injury problem as of a given time.
From period to period, however, they reflect
changes in the volume of employment, shifts in
industrial activity, and technological changes in
industry as well as changes in the level of work
safety. They are not, therefore, fully satisfactory
measures of progress or retrogression in accident
prevention, particularly in long-term comparisons.
The injury rates for manufacturing are based
upon broad and well-distributed samples and are
presented in relatively detailed homogeneous clas­
sifications. In some areas of nonmanufacturing,
however, coverage limitations prevent the presen­
tation of rates in the most significant detail and
thereby impose some limitations upon the data as
the basis for evaluation of an individual estab­
lishment’s experience.
No injury-severity data are collected in the
quarterly survey, since the final degree of disability
for many injuries cannot be determined in the
short period allowed for reporting after the end
of the quarter.


Technical References

1. A m erican S ta n d ard M ethod o f Recording an d M e asu rin g Work In ju r y Experience .
New York: American Standards Association, Z 1 6 .1 — 1954 (Reaffirm ed 1959).
Defines terms, establishes standard measures of accident experience, and provides
interpretations and guiding examples for use in applying the various standard indexes
of safety performance.
2. A m erican S ta n d ard Method o f Recording B a s ic F ac ts R elatin g to the N atu re a n d Occurrence
o f Work In ju r ie s . New York: American Standards Association, Z 16.2 — 1962 .
Describes basic concepts, definitions, analytical categories, and the rules to be
followed in selecting the categories for accident analysis purposes. Includes detailed
series of coded analytical categories for use in classifying accident information.
3. Blake, R. P., In d u stria l Safety, 3rd ed. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963.
Contains chapters describing methods used in appraising safety performance and
identifying injury sources and causes.
4. National Safety Council, Accident Prevention M a n u a l fo r In d u str ia l O perations , 5th ed.
Chicago, 111.: National Safety Council, 1964.
The chapters on “ Accident Records and Injury Rates” and “ Accident Investiga­
tion, Analysis, and Costs” are of special interest.
5. Simonds, R. H., and J. V. Grimaldi, Safety M an agem en t , revised ed. Homewood, 111.:
Richard D. Irwin, Inc., 1963.
Chapter 8 describes methods for locating and defining accident sources.
— M a u r ic e F . B r e sn a h a n

Office of Productivity, Technology and Growth

222— 617— 66—




Chapter 27. Work Injuries in Maritime Activities

Description of Survey

The Longshoremen’s and Harbor Workers’
Compensation Act, administered by the Secretary
of Labor, provides a system of compensation for
death or injury suffered by employees engaged in
maritime employment upon the navigable waters
of the United States (including any drydock).
Employees covered by the act are largely em­
ployed in longshoring and in ship and boat con­
struction or repair. These are industries which
have had extremely high work-injury rates, as
revealed by occasional B L S studies over many
years. In 1958, Public Law 85-742, amending the
section of the act which relates to safety regula­
tions, was enacted to aid in the reduction of these
extremely high rates. It placed in the Depart­
ment of Labor some new administrative functions
which are carried out by the Bureau of Labor
Standards (L S B ). The amended act authorizes
the Secretary of Labor to issue regulations or or­
ders which will protect the life, health, and safety
of covered employees. Among the regulations is­
sued under this authority are those requiring the
maintenance of records and the making of reports
which will permit preparation of standard workinjury frequency and severity rates. B L S acts as
the agent of L S B in collecting reports and compil­
ing rates under the regulations.
Collection of standard work-injury data for
maritime employment was begun early in 1960.
B L S has been collecting statistics for L S B since
that time, covering workers subject to the act.
Provisions of the act are limited generally to em­
ployees aboard vessels afloat or in drydock. Many
of the establishments included in the survey are
also included in the B L S basic work injury sur­
veys, which cover all employees.1 In order to re­
duce the number of requests for information, B L S
requests on one report form both the information
needed to serve L S B and the information for all
employees of the reporting e s t a b l i s h m e n t s ,
whether or not they are subject to the provision of
the act.

There are three parts to this survey; two closely
resemble the B L S quarterly and annual surveys,
the third collects information for classification by
activity. All three parts are conducted by mail.
However, field personnel of L S B are available to
advise respondents on problems in reporting. At
the beginning of the year, respondents are asked
to provide information necessary to classify the
establishment in the Standard Industrial Classifi­
cation system2 and also the information necessary
to classify the activity of the group of employees
subject to safety and health provisions under the
Longshoremen’s and Harbor Workers’ Compen­
sation Act. Each quarter respondents are asked
to provide injury and employment data basic to
injury-frequency rates for each of the 3 months
of the quarter for the entire establishment and
separately for employees subject to the act. At
the end of the year, they are asked for data re­
quired to compile injury-severity rates, again for
the entire establishment and for employees subject
to the act.
As in the similar basic B L S surveys, the Ameri­


can Standard Method of Recording and Measur­
ing Work In jury Experience (Z16.1) defines the

measures used in the maritime survey. The fre­
quency rate used is the number of disabling in­
juries per million employee-hours worked, and the
severity rate is the number of days lost or charged
per million employee-hours worked. In addition,
the average number of days lost or charged per
case and the percent of disabling injuries resulting
in death, permanent impairment, and temporarytotal disability are computed.
Injury and employment data are collected in
two sections of the forms. Section A, identified
as “Employments subject to the safety and health
1BLS basic work injury surveys are annual and quarterly
surveys in manufacturing and selected nonmanufacturing in­
dustries. See chapter 26, “Frequency and Severity of Work
Injuries” , pp. 197-207.
2See appendix B.



regulations issued pursuant to the Longshore­
men’s and Harbor Workers’ Compensation Act”
must be completed to meet provisions of the act,
but section B, covering all employees, is intended
to meet B L S needs for the basic surveys and re­
sponse to it is voluntary.

hours and injuries of the establishment to be used
in the basic B L S programs. Submittal of the
latter information is voluntary.

Data Sources

Three report forms are used to collect the
information outlined above as necessary in this
program. B L S 2672-A requests information of
the services or products of the entire establish­
ment and for identification of specific maritime
operations and services performed with an in­
dication of which of them accounts for the largest
number of man-hours. It also provides for a de­
scription of the estimating procedure, if it is neces­
sary to estimate the number of man-hours worked
in employments subject to the act. This form is
completed early in each year, with the activities
described referring to the previous year. B L S
2672-B requests for each month of a quarter the
number of disabling work injuries (deaths and
others separately) and the numbers of employees
and man-hours. It is collected in the month fol­
lowing the end of each quarter. B L S 2672-C re­
quests all the additional information indicated
above as being needed at the end of the year and is
submitted in the 6 weeks following the end of the
calendar year.
The Bureau of Labor Standards furnishes the
Bureau of Labor Statistics with lists of employers
required to report in the program. Beport forms
are mailed to them by B L S and all returns are
by mail. However, field personnel of L S B are
available to advise respondents on problems of
reporting. At the end of the period allowed for
submitting reports, B L S gives L S B the names of
employers from whom the required portions of the
reports have not been received so that a follow-up
may be made and the required reports obtained.

There are about 2,000 establishments in the
United States with employees in activities that
make them subject to the act. Certain of these,
with less than 6,000 employee-hours per year sub­
ject to the act, are administratively excused from
reporting by administrative rule. The remaining
1,600 establishments provide the information in
the survey. From employment and injury rec­
ords, they provide information on the number of
employees, the total number of hours worked, and
the number and kinds of injuries. Additional
information on the duration of disability is ob­
tained at the end of the year, together with neces­
sary information for determining time charges
for cases of permanent impairments. The infor­
mation on number of employees, number of hours
worked, and number of deaths and of other dis­
abling injuries must be obtained for each month
of the year to permit calculation of monthly in­
jury-frequency rates. The collection of this infor­
mation is quarterly, however, and data for each of
the three months are obtained at one time. Data
for the computation of severity rates are needed
only once during the year. Therefore, at the end
of the year a summary of the number of disabling
injuries is obtained in somewhat greater detail
than in the quarterly reports. For many cases in
which the full duration of disability was not
known when the quarterly reports were submitted
it can be accurately reported at the end of the
year. For cases still open at the end of the year,
additional information on the characteristics of
the injury and on the estimated total duration of
disability is needed. Additional information on
permanent impairments also is needed at this time
to assign time charges as required by the Z16.1
Standard for the calculation of severity rates. In
all of these matters information is needed sepa­
rately for hours and injuries coming under the act,
as the amendments to the act require, and for all

Collection Methods

Estimating Procedures
As the completed report forms are received they
are reviewed for completeness, internal con­
sistency, and reasonableness. Questionable re­
ports are returned for confirmation or correction.
From the information on B L S 2672-A, an indus­
try code is assigned to the entire establishment—



from the 1957 edition of the Standard Industrial
Glassification Manual and its 1963 Supplement—
and an activity code to operation under the
act—from the list appearing on the form. In
connection with this editing of B L S 2672-C, time
charges are assigned for cases of permanent im­
pairment and predictions of total days of disa­
bility made for cases not yet closed. The time
charges are assigned as defined in the American
Standard Method of Measuring and Recording
Work In ju ry Experience. The necessary approxi­

mations of total days of disability are made on the
basis of all information available, the doctor’s pre­
diction of future disability and experience with
similar cases.
The assigned codes and edited data are trans­
ferred to punch cards. Summaries are made and
rates computed on data processing equipment.
The measures computed are:
Number of disabling injuries
__________ X 1,000,000_________
Total number of employee-bours1
Total days lost or charged X 1,000,000
Severity rate=Total number of employee-hours ’
Total number of days lost or charged
Average severity=
Number of disabling injuries
Frequency rate=

Bates are calculated for a considerable number
of regions, districts, and ports. These rates are
not published by B LS, but are transmitted to L S B
for publication and internal use. While B L S
advises L S B of the technical characteristics of
the rates substantive analysis is done entirely by
LSB . Using these rates and a wealth of other
information gained in carrying out its responsi­
bilities in the area of maritime safety, L SB has
been able to provide a considerable amount of
safety guidance to employers. In the first 5 years
of the program significant reductions in injury
rates have been accomplished.
While the principal presentation of the data
has been through L SB , data reported for entire
establishments have been incorporated in rates
compiled in the basic B L S programs. Also, rates
for the marine cargo handling industry have been
published in a separate B L S release. This in­
dustrial classification is a new one, appearing for
the first time in the 1963 Supplement to the
Standard Industrial Glassification Manual. 3 A
full range of rates for 1963 was published for five
regions and for three activities in the new industry.

Uses and Limitations
The largest use of these rates has been that for
which the program was designed, assistance to the
Bureau of Labor Standards in carrying out its
responsibilities under the safety and health provi­
sions of the Longshoremen’s and Harbor Workers’
Compensation Act. Not only the statistical results,
but also the activities of the L S B field staff in
assisting respondents, have been useful in accom­
plishing this obj ective. Further, the results of the
surveys have been available directly to employers
and to their safety officers as indicators of the na­
ture and extent of the injury problems, aids in di­
recting safety efforts, and measures of success or
Full-establishment reports have strengthened
the basic B L S injury statistics, notably in the
marine cargo handling and ship and boat construc­
tion and repair industries. As was mentioned
above, B L S has been able to publish rates for the
new marine cargo handling industry in a separate
The rates obtained in these surveys are intended
for administrative use in area of responsibility of
LSB . They represent high-hazard activities with­
in industries and cannot be compared directly with
the whole-industry rates which are available from
the basic B L S surveys. The special-purpose na­
ture of the rates limits their comparability to gen­
eral-purpose statistics in two ways: The activities
classified are not Standard Industrial Classifica­
tion industries and the units reported are not estab­
lishments as defined in the S I C Manuel. Thus
the relation of the coverage in these surveys to the
coverage in other surveys cannot be stated with
precision. Again, the geographic areas repre­
sented by most of the rates coming from the sur­
veys are administrative areas. Statistics for more
generally used areas cannot be matched with these
areas for comparison. To the extent that fullestablishment reports from the surveys can be used
in the basic B L S compilations, these limitations
are made less important.
3 Industry 4463, “ Marine cargo handling” , includes establish­
ments engaged in activities directly related to marine cargo han­
dling from the time cargo, for or from a vessel, arrives at shipside, dock, pier, terminal, staging area, or intransit area until
cargo loading or unloading operations are completed. This in­
dustry includes the operation and maintenance of piers, docks,
and associated buildings and facilities; but lessors of such fa­
cilities are classified elsewhere.


Technical References

1. A m erican S ta n d ard Method o f Recording an d M e asu rin g W ork In ju r y E xperien ce , Z 1 6 .1 , Approved
December 16, 1954, Reaffirmed June 23, 1959. New York, American Standards Association.
Basis of the injury frequency and severity rates compiled for maritime activities and for
basic injury statistics.
2. Recording an d R eporting W ork -In ju ry Frequency an d Severity D a ta Concerning R epairm en a n d Other
H arbor W orkers , 29 Code of Federal Regulations, Part 1506.
Legal requirement for the statistical activities.
3. U.S. Department of Labor, Bureau of Labor Standards, S afety a n d H ealth R egulation s fo r Longshoring June 1960 (and Supplement, 1961).
This publication, and the three below include Public Law 85-742 and regulations.
4. U.S. Department of Labor, Bureau of Labor Standards, S afety a n d H ealth R egulation s fo r S h ip
B reak in g , Safety a n d H ealth R egulation s fo r S h ip B u ild in g , Safety a n d H ealth R egulation s fo r S h ip
R e p a ir . 1965.

— M a u r ic e F . B r e s n a h a n
Office of Productivity, Technology and Growth

Chapter 28. Injury Rates and Accident Causes
The risks to health and safety faced by workers
because of their occupations were among the
earliest subjects of interest to the Bureau. Ac­
cident records were cited in reports on various
industries before the turn of the century, and by
the First World War, a number of reports had been
published on dangerous machinery, lead poisoning,
and other industrial hazards. While the small
number of industries for which the basic recurring
injury surveys were conducted was slowly increas­
ing, occasionally more intensive surveys could be
made. In the twenties and thirties, a few specific
industry accident experience reports were pre­
pared on the basis of special mail surveys. Indus­
tries chosen were those with high-frequency rates
in the Bureau’s continuing annual injury statistics
During the Second World War, similar special
studies were made for foundries, longshoring, and
the slaughtering and meat-packing industry. For
these studies, the mail survey of injury incidence
was supplemented by personal visits to plants to
determine the nature of injury and the causal fac­
tors associated with the individual accidents lead­
ing to the injuries. These studies proved to be
very valuable in planning safety work. The series
of studies was continued after the war, and they
make up the present special studies program.

Description of Survey
These intensive studies collect disabling injury
data in more detail than is possible in the Bureau’s
recurring surveys. In addition, they gather in­
formation on the nature and causes of the accidents
producing the injuries, on characteristics of the
injured worker, and on the environment in which
he was working when injured. The number of
studies is limited and each covers a selected indus­
try or occupation over a specified period of time.
In recent years, one industry a year has been

Injury-Incidence Data
The basic measures of injury incidence are the
injury-frequency rate, the injury-severity rate, and
average severity. These are discussed in chapter
26 and defined in the American Standard Method
of Recording and Measuring Work Injury E xperi­
ence (Z16.1). The injury-frequency rate meas­

ures the number of disabling injuries per million
man-hours of exposure. The injury-severity rate
is the number of days lost, or, for permanent dis­
abilities and death, charged according to a pre­
scribed scale, per million man-hours of exposure.
Average severity is the number of days lost or
charged per disabling injury.
In the detailed special surveys, it is possible to
compute these measures for a number of sectors of
the industry covered. Some of the sectors may
be geographic regions, departments common to es­
tablishments in the industry, occupational groups,
or groups classified by size of establishment (num­
ber of employees).
Collection by mail survey of information basic
to these measures from establishments in the in­
dustry covered is the first of two parts of a typical
special study, and the report on the survey is the
first of two reports.

Accident Cause Data
While the incidence rates indicate the location
and seriousness of hazards in workplaces, the
identification of the nature and causes of accidents
constitutes the first substantial step toward elimi­
nation of hazards. This step is taken in the second
part of the special studies, the “cause studies,” and
in the reports that come from them.
The statistics and analysis based on the cause
studies are intended to assist the accident preventionist by identifying the events and circumstances
which most commonly lead to the occurrence of
injuries. They identify the most significant haz­
ards and indicate the specific accident-prevention
activities which most need to be emphasized. This
is accomplished by cross tabulations of characterisistics of the accidents and of the injuries they


As is the case of definitions of injury frequency
and severity rates, the commonly accepted analysis
procedure is contained in an American Standard,
the American Standard Method of Recording
B asic F acts Relating to the Natu/re and Occurrence
of Work Injuries (Z16.2). Bureau staff members

had an important part in the development and
updating of this Standard, which is in general use
in the United States. Under the Standard, eight
characteristics of the injury and accident are re­
corded, without reference to which are of greatest
importance in the particular case.1 Standard
terms for the item recorded, with tabulating codes,
are included in the Standard in some detail and
proper use of the standard analysis requires fairly
close adherence to them.
These surveys cover the coterminous United
States. Usually they relate to Standard Indus­
trial Classification categories. In some cases, as
school lunchrooms and hospitals, both Government
and private establishments are included.
Experience indicates that the most effective
safety activities are those carried out on an in­
dustrywide basis. For this reason, the special
studies are planned with the possibility of an in­
dustry safety campaign in mind. Of two indus­
tries with equally high-frequency rates, the one
which has a trade association or some other indus­
try group to organize and support a safety pro­
gram will be considered the better prospect for a
special survey.

Data Sources
Information required for the first part of these
surveys is very similar to that required for the re­
curring surveys of the frequency and severity of
1An example from the Standard will indicate briefly the
nature of the characteristics recorded:
“A circular-saw operator reached over the running saw to
pick up a piece of scrap. His hand touched the blade, which was
not covered, and his thumb was severely lacerated.
Analysis category

work injuries described in chapter 26, except that
it is in much more detail.
Nearly all the information required for the
measures of injury incidence is collected by mail,
although data for a few large establishments that
are virtually certain to fall in the smaller sample
for the cause part of the survey are requested in
personal visits in connection with that part. The
mail-report form asks for three kinds of informa­
tion. First, details on a number of characteristics
of the reporting unit—the establishment—are re­
quested. These vary considerably with the indus­
try covered and are those characteristics which a
priori seem to be related to differences in hazards.
They usually will include location and details of
operation, product, or service. Second, the aver­
age number of workers and the total number of
hours of exposure for a 1-year period are requested,
not only for the establishment as a whole but also
for departments, operations, or occupational
groups within the establishment. Third, provi­
sion is made for entry of information about the
injury and the injured person for every disabling
injury during the year studied.
This information is requested from a large sam­
ple of establishments, or, if the industry is fairly
small, from all establishments in the industry.
This permits calculation of injury frequency and
severity rates in great detail and some analysis of
characteristics of injuries for industry sectors.
Information required for the second part is
much more complex. Individual case records
must be obtained covering each of the characteris­
tics considered by Z16.2. Where establishment
records do not include all the facts desired, it is
frequently possible to obtain them by interviewing
the persons injured or their supervisors or by
observation of the place of the accident.
Information for this part of the studies is ob­
tained in personal visits to the reporting establish­
ments by Bureau analysts. Because the visits are
costly, the sample for the cause part of the survey
generally is limited to about 200 establishments.


Nature of injury______________ Laceration
Part of body affected___________ Thumb
Source of injury______________ Circular saw
Accident type__________________ Struck against
Hazardous condition___________ Unguarded
Agency of accident____________ Circular saw
Agency of accident part________ Blade
Unsafe act____________________ Cleaning, adjusting mov­
ing machine.”


Coflection Methods
As in the similar recurring surveys, information
for the first part of the surveys most often is trans­
ferred from establishment records to the mail-



report forms by safety or other personnel of the
establishments. Report forms and explanatory
materials are mailed to the establishments follow­
ing the end of the year to be covered. A second
mail request is sent to those not replying to the
first. I f less than 70 percent of those solicited
reply, further requests are made. Completed re­
ports returned to the Bureau are reviewed for
completeness, internal consistency, and reasonable­
ness. When a questionable entry is found, the
person making the report is asked by letter to
confirm or correct the entry.
The information in the second part of the sur­
veys is considered too complex for successful col­
lection by mail. The Z16.2 Standard is newer and
less generally used than the Z16.1 and many estab­
lishments do not use it or keep records of all the
facts it requires. Therefore, Bureau analysts col­
lect this information in personal visits to the
establishments. They record on worksheets all the
information available in company records and sup­
plement this by interviewing the injured persons,
their supervisors, plant medical or first-aid person­
nel, and others; and by observing the place of the
accident. The analysts have the mail report list­
ing all disabling injuries for the period studied,
and so, know in advance the number of records

Because of the detail in which the data are pre­
sented, the special studies require larger samples
than do the recurring surveys. The high-rate in­
dustries covered are frequently small in total num­
ber of establishments and employees. When this
is true, it is often possible and advantageous to
cover the entire industry in the mail survey.
Where the total number of establishments is very
large, a sample is selected for the mail portion of
the survey. The number of establishments in the
mail survey is usually between 5,000 and 10,000.
In the second part of the study, field agents
collect the data. Since a large sample would be
prohibitively expensive and would demand more
2The formulas for these injury measures are:
Frenuencv r a t e - Nnmber of dlsabling lnJuries X 1.°00 0(>(>

Total number of employee-hours worked
Severity ratc — Total of days lost or charged X 1,000,000
“ Total number of employee-hours worked

analyst time than is available, a sample of about
200 establishments is selected from all establish­
ments reporting injuries for this part of the sur­
vey. Because the mail reports are available for
the selection of this sample, a selection may be
made which guarantees a sufficient number of
cases for analysis in each stratum at minimum cost.

Estimating Procedures
The reports received by mail are edited and any
questions arising in them are settled through cor­
respondence with the reporting establishment.
Codes are assigned for the establishment charac­
teristics included in the report and to the depart­
ments or occupational groups for which hours
worked and injuries are reported. Frequency and
severity rates2 and counts of injuries are cross
tabulated by the coded characteristics.
The records made by analysts visiting establish­
ments in the second part of the survey are re­
viewed in the Bureau’s office. Codes are assigned
for the eight “analysis categories” of the Ameri­
can Standard and for some 10 to 15 other charac­
teristics of the establishment, injury, and the in­
jured person, for each case. The final tabulations
group the cases according to these factors and
provide a large number of analytical cross classi­
fications. These tables, running 50 to 60 pages, are
used in the Bureau’s analysis and reproduced in
the report on the survey.

Analysis and Presentation
The written report is intended to assist in the
prevention of accidents by indicating the nature
and extent of the problem, the frequency of vari­
ous kinds of injuries, and the circumstances most
frequently associated with accidents and resulting
Frequency and severity rates for injuries are
compared with those of other industries. Rates
for sectors of the industry also are compared.
This permits an assessment of the seriousness of
the general injury problem in the industry and
identification of the geographic, industry, occupa­
tional, and other areas in which corrective action


is most needed. The nature of injuries and parts
of the body injured most frequently in the various
sectors also are identified.
Better understanding of the nature and causes
of injuries is possible through analysis of data
from the second part of the study. The greatest
number of meaningful cross classifications of such
characteristics as nature of injury, part of body
affected, activity at time of injury, and so on can
be selected; the comparative importance of the
various injuries explored; and the nature of the
more and less severe injuries established. The
second part of the study increases the understand­
ing of factors involved in accidents resulting in
injuries. I f factors frequently present in, and
possibly necessary to, accidents can be eliminated,
repetition of the accidents may be avoided. The
analysis is aimed not at understanding of the
underlying causes of injuries, but at breaking
chains of events which have led to injury through
accidents. Unsafe acts and hazardous conditions
associated with accidents are the material of this
analysis. Those most commonly related to in­
juries in various sectors of the industry, to injuries
of various kinds, and to various activities are ex­
plored. Hypotheses concerning relationships
among variables associated with accidents are
tested statistically.
The analyses emphasize the circumstances and
combinations of circumstances which call for acci­
dent-prevention activities at every work site.
They also point out situations of high-accident
potential which, though less common, may be im­
portant in individual establishments.
A B L S “Report” is prepared for each of the
two parts of the survey. A summary of the most
important findings is followed by a description
of the survey with particular attention to the
definitions and standards used and the nature and
limitations of the statistics presented. This is
followed by the analysis just described and many
detailed tables of data.
While the “Report” is the principal medium for
releasing the survey results, several other means
are used to insure the widest possible distribution
of the information in the safety movement, where
it may be effectively used. A summary of the
findings is published in the Monthly Labor Review
and the findings are freely available for publica­
tion in trade or professional journals. A copy of
222-617— 66----- 16


the report is sent to each respondent on request.
When possible, the staff member in charge of the
survey reports directly to interested groups in the
industry in speeches to conferences or conventions.

Uses and Limitations
The ultimate purpose of these surveys is to
stimulate and assist in the prevention of accidents
and their resulting injuries. They aim to satisfy
this purpose by establishing the importance of
work injuries in the industry, identifying the in­
dustry sectors where problems are most severe, and
providing a first step to elimination of accidents by
identification of the factors involved in them.
For the safety man in the individual establish­
ment, the report provides a number of leads to ac­
cident prevention and provides an industry sum­
mary as a basis of comparison for the experience
in his own departments and occupational groups.
It also provides him with an indication of potential
accident problems, identifying general hazards
which have produced injuries in other establish­
ments but not yet in his own.
There is some evidence of the success of the sur­
veys. In 1968 a number of charts were made of
the trends of frequency rates in industries covered
by special studies, before, during, and after the
special study. They showed reductions in rates not
only f ollowing publication of the reports, but also
during the time the studies were in progress. Pre­
sumably, receipt of the report form and discussion
of the survey in trade journals, as well as the re­
ports themselves, provided an impetus to greater
Safety efforts.
Because of the small number of special surveys
that are conducted, the detail and analysis given
in the reports cannot be available for the great
maj ority of industries. Also, as safety campaigns
are conducted and conditions in the industries
change, the reports become less applicable. After
a few years, frequency rates tend to rise from the
low levels achieved during and just after the sur­
veys, but only in rare cases is it possible to repeat
a survey within 15 years.
Although the reports identify hazardous condi­
tions and unsafe acts frequently associated with ac­
cidents, they do not provide a judgment as to which
were most important or whether it was the unsafe



act or the hazardous condition that was the major
contributing cause; in most cases, both were found
to have been present. Neither can the analysis
determine underlying causes of accidents beyond
the immediate circumstances. The American
Standard Z16.2 consciously avoids making these
judgments on the premise that a survey cannot
successfully identify underlying causes. Again,
the reports do not provide engineering solutions to
accident prevention, but are limited to providing
the safety engineer with clues and indications to
conditions that merit his attention because of fre­
quent association with accidents. Safety engineer­
ing practices, safety codes, and built-in safety fac­

tors in equipment are not within the scope of the
Bureau’s work injury studies.
The large samples used minimize sampling error.
However, there is a problem of bias of nonresponse.
A response of 70 percent or more is used as a signal
to discontinue followups, though reports received
above this level are used. No survey of nonre­
spondents is made to determine their injury expe­
rience. The sample results are considered as
slightly understating the levels of frequency and
severity rates.
— M a u r ic e


B r esn a h a n

Office of Productivity, Technology and Growth

Other Programs
Chapter 29. Foreign Labor Conditions and International Comparisons
Almost from its inception, the Bureau of Labor
Statistics has conducted research on labor condi­
tions and developments abroad. Analysis of re­
search data on labor in foreign countries and the
making of international comparisons constitute
only two of the international activities carried on
by the Bureau,1 but from the point of view of
technique they are the ones of most interest to
users of research reports. The United States
undertook these research programs because (1)
information on labor conditions published by a
majority of these countries is not available in Eng­
lish; (2) summary information is often not avail­
able at all and is seldom available in a form readily
usable by American businessmen, labor unions,
Government officials, and others; (3) the sheer
variety of source materials from nearly 200 inde­
pendent countries and dependencies is impossible
for most users to keep up with; and (4) it is often
impossible for anyone but an expert to have trust­
worthy judgment of the quality of source mate­

Description of Reports
The foreign labor research reports published by
the Bureau are partly general and partly statisti­
cal. The principal types may be listed as follows:
1. Labor Developments Abroad, a monthly pub­
lication covering not only important developments
in labor abroad, but also the principal labor and
price statistics in leading important countries.
2. Labor monographs on single countries, in­
cluding general labor summaries and labor law and
practice reports. These reports are factual, but
not primarily statistical.
3. Brief labor digests on single countries.
4. Bibliographies, chiefly for a country or an
area and often with annotations or brief sum­

International comparisons presented in fea­
ture articles and bulletins. These are statistical
in nature and contain, within limitations of the
medium, a careful adjustment for differences in
definitions and methods. So far, principal com­
parisons published have concerned labor cost, un­
employment, job vacancies, and worktime required
to buy selected consumer goods and services.

Data Sources and Collection Methods
The Bureau receives a mass of statistical and
other material on labor developments abroad and
maintains this material in files classified by country
and by subject. The material includes (1) cur­
rent reports from labor attaches and other officers
of the Foreign Service in American diplomatic
missions and consular offices throughout the world;
and (2) periodicals and other publications issued
by labor ministries, bureaus of statistics, and
other governmental agencies of foreign countries,
the International Labour Organization and other
international organizations, and by banks, eco­
nomic research organizations, and other private
agencies in the United States and foreign coun­
tries. When specific supplementary information
is needed, the Bureau often requests it through
Foreign Service reporting channels.
The Bureau also maintains contact with re­
search institutions and individual scholars spe­
cializing in international labor, and with U.S.
firms which employ labor abroad. In addition,
staff members of the Bureau are briefed by labor
attaches and other Foreign Service officers, when
these officers arrive in Washington from their over­
seas posts; and, from time to time, members of the
1 Other major B L S activities in the international field include
providing orientation and factual advice to U.S. policy and pro­
gram officers on labor in foreign countries, and providing training
or orientation to foreign statisticians and other foreign visitors
(especially participants in exchange programs and technical co­
operation programs), on U.S. labor statistics methodology and
on the economic conditions of American workers.




research staff study foreign labor conditions at
first hand on carefully planned trips abroad.

Analysis and Uses
With the exception of international comparisons,
the major classification criterion in the presenta­
tion of reports is geographic. For an individual
country, the organization is by subject. The
sequence and coverage of subjects varies somewhat
in different reports, but usually includes the fol­
lowing general headings: (1) social, economic, and
political background; (2) manpower (including
labor force, employment, unemployment, man­
power planning and utilization); (3) labor law
and administration; (4) wages, supplementary
benefits, hours and working conditions (including
prices and real earnings); (5) labor organizations;
(6) management organizations; and (7) labormanagement relations.
Most reports are concerned with either the basic
situation at the time the report was written or
with a recent development or trend. With a few
exceptions, of which real wages is one, less atten­
tion is given to historical development of long­
term trends. The extent of material to be covered
is so great that country monographs and, indeed,
all but the most summary information often can
be written for a given country only about every
10 years, and many countries have never been
covered. Beports often contain an annotated
bibliography for those interested in more detail.
International comparisons studies were begun in
1960 on a regular basis. They differ from other
foreign labor reports in being primarily statistical
in method. In purpose and uses they also differ,
since their chief ultimate objective is learning
from foreign experience findings which may be
useful in the United States. The international
comparisons are prepared by subject, and the
methodology is intricate, requiring specific infor­
mation about details of definition. This informa­
tion is seldom, if ever, fully covered by regular
Foreign Service reporting or published sources,
and great reliance is therefore placed by the
Bureau on specific requests to the Foreign Service
for required information.
Among the subjects the B L S has covered in

international comparisons is labor cost. Articles
containing basic information on “ The Bole of
Labor Cost in Foreign Trade” and on “ Interna­
tional Comparisons of Unit Labor Cost: Concepts
and Methods” have been published. No absolute
comparisons have yet been issued for any industry,
but trend data in index form have been published
for all manufacturing in eight principal trading
partners of the United States. Belative labor
cost trends are useful for analysis of the balance
of payments. Other topics studied have included
unemployment, work stoppages, and job vacancies.
Articles on worktime required to buy selected con­
sumer goods and services were published in the
period 1949 to 1952. This is a subject in which the
Bureau has had a continuing interest.
Unit labor cost is the ratio of total labor cost
(or expenditure), expressed in money terms, to
total output produced by the labor (in concert
with other factors of production), in physical
terms. The unit of measurement may be expressed
as, for example, dollars per ton or francs per
gallon. The greatest problem involved in com­
paring unit labor cost among countries arises from
the fact that labor cost data are available by
industry, whereas output is reported by product,
and since nearly all industries produce a number
of products, the question of labor cost allocation
(or product synthesis) arises. Previous efforts
suggest that the most fruitful approach probably
will be cost per unit of output on an industry-by­
industry basis.
The trend data published for all manufacturing
were taken, in most cases, from official statistical
reports issued by the individual countries. Be­
cause the data are not entirely comparable from
country to country and because more than one set
of data exists for most free industrial countries,
considerable effort was spent on preparing esti­
mates from different sets of data and analyzing
the differences in results before selecting the series
finally used. In most countries where it was avail­
able, the Bureau decided (1) to use as the measure
of labor cost the compensation of employees in
manufacturing, from the national accounts, and
(2) to use as the measure of total output the real
gross national product originating in manufactur­
ing. The chief reason for this decision has been
the greater comparability and comprehensiveness
of these measures.


As regards unemployment, the interest in inter­
national comparison centers on the percentage of
the labor force which is unemployed. Data pub­
lished by the individual countries on a current
basis is of different types, but most industrial
countries of the free world have made one or more
sample surveys of the labor force, using defini­
tions and methods fairly comparable to those in
the United States. These latter data, adjusted
as carefully as feasible to uniform definitions and
brought up to date by the best data currently
available, form the basis for the international
comparisons of unemployment rates published by
the Bureau.
For both international comparisons studies and
reports on foreign labor conditions, businessmen
are the largest single group of recipients, and their
uses of foreign labor information lie in two chief
areas: (1) their own actual and prospective opera-


tions abroad; and (2) the competition which they
are experiencing or can expect from abroad. The
uses of Government officials, labor unions, research
institutions, and others are more varied and often
less direct. But every year the world grows
smaller as Govemment-to-Govemment and people-to-people contacts become more frequent and
close knit. Adequate manpower planning, devel­
opment, and utilization are crucial to economic
progress in developing countries. The A F L -C IO
and its affiliated component unions are assisting
directly and through international organizations
literally hundreds of labor unions in Latin Amer­
ica, Asia, and Africa. Knowledge of the labor
and manpower activities of other countries is thus
increasingly important.
— W il l ia m C. S h e lt o n
Office of Foreign Labor and Trade

Chapter 30. Economic Growth Studies
The U.S. Department of Labor, in cooperation
with other Government agencies and private re­
search organizations, has undertaken an extensive
program of economic growth studies. These
studies are aimed at providing a more compre­
hensive and integrated framework than previously
has been available for analyzing the problems of
long-run economic growth in relation to employ­
ment opportunities.
Because this program has important implica­
tions for many agencies of the Government, an
interagency planning and coordinating committee
provides the guidelines for the program. The
chairman of the committee is from the Council of
Economic Advisers; other members come from
the Departments of Labor, Commerce, and the
Bureau of the Budget. Many elements of the
projections are being developed by other agencies
of the Government and by private research
groups. The integration of these separate proj­
ects into a consistent set of projections has been
undertaken by the U.S. Department of Labor, op­
erating under the guidelines of the interagency
coordinating committee.
A primary objective of the program is to de­
velop projections, under alternative assumptions,
of the rate and patterns of growth in the economy.
These would then provide a framework for assess­
ing a number of important economic problems,
including problems of manpower utilization over
the next decade. The projections will cover both
the supply of and demand for labor in consider­
able industry detail.
The estimated supply of labor will be based on
interrelated projections of the population, by age
and sex; school and college enrollment; educa­
tional attainment of the work force; household
and family formation; and labor force participa­
tion rates by age and sex.
The demand for labor will be based on projec­
tions, under alternative assumptions, of total in­
come and output of the economy and how this
output may be distributed among the various cate­
gories and detailed components of final demand

for consumption, investment, government expend­
itures, and foreign trade.
Estimates of final demand will then be con­
verted into direct and indirect output of all the
supporting industries which contribute materials,
parts, components, fuels, and transportation and
distribution services embodied in the final prod­
ucts. The basis for such computations is provided
by a study of interindustry sales and purchases in
the economy, and the projection of these inter­
industry relationships over the next decade to re­
flect anticipated changes in technology and, if
possible, relative costs. These interindustry rela­
tionships can then be used to convert projections
of end-product deliveries to estimates of output
requirements for each industry, covering inter­
mediate as well as final products. The industry
output requirements, along with projections of
hours of work and unit labor requirements, will
provide the basis for deriving comprehensive and
consistent estimates of the demand for labor on an
industry-by-industry basis. In addition, the pro­
jections of population, labor force, industrial out­
put, etc., along with projections of unit capital
requirements, will be used to derive estimates of
public and private investment and capital stock
required by an expanding economy.
The final stage in the program is the translation
of the industry employment estimates into occu­
pational requirements based on projections of
occupational patterns, by industry. These data
can be used to evaluate potential areas of substan­
tial surplus or shortage, and provide the basis for
occupational guidance and development of longerrun training programs.1
The elements in the integrated set of projections
previously indicated are being developed by vari­
ous units of the U.S. Department of Labor and
other Government agencies and private research
organizations. In particular, the Office of Busi­
ness Economics, U.S. Department of Commerce,
has prepared the table of interindustry relation­
ships which play a central role in the projections.
Some of these elements are products of the regular
economic and statistical research activities of the
agencies involved, but a substantial proportion are
1 See chapter 7, “ Industry-Occupational Matrix.”


being developed to meet the specific needs of this
These research projects include the following:
Bureau of Labor Statistics: Labor force projections,
productivity projections, industry price indexes, capital
flow table.
Office of Business Economics: Projections of capitaloutput ratios, residential construction projections, balanceof-payments projections, and projections of income and
product accounts.
U.S. Department of Agriculture: Projections of inputs,
for farm sectors and demand for farm products.


Bureau of Mines: Projections of inputs, for mining
Harvard Economic Research Project: Projections of
consumer expenditures, and projections of inputs for
selected manufacturing industries.
George Washington University and Council of State
Governments: Projections of State and local government
revenues and expenditures.
National Planning Association: Projections of Federal
Government revenues and expenditures.
— R onald E . K u t sc h e r
Office of Productivity, Technology and Growth

Appendix A. The BLS Seasonal Factor Method
An economic time series may be affected by reg­
ular intra-yearly (“seasonal” ) movements which
result from climatic conditions, model changeovers, vacation practices, holidays, and similar
factors. Often such effects are large enough to
mask the short-term underlying movements of the
series. By isolating and removing the effect of
such intra-yearly repetitive movements, the cur­
rent evaluation of a series may be made more
Seasonal movements are found in almost all
economic time series. They may be regular, yet
they do show variation from year to year and are
also subject to changes in pattern over time. Be­
cause these intra-yearly patterns are combined with
the underlying growth or decline and cyclical
movements of the series (“trend-cycle” ) and also
random irregularities, it is difficult to determine
the pattern with exactness.
More than a half-century ago, attempts were
made to isolate seasonal factors from time series.
Some early methods depended upon smoothing
curves by using personal j udgment. Other formal
approaches were periodogram analysis, regression
analysis, and correlation analysis. Because these
methods involved a large amount of work, rela­
tively little application of seasonal factor adjust­
ment procedures was carried out.
In the mid-1950’s, new electronic equipment
made more elaborate approaches feasible in sea­
sonal factor methodology as well as in other areas.
Using an electronic computer, the Bureau of the
Census developed seasonal factors based on a ratioto-moving-average approach. This was a major
forward step, as it made possible the uniform
application of a method to a large number of series
at a relatively low cost.1 Subsequent improve­
ments in methodology and in computer technology
have led to more refined procedures which are
both faster and cheaper than the original
The Bureau of Labor Statistics began its work
in seasonal factor methodology in 1959, primarily
to correct a deficiency in the method then used.
Prior to this time, when additional data became

available and seasonal factors were generated from
the lengthened series, the new factors sometimes
differed markedly from the corresponding factors
based on the shorter series. This could affect any
portion of the series. It was difficult to accept a
process by which the addition of recent informa­
tion could significantly affect the seasonal factors
for periods as much as 15 years earlier, especially
since this meant that factors could never become
final. The first method developed by B L S and
introduced in 1960 had two goals: first, to stabilize
the seasonal factors for the earlier part of the
series; second, to minimize the revisions in the
factors for the recent period.
Since 1960, the Bureau, through continued re­
search, has made numerous changes and improve­
ments in its techniques and in methods of applying
them. These changes were described2 as they
were introduced. The method introduced in Ja n ­
uary 1964 is described in the sections which follow.

Characteristics of BLS Seasonal
Factor Method
The B L S method is a ratio-to-moving-average
method. It assumes that the three component
parts—trend-cycle, seasonal, and irregular—are
multiplied together (multiplicative assumption) to
give the original observations.3 (See illustration
in charts 1 and 2.) The B L S method differs from
other similar methods in the following respects:
1. The initial trend cycle is improved by restor­
ing to it any residual trend-cycle which may have
found its way into the irregular component. This
adjustment for the deficiency in trend-cycle is de­
veloped explicitly and is available for review in
the form of a table of trend-cycle correction values.
2. The B L S method provides for changes over
time in the producing mechanism (changes in
samples, method of collection, method of estima­
1 Julius Shiskin, Electronic Computers and Business Indicators.
New York, National Bureau of Economic Research, 1957, Occa­
sional Paper No. 57.
2 See U.S. Department of Labor items in the list of Technical
References at the end of this appendix.
3 A parallel method assuming an additive relationship among
the components has also been prepared for experimental use.



tion) or in economic factors (introduction of guar­
anteed annual wage plans, etc.) .4
3. The B L S method calibrates each observation
in order to assign a supplementary weight which is
used in the various averaging processes. These
“credence f actors” reduce the impact of observa­
tions having large irregularities. They increase
the smoothness with which the seasonal factors
change over time, and they also keep large irregu­
larities out of the final trend-cycle.
4. A second way in which the B L S method at­
tempts to protect the final trend-cycle from large
irregular fluctuations is by using modified original
values for computing the centered 12-month mov­
ing average in the later stages of the procedure.
The credence factors are used in obtaining the
modified original values.
The B L S method is complicated and requires a
very large number of arithmetic operations. Such
a procedure would not be practical without an elec­

tronic computer. Since an earlier version of the
B L S method has been accepted by State agencies,
foreign governments, and private industrial con­
cerns, the method has been programmed for a
small computer system which is widely available.
Special emphasis has been placed on keeping the
handling and clerical requirements to an absolute
minimum, on providing as many aids as possible
for analysts or others using the results, and on
making the application of the computer program
as simple and efficient as possible. Only the data
4 A 5-year moving period is used to reflect such changes. The
use of the 5-year moving period allows the full impact of a
change to be reflected in a relatively short period of time. A
review of the basic U.S. labor force series, for example,
indicates that in the early 1960’s the standard deviation of the
irregular component is only 0.6 as large as it was in the early
1950’s. This reflects the improvements made in the survey,
such as the expansion in the number of sample areas, the in­
crease in the number of households covered, the changes in the
estimating methods, and the improvements in the design of
selecting households so that there would be a three-quarters
overlap in adjacent months and a one-half overlap over the year.






Chart 2.



and one specification card are needed to produce a
set of completely labelled tables for each series.
Any number of series with varying characteristics
(length of series, starting and ending dates, magni­
tude of original observations, final table patterns
desired) can be processed in a single running with­
out manual intervention. The computer program
for applying the B L S procedure is available upon

Basic Approach
The B L S method attempts to separate an eco­
nomic time series into three constituent parts:
the underlying movement or trend-cycle which is
a combination of the long-term, cyclical, and subcyclical movements ( T ) ; the annual repetitive
movement within the year, or “seasonal” ( S ) ; the
irregular or residual unexplained movement (I).

These three components, when multiplied together,
completely and exactly exhaust the original obser­
vations (O ). (See charts 1 and 2.) The exact
allocation among the components is somewhat arbi­
trary, because there are no simple criteria or gen­
erally accepted techniques for separating them.
The B L S method attempts to strike a good balance
between the conflicting objectives of smoothness
in the trend-cycle, stability in the seasonal, and
randomness in the irregular. The process is an
iterative one, with each successive iteration pro­
viding an improved estimate for each of the com­
ponents of the original series.
5 Upon request, a t nominal cost, the Bureau of Labor Statistics
will make its seasonal factor computer program available in the
following m anner: (1) A deck of punched cards (2321 cards)
containing the program instructions (in condensed instruction
format) ; (2) A deck of punched cards (11 cards) containing
a test problem (data and specifications) which produces 18
tables for checking purposes; (3) Two copies of a document to
be used as an operating guide and reference manual.


Three iterations are used, each having two
phases. The first phase of each iteration starts
with a centered 12-month moving average as the
estimate of trend-cycle. Seasonal-irregulars are
then secured and partitioned into seasonal and ir­
regular components. The second phase in each
iteration starts with an improved trend-cycle in
which the centered 12-month moving average has
been improved by recovering the residual trendcycle from the irregulars of the first phase. The
seasonal and irregular components are then de­
veloped as in the first phase.
The Bureau’s method uses “credence factors”
based on the irregulars from the second phase of
each iteration to reduce the impact of large irregu­
larities in the original observations. These are
supplementary weights in which observations with
sm all6 irregulars are given more weight than ob­
servations with large irregulars in calculating the
trend-cycle and the seasonal factors.

Detailed Procedure of BLS Seasonal
Factor Method
First Iteration
First Phase. The initial estimate of trend-cycle
(T) is the centered 12-month moving average of
the original observations. The trend-cycle value
for each month is divided into its corresponding
original value (O) to produce a series of seasonalirregular (S I) ratios ( 0 / T = S I ). Treating each
month separately (i.e., all January’s, all Febru­
ary’s) , the S I ratios are arrayed by year and mov­
ing 7-term weighted averages7 (S ') are secured
(in percentage form) as estimates of the unforced
seasonal factors. The unforced seasonals (S ') for
each year are then adjusted to yield a monthly
average of 100.0 for the calendar year.8 These
forced seasonal factors (S ) are the seasonal factors
for the first phase of the first iteration. Each sea­
sonal-irregular ratio (S I) is then divided by its
forced seasonal factor (S ) to yield the random or
irregular ratio ( S I / S = I ) . The irregular ratios at
this stage may contain some residual trend-cycle.9
To separate the residual trend-cycle from the truly
random variation contained in the irregulars, a 9term weighted moving average10 (trend-cycle cor­
rection) of the first phase irregulars is secured.
This completes the first phase of the first iteration.


Second Phase. The second phase of the first
iteration starts with an improved estimate of the
trend-cycle. This is secured by multiplying the
trend-cycle used in the first phase (the centered
12-month moving average of original observa­
tions) by the trend-cycle correction (weighted
moving average of first phase irregulars). Using
this improved trend-cycle, the second phase re­
peats the computational steps of the first phase
to develop new S I ratios, new unforced seasonals
6 References to “large” or “ small” Irregulars have to do with
the deviation of the irregular component from 1.000 in relation
to the standard deviation of such irregulars. A value of 1.423
or .577 would usually be considered large, while .997 or 1.003
would be small.
7 The weight patterns used are:
Weight pattern assigned to
seven consecutive
seasonal-irregular ratios
1st year........................................ .281 .270 . 242 207 ..........................
2nd year.............................................213 . 221 .213 .191 .162 ..................
3rd year......... ................................... 160 .179 .185 .179 .160 .137 ........
Middle years..................................... 120 .141 .157 Jfi4 .157 .141 .120
3rd from end.............................................. 137 .160 .179 JU85 .179 .160
2nd from end...................................................... 162 .191 .213 .J221 .213
End year....................................................................... 207 . 242 . 270 . 281
The underlined value indicates the year to which the} weighted average
8 The forcing is performed in two sta g e s: If the unforced sea­
sonals do not start in January, the first 12 unforced seasonals
are summed and the total divided into 1200 to provide a forcing
factor. This factor is then multiplied by the unforced seasonals
for the partial year only (through the first December value)
to provide the forced seasonals for the incomplete year at the
beginning of the series. A similar procedure is followed at the
end of the series if the unforced seasonals do not end in Decem­
ber. For each of the full calendar years between, a forcing
factor is computed by dividing the sum of the unforced seasonals
for the year into 1200. This factor is then multiplied by the
unforced seasonals in that year to produce the forced seasonals.
®This is because the 12-month moving average is not very
good at following abrupt or curvilinear changes in the level of the
original series, particularly in the vicinity of peaks and troughs.
Since the trend-cycle, seasonal, and irregular components com­
pletely and exactly exhaust the original series, any deficiency
of the first estimate of trend-cycle is transferred to the seasonal
and/or the irregular.
However, the seasonal factors are
secured by averaging seven SI ratios, each one year apart. This
averages the deficiency for seven different years, with the result
that the deficiency of the 12-month moving average, as an esti­
mator of trend-cycle, appears mainly in the irregular component.
The periods where the 12-month moving average is a poor esti­
mator of trend-cycle, usually have a run of consecutive irregular
ratios all on the same side of the base reference ratio of 1.000
instead of being scattered above and below this base.
70The weight patterns used are:
Weight pattern assigned to nine consecutive irregulars
1st month......................... S70 . 341 .256 .115-. 082 ...................................
2nd month_____ _____ 254 .£76 .254 .191 .086-. 061 ..........................
3rd month.......................160 . 214 . 231 .214 .160 . 072-. 051 ................
4th month........................ 067 .150 .199 . 216 .199 .150 . 067-. 048
Middle months............-.050 . 071 .157 . 209 . 226 . 209 .157 . 071-. 050
4th month from end........... --.048 . 067 .150 .199 . 216 .199 .150 . 067
3rd month from end.......................-.051 .072 .160 . 214 . 221 .214 .160
2nd month from end................................—.061 .086 .191 .254 . 276 . 254
End month............-..........................................-.082 .115 .256 .341 .370
The italicized weight indicates the month to which the weighted average



( S '), new forced seasonals (S ), and new irregu­
lars ( I ) . A t this point, the generated trend-cycle
and seasonal components represent the components
of the original series fairly well except for the
effect of highly deviant original values. In the
B L S method, the influence of such values is
diminished by using supplementary weights in
addition to the regular weights given in footnotes
7 and 10. The use of these supplementary weights
(called “credence factors” ) gives less impact to the
extreme observations (those with large irregular
components) and relatively more to the neighbor­
ing values with smaller irregular components.11
With the preliminary credence factors, the S I
ratios of the second phase are repartitioned into
seasonals and irregulars. The adjusted unforced
seasonal is a 7-term moving average of the S I
ratios, using as relative weights the product of the
standard 7-term weights (given in footnote 7) and
the preliminary credence factor associated with
each value. These unforced seasonals (S ') are
forced in the usual manner to total 1200 for the
calendar year, after which adjusted irregulars are
calculated by dividing the adjusted forced sea­
sonals (S ) into the S I values.
This repartitioning of the S I ratios removes the
large irregular variation from the seasonal com­
u The credence factors are computed as follows. First, a test
is made to locate extreme irregulars falling outside of the 3 sigma
limit. These irregulars are replaced by a value of 1.000 for a
new sigma calculation. A moving 61-term standard deviation of
the irregulars is computed for calibrating the irregular associated
with the middle (31st) term. For the 30 terms at the beginning
(end) of the series, the first (last) centered value is used for
the calibration. Each irregular is then standardized by getting
its absolute difference from the mean of the 61 terms used to
secure its standard deviation, and dividing this difference by the
standard deviation.
A preliminary “ credence factor” is assigned to each value,
based on its standardized irregular, as follow s: For a standard­
ized irregular of 1.000 or less, the credence factor is 1.000. For
a standardized irregular of 2.800 or more, the credence factor is
0.000. For a standardized irregular between 1.000 and 2.800
the credence factor is 1.555-.555 times the standardized irregu­
lar. The 2.8 sigma limit makes it extremely unlikely for a
“good” value, not affected by a strike or other such aberration,
to be assigned zero credence. Only one-half of one percent of the
values in the normal distribution lie beyond this limit. On the
other hand, “ bad” values which deserve to be disregarded have
a much higher probability of falling outside the limit.
12 The amount of reduction for each observation is such that
the deviation of its new irregular from 1.000 is the product of the
credence factor and the deviation of its old irregular component.
Thus, each modified original value is the product of the trendcycle and seasonal components developed in the first iteration,
times an irregular which has been scaled down if it exceeded one
standard deviation. The modified original values are used only
to secure an improved 12-month moving average with which to
sta rt the second iteration ; seasonal-irregulars are always cal­
culated from the actual original values.

ponent and puts it in the irregular component
instead. However, the trend-cycle is still contami­
nated by deviant original values, because it was cal­
culated before the credence factors were developed.
The second iteration is designed to provide an
improved 12-month average for use in developing
revised components. In preparation for this, the
adjusted irregulars of the first iteration are used
to develop revised (intermediate) credence factors.
These intermediate credence factors are computed
from the adjusted irregulars in exactly the same
way (described in footnote 11) as the preliminary
credence factors are obtained from the earlier sec­
ond phase irregulars.
The first use of the intermediate credence fac­
tors is to create a modified original series having
no large irregularities. Each original value hav­
ing a credence factor less than 1.000 is replaced by
a modified value in which the irregularity has been
reduced.12 The creation of the modified original
series marks the end of the first iteration.

Second Iteration
First Phase. The second iteration is similar to
the first iteration, except that the intermediate
credence factors are used along with the weight
patterns of footnotes 7 and 10. The second itera­
tion, like the first, has two phases. The first phase
begins with the centered 12-month moving average
of the modified original series previously de­
scribed. S I ratios are secured by dividing the
actual original series (not modified) by this 12month average. The S I ratios are arrayed by
month and moving 7-term averages, using the
weights of footnote 7 and the intermediate cre­
dence factors, are taken to yield unforced sea­
sonals (S '). (The credence factors prevent ex­
treme observations from affecting the seasonals.)
The forcing process is then applied to yield sea­
sonals (S ) which average 100.0 for the calendar
year. Irregulars (I) are secured by dividing the
latest S I ratios by their corresponding forced
seasonals (S ). These irregulars may include some
residual trend-cycle because of the failure of the
12-month moving average to fully penetrate the
peaks and troughs of the modified original series.
A trend-cycle correction is computed by arraying
the irregulars in normal time sequence and taking
moving 9-term averages, using the weights of foot-


note 10 and the intermediate credence factors.
This completes the first phase of the second
Second Phase. The second phase of the second
iteration starts with an improved trend-cycle,
which is the product of the centered 12-month
moving average of the modified originals and the
trend-cycle correction previously described. The
second phase repeats the steps and procedures of
the first phase to develop new S I ratios, new un­
forced seasonals (S ') making use of the inter­
mediate credence factors, new forced seasonals
( S ) , and new irregulars (I).
These second phase irregulars are used to calcu­
late final credence factors in the same way as be­
fore (see footnote 11). Then a new modified orig­
inal series is secured in the same manner as before
(see footnote 12), using the final credence factors.
This completes the second iteration.

Third Iteration
The third iteration carries the refinement proc­
ess still further. It follows the same steps as in
the second iteration, from the centered 12-month
moving average of the newly modified original
series up to the development of the irregular com­
ponent near the end of the second phase. This
completes the partitioning of the series into the
final trend-cycle, seasonal, and irregular com­
ponents.13 As supplementary information to aid
18 Additional iterations yield little further modification. The
decision to stop with 3 iterations was based on the very small
changes occurring after the third iteration, the reasonable fit of
the trend-cycle to the original data, and the cost of additional


in analysis and evaluation, the final irregular is
standardized using 61-term moving sigmas, which
are printed also. The seasonally adjusted series
is calculated by dividing the original series (O)
by the final seasonal factors (S ).

Cautionary Notes
In applying the method, the user should be
aware that the result of combining series which
have been adjusted separately will usually be a
little different from the direct adjustment of the
combined series. For example, the quotient of
seasonally adjusted unemployment divided by sea­
sonally adjusted labor force will not be quite the
same as when the unemployment rate is adjusted
directly. Similarly, the sum of seasonally ad­
justed unemployment and seasonally adjusted
employment will not quite match the directly
adjusted labor force. Separate adjustment of
components is usually preferable if their seasonal
patterns are different, provided the increased
measurement errors in the components are not
excessive and that the amount of work does not
proliferate unduly.
Finally, it is worth noting that the availability
of a fast, efficient procedure for making seasonal
computations can easily lead to the processing of
large numbers of series without allotting enough
time to review the results. No standard proce­
dure can take the place of careful review and eval­
uation by a skilled analyst. A subjective review
of all results is strongly recommended. The com­
puter program for applying the B L S method
facilitates such review by providing the needed
materials in a logical and easily used format.



Technical References

1. Barton, H. C., Jr., “ Adjustment for Seasonal Variation,” F ed eral Reserve B u lletin , June 1941, pp.
The classic account of the FR B ratio-to-moving-average method, in which the analyst uses
his skilled judgment to draw freehand curves at key stages of the procedure.
2. Macaulay, Frederick R., The Sm oothing of Tim e S e r ie s . New York, National Bureau of Economic
Research, 1931. Publication of N B E R No. 19.
An early discussion of moving averages and of the criteria for choosing one average rather
than another.
3. Organization for Economic Co-operation and Development, S e a so n a l A d ju stm e n t on Electronic
C om puters . Paris, 1961.
The report and proceedings of an international conference held in November 1960. Describes
experience in the United States, Canada, and several European countries. Includes theoretical
sections relating to calendar (trading day) variation and general properties of moving averages.
4. Shiskin, Julius, Electronic Com puters an d B u sin e ss In d icato rs. New York, National Bureau of
Economic Research, 1957. Occasional Paper No. 57. Also published in J o u r n a l o f B u sin e ss,
Vol. 30, October 1957, pp. 219-267.
Describes applications of the first widely used computer program for making seasonal
5. U.S. Department of Labor, Bureau of Labor Statistics, The B L S Method fo r D eriving S e a so n a l
F actors. Paper presented at the Interstate Conference on Labor Statistics, Newport, R.I.,
June 16, 1960.
6. ----- . The B L S S e a so n a l F actor M ethod. Paper presented at the 1960 Annual Meeting of the
American Statistical Association, Palo Alto, Calif., Aug. 23, 1960.
7. ----- . New S e a so n a l A djustm ent F actors fo r L ab or Force Com ponents. Special Labor Force Report
No. 8, 1960. Includes article by Morton S. Raff and Robert L. Stein published in the M onthly
L ab or Review , August 1960, pp. 822-827.
8. ----- . The B u re au of Labor S tatistics S e a so n a l F actor Method. Paper presented at the Interstate
Conference on Labor Statistics, Atlantic City, N.J., June 21, 1962.
9. ----- . The B u re au of Labor S tatistics S e a so n a l F actor Method (1963 Revision ). Paper presented at
the Interstate Conference on Labor Statistics, San Francisco, Calif., June 27, 1963.
10. ----- . The B L S S e a so n a l Factor Method , I t s A pp lication by a n Electronic Com puter. June 1963.
11. ----- . The B L S S e a so n a l Factor Method (1964) • June 1964.
12. ----- . The B L S S e a so n a l F ac to r Method (1 9 6 6 ). May 1966.

—M orton S. R a ff


A b e R othman

Office of the Chief Statistician

Appendix B. Industrial Classification
Much of the usefulness of B L S statistics pre­
sented by industries is due to the fact that they can
be compared with other types of data for the same
industries. This is possible because B L S and
other Federal and State agencies follow as closely
as possible a single system to define and classify in­
dustries in the U.S. economy. The government
publishes a Standard Industrial Classification
Manual (S IC ) of industries based on principles
set forth by a technical group made up of govern­
ment and industry experts.1 The Bureau of L a ­
bor Statistics took part in the development of the
SIC over a long period of years and continues to
work actively with the Bureau of the Budget and
other agencies in seeking to improve the system.
Four basic principles were followed in develop­
ing the classification: 2
(1) The Classification should conform to the
existing structure of American industry. (2) The
reporting units to be classified are establishments,
rather than legal entities or companies. (3) Each
establishment is to be classified according to its
major activity. (4) To be recognized as an in­
dustry, each group of establishments must have
significance from the standpoint of the number
of persons employed, volume of business, and other
important economic features, such as the number
of establishments.
As there are thousands of products and activi­
ties, the system provides for grouping these into
categories, both narrow and broad, to enhance the
value of industrial statistics for users interested
in different levels of detail.3
Using the S IC as a guide, the Bureau classi­
fies the reports received from each factory, shop
or store according to major product or activity.
The S IC is used in the same way by the agencies
supplying the Bureau with its universe lists and
benchmark data. Hence, a high degree of orderli­
ness and consistency is attained, which benefits not
only the users of all B L S establishment statistics,
but also the users of all Government figures.
Certain operational problems make it imprac­
ticable, however, to secure complete uniformity by
this process. Also, specific modes of applying the
SIC differ from one statistical program to another.
For example, there may be differences in the way
in which “major activity” is determined; or

changes in the major activity of individual estab­
lishments which occur over time may be handled
in statistical time series in different ways. Conse­
quently, the use of the same manual and following
a common set of principles of application do not
always result in identical industry classifications of
a given establishment by all agencies, or even by
all programs within B L S . Therefore, any major**
deviations from the normal method of handling
industrial classification will be described in the
chapters on B L S establishment statistics, such as
those on employment, work injuries, and the like.
The standard definition of establishment is
stated as follows:
An ‘establishment’ is an economic unit which produces
goods or services—for example, a farm, a mine, a factory,
a store. In most instances, the establishment is at a
single physical location; and it is engaged in only one, or
predominately one, type of economic activity for which
an industry code is applicable.
Where a single physical location encompasses two or
more distinct and separate economic activities for which
different industrial classification codes seem applicable,
such activities should be treated as separate establish­
ments and classified in separate industries, provided it is
determined th at: (1) such activities are not ordinarily
associated with one another at common physical locations;
(2) no one industry description in the Standard Indus1 U.S. Bureau of the Budget, Standard Industrial Classification
Manual, 1957.
a Ibid, p. 2.
* The SIC provides for different levels of aggregation. The
broadest level divides the economy into ten D ivisions: A. Agri­
culture, forestry and fisheries; B. M ining; C. Contract construc­
tion ; D. M anufacturing; E. Transportation, communications,
electric, gas, and sanitary services; P. Wholesale and Retail
T rad e ; G. Finance, insurance, and real e sta te ; H. Services; I.
Government; and J . Nonclassifiable establishments. At the
2-digit level all products and services are combined into 79
“major groups.” Thus, in the Manufacturing Division, estab­
lishments engaged in manufacturing machinery, apparatus and
supplies for the generation, storage, transmission, transform a­
tion, and utilization of electrical energy are combined into Major
Group 36, “ Electrical Machinery, equipment and supplies.”
The 3-digit level provides several hundred categories. In the
electrical machinery m ajor group the SIC provides 8 groups of
in dustries: 361. Electric transmission and distribution equip­
m ent; 362. Electrical industrial ap paratu s; 363. Household
appliances; 365. Electric lighting and wiring equipment; 365.
Radio and television receiving sets, except communication ty p es;
366. Communication equipment; 367. Electronic components
and accessories; 369. Miscellaneous electrical machinery, equip­
ment and supplies.
Thousands of products and activities are distinguished a t the
4-digit level. For example, in Group 367, 4 industries are de­
fined : 3671. Radio and television receiving type electron tubes,
except cathode ra y ; 3672. Cathode ray picture tubes; 3673.
Transmitting, industrial and special purpose electron tu b es; 3679.
Electronic components and accessories, not elsewhere classified.




trial Classification includes such combined activities; (8)
the employment in each such economic activity is signifi­
cant; and (4) reports can be prepared on the number of
employees, their wages and salaries, and other establish­
ment type data. An establishment is not necessarily
identical with the business concern or firm, which may
consist of one or more establishments. Also, it is to be
^ ibid.,



distinguished from organizational subunits, departments,
or divisions within an establishment. Supplemental in­
terpretations of the definition of an establishment are
included in the industry descriptions of the Standard
Industrial Classification.4
— R obert B . S t e f f e s
Office of Manpower and Employment Statistics

Appendix C. Geographic Classification
United States and States
Since the publication of the previous edition, two
States have been added to the Union. All statis­
tical series of the Bureau were expanded in 1959
to include Alaska and Hawaii. Due to the rela­
tively small populations of these new States, the
effect on national time series was small. Where
statistics are published by States, data for these
States and other small States are shown where
reliable data are available.
Generally speaking, B L S assigns an establish­
ment to the State in which its employees are re­
ported for payroll tax purposes, since these sources
are used both for deriving benchmark levels and
for drawing samples.

Standard Metropolitan Statistical Areas
Some of the Bureau’s data are presented not only
on a national and State basis, but also for selected
metropolitan areas. Comparability of area statis­
tics among Federal agencies is maintained by
means of a set of published standard definitions.1
Defining metropolitan areas for statistical pur­
poses is done on the basis of information about
population and about economic and social ties
among cities. To qualify as a Standard Metro­
politan Statistical Area under current rules, an
area must have at least one city with 50,000 or
more inhabitants, or two contiguous cities meeting
certain other criteria.2 Which counties are to be
included in the standard area is determined on the
basis of criteria relating to metropolitan character
and extent of economic and social communication
among counties.3
The Bureau of Labor Statistics is represented on
the Federal Committee on Standard Metropolitan

Statistical Areas, and plays an active role in the
Technical Committee on Area Definitions. Both
interagency groups are under the chairmanship of
the Bureau of the Budget.

State Economic Areas
State economic areas are relatively homogeneous
subdivisions of States developed by the U.S. De­
partment of Commerce’s Bureau of the Census, in
cooperation with the U.S. Department of Agricul­
ture’s Bureau of Agricultural Economics, and sev­
eral State and private agencies. They consist of
single counties or groups of counties which have
similar economic and social characteristics. The
boundaries of these areas have been drawn in such
a way that each State is subdivided into relatively
few parts, with each part having certain significant
characteristics which distinguish it from adjoining
— R obert B. S t e f f e s
Office of Manpower and Employment Statistics

1 U.S. Bureau of the Budget, Standard Metropolitan Statistical
Areas, 1964.
2 The population criteria a r e :
1. Each standard metropolitan statistical area must include
at le a s t :
(a) One city with 50,000 or more inhabitants, or
(b) Two cities having contiguous boundaries and constitut­
ing, for general economic and social purposes, a single community
with a combined population of at least 50,000, the smaller of
which must have a population of at least 15,000.
2. If two or more adjacent counties each have a city of 50,000
inhabitants or more (or twin cities under 1 (b )) and the cities
are within 20 miles of each other (city limits to city lim its),
they wll be included in the same area unless there is definite
evidence that the two cities are not economically and socially
3 Ibid., pp. 1-2.
4 U.S. Bureau of the Census, Census of Population 1960. Re­
port F C (1) 1A. Number of Inhabitants.
U.S. Summary,



Accessions, definition, 35
Accidents. See Injury Rates and Accident Causes; Work
Advisory Committee on Education, 45
Advisory Committee on Employment Statistics (President
Hoover), 15
Advisory Committee on Standard Budget Research, 68
Agreements. See Collective Bargaining Agreements
Agriculture, U.S. Department of, 54, 96,108,183, 205, 221,
Air Force, U.S. Department of the, 52
Aldrich Report. See W h o lesale P r ic e s , W ages, a n d T r a n s ­
p o rta tio n

American Bankers’ Association, 189
American Gas Association, 203
American Hospital Association, 24
American Public Health Association, 64
American Recommended Practice for Compiling Industrial
Accident Causes, 198
American Standard Method of Recording and Measuring
Work Injury Experience, 198, 208, 212
American Standards Association, 198,211, 213
A n v il an d th e P low , The, 2
Apprenticeship and Training, U.S. Bureau of, 47
Atomic Energy Commission, U.S., 159
Automobile Manufacturers’ Association, 203
Barton, H. C., Jr., 228
Bauman, Alvin, 135
Blake, Roland P., 207
Bloch, Joseph W., 163,166,168
B L S Seasonal Factor Method, 222 ; historical background,
222; characteristics, 222; basic approach, 224; proce­
dure, 225 ; limitations, 227 ; computer program, 223 ;
credence factors, 226 ; weight patterns, 225 ; publications,


BOASI. See Social Security Administration.
Brackett, Jean O., 67, 68
Brady, Dorothy S., 61
Bregger, John E., 14
Bresnahan, Maurice F., 207, 211,216
Budget, U.S. Bureau of the, 4,16, 69,107,114,117,126,160,
189, 220, 229,231. See also Standard Industrial Classifi­
cation, Standard Metropolitan Statisical Areas.
Budget for a Retired Couple, 63
Budgets, standard. See Standard Budgets.
Building trades, wages. See Union Wage Scales.
Bureau of Employees’ Compensation, U.S., 199
Business and Defense Services Administration, U.S., 189
Business Research Advisory Council, BLS, 4

Census, U.S. Bureau of the—population projections, 50,51;
Current Population Survey, 6, 51; seasonal adjustment
procedure, 222; State economic areas, 231, occupational
classification, 70; production index, 108; authority to re­
quire reports, 2; expenditures on plant and equipment,
189; See also Census of—
Census of—State and local governments, 24, 42
Population, 8, 46, 50, 52, 58, 59, 72, 231
Business, 15, 46
Manufacturers, 15, 31, 46, 47, 96,108,156,180,182,195
Mineral Industries, 96,108,183
Chamber of Commerce, U.S., 61
Chamberlain, Neil W., 163
City Workers’ Family Budget, 63
Civil Service Commission, U.S., 24,46,52,117
Clague, Ewan, III, 87,104,110
Clorety, Joseph A., 67
Cohany, Harry P., 174
Cohen, Samuel E., 125
Collection methods. See individual subjects
Collective bargaining agreements, 165; historical back­
ground, 165; concepts, 165; scope, 165; collection and
analysis, 166; uses, 167; limitations, 167; technical
references, 168. See also current wage developments
Commerce, Department of—trade statistics, 96, 221; See
also Census, Bureau of the
Committee on Government Statistics and Information
Services (COGSIS), 1
Commodity classification, 93,108
Community wage surveys, 115
Computers, inventory of, 189
Confidentiality of reports, 2
Construction labor requirements, 194; historical back­
ground, 194; description, 194; data sources and col­
lection methods, 194; sampling and estimating
procedures, 195; analysis and presentation, 196; uses
and limitations, 196; off-site labor, 194
Consumer expenditure surveys, 54; historical background,
54; description, 54; field organization, 55; question­
naires, 55; sampling, 56; analysis and presentation, 60;
uses, 60 ; limitations, 60; technical references, 61; Ad­
visory Committee, 60
Consumer price index, 69; historical background, 69;
description, 70; concepts, 70; weighting structure, 70;
data sources, 72; collection methods, 72; sampling, 75;
city sample, 75; consumer unit samples, 76; item
samples, 76; outlet sampling, 78; calculation procedures,
82; base period, 84; imputation procedures, 84; average
prices, 84; item indexes, 84; analysis and presentation,



Consumer price index—Continued
85; uses, 85; limitations, 85; taxes, 7 0 ; market basket,
7 0 ; seasonal factors, 82; quality change, 7 4 ,7 6 ; technical
references, 87
Consumer Purchases, Study of, 54
Cooper, Sophia, 51
Copeland, William L., 44
Council of Economic Advisers, U.S., 220
Council of S tate Governments, 221
County Business Patterns, 156
“Coverage ratio,” definition, 182
CPI. See Consumer Price Index
Cunningham, Francis S., 85, 87,104
Current Population Survey, 6, 51, 176; See also Labor
Force, Employment, and Unemployment
Current Wage Developments, 141; historical background,
141; description, 141; data sources and collection
methods, 142; sampling and estimating procedures, 142;
presentation and analysis, 143; uses and limitations,
144; collective bargaining agreements, 167.
Daugherty, Jam es C., 87
Davis-Bacon Act, 195
Defense, U.S. Department of, 6, 114; See also Air Force,
U.S. Department of the,
Definitions, standard, 4
Diatchenko, V., 186
Disabling Injury, definition, 198
Discharges, definition, 35
Douty, H. M., 125
Dragonette, Joseph E., 186
Dunlop, John T., 186,192
Earnings. See Earnings and hours; occupational pay;
union wage scales ; industry employment, hours, and
earnings; current wage developments; employer ex­
penditures for supplementary compensation practices
Earnings and hours, frequency distributions of, 1 2 6 ;
historical background, 1 2 6 ; description, 1 2 6 ; data
sources, 1 2 6 ; collection methods, 1 2 7 ; sampling, 1 2 7 ;
estimating procedures, 134; analysis and presentation,
134; uses, 135; limitations, 135; technical references,
1 3 5 ; questionnaire, 128
Economic areas. See State Economic Areas
Economic growth studies, 220
Economists, surveys of, 41
Education, U.S. Office of, 24,47
Eligibility for Union Membership, 173
Employee benefit plans. See Collective Bargaining
Employee earnings. See earnings and hours; wages
Employer expenditures for supplementary compensation
practices, 1 4 5 ; historical background, 1 4 5 ; description,
146; payroll hours, 146; establishment policies, 147;
data sources, 14 7 ; collection methods, 1 4 7 ; sampling,
1 4 7 ; estimating procedures, 1 5 6 ; analysis and presenta­
tion, 1 5 6 ; uses, 1 5 7 ; limitations, 1 5 7 ; technical refer­
ences, 157
Employment. See Labor force, employment and unem­
ployment; industry employment, hours and earnings;

labor turnover; scientific and technical personnel; oc­
cupational outlook; projections of the labor force; in­
dustry-occupational m atrix; technological change; con­
struction labor requirements; economic growth studies;
foreign labor conditions
Employment, definition of, 7
Employment Security, Bureau of, U.S. Department of
Labor, 16, 23, 34, 159
Engineers, surveys of, 41
Establishment, definition of, 4,16, 229
Estimating methods, BLS. See individual subjects
Evans, W. Duane, 95,104,107,110,186
Fabricant, Solomon, 179
Fair Labor Standards Act, 114,126,135
Federal Aviation Agency, U.S., 46
Federal Glassification Act, 115
Federal Committee on Standard Metropolitan Statistical
Areas. See Technical Committee on Area Definitions
Federal government employment, 16,24
Federal Mediation and Conciliation Service, U.S. Depart­
ment of Labor, 144,159,165
Federal Reserve Board, U.S., 47
Federal statistical programs. See Budget, U.S. Bureau
of th e; Committee on Government Statistics and Infor­
mation Services; advisory committees; President’s com­
mittees ; Price statistics review committee; and names
of Federal agencies
Ferber, Robert, 61
Fish and Wildlife Service, Department of the Interior
Fisheries, Bureau of, 96
Food and Nutrition Board. See National Research Coun­
Force-account construction, 17
Foreign Labor Conditions, 217; description, 217; data
sources, 217; collection methods, 217; analysis and uses,
Foreign Service, 218
Fringe Benefits. See Employer Expenditures for Sup­
plementary Compensation Practices, Occupational Pay
and Supplementary Benefits
Fulton, Joseph, 46
Geographic Classification, 231. See also Standard Metro­
politan Statistical Areas
George Washington University, 221
Goldman, Morris R., 110,196
Goldstein, Sidney, 23
Goodman, Roe, 75
Gordon Committee. See President’s Committee to Ap­
praise Employment and Unemployment Statistics
Greenberg, Leon, 186
Greenspan, Harry, 53
Grimaldi, J. V., 207
Gross National Product (GNP), 175
Hald, Helen F., 105,113
Hartman, P. T., 163
Harvard Economic Research Project, 221


Hazards, Industrial. See Work Injuries
Health-Service Occupations, Surveys of, 41, 43
Hires, new, Definition, 35
History of BLS Programs. See individual subjects
Hoffenberg, Marvin, 95,104,107,110
Hoover, Ethel D., 87
House, Frequency Distributions of Employees. See Earn­
ings and Hours
Housing Unit Survey, Comprehensive, 58, 76; Rural, 58
Humes, Helen. See Lamale, Helen H.
Indexes. See Consumer Prices, Wholesale Prices, Produc­
tivity, etc.
Industrial Classification of Reports. See Standard Indus­
trial Classification (SIC)
Industrial Relations. See Work Stoppages; Collective
Bargaining Agreements; Union Membership; and topics
under “Wages”
Industry Employment, Hours and Earnings, 1 5 ; histori­
cal background, 15; description, 1 6 ; concepts, 1 6 ; data
sources, 2 3 ; benchmark data, 2 3 ; collection methods, 2 4 ;
“shuttle” schedule, 2 4 ; sampling, 2 5 ; estimating proce­
dures, 2 6 ; hours and earnings, 2 7 ; spendable earnings,
2 9 ; seasonal adjustments, 2 9 ; presentation, 2 9 ; uses, 3 1 ;
limitations, 3 1 ; technical references, 33
Industry Wage Surveys, 114
Industry-Occupational Matrix, 52; historical background,
52; data sources, 52; analysis and uses, 53
Industry-Sector Indexes, 106; description, 106; concepts,
106; universe, 107; prices and base period, 107; sam­
pling, 107; weights, 108; calculation, 108; analysis and
presentation, 108; uses, 108; limitations, 109; technical
references, 110; classification, 107
Injury Rates and Accident Causes, 212; historical back­
ground, 212; description, 212; accident cause data, 212;
data sources, 213; collection methods, 213; sampling,
214; estimating procedures, 214; analysis and presenta­
tion, 214; uses, 215; limitations, 215; See also Work
Input-Output tables, 106,109, 220
In terin du stry relationships, 106, 109, 196, 220

Internal Revenue Service, 60,183,192
International Association of Industrial Accident Boards
and Commissions (IAIABC), 198
International Comparisons, 218
International Labour Office, 61,163,186,192,217
Interstate Commerce Commission, 24,46,183,205
Jaffe, Sidney A., 87
Johnston, Denis F., 51
Kafka, Fritz, 186
Kalish, Carol B., 14
Kanninen, Toivo P., 125
Kendrick, John W., 179,186
Ketterling, Virgil F., 179
Kish, Leslie, 75
Kutscher, Ronald E., 221
Labor and non-labor payments, 178
Labor Developments Abroad, 217


Labor Force, Employment and Unemployment; historical
background, 6; description, 6; concepts, 7; collection
methods, 10; sampling, 8; estimating methods, 11; pres­
entation, 12; uses, 12; limitations, 13; accuracy, 13;
technical references, 14
Labor Force Projections, 50; uses, 50; methodology, 50;
data sources, 51; publications, 51; assumptions, 50
Labor Force Reports, Special, 12
Labor-Management and Welfare-Pension Reports, Office
of, Department of Labor, 142,169
Labor-Management Reporting and Disclosure Act, 169
Labor Monographs, 217
Labor Requirements, Construction. See Construction
Labor Requirements
Labor Research Advisory Council, BLS, 4
Labor Standards, U.S. Bureau of, 208,211
Labor Turnover, 34; historical background, 34; descrip­
tion, 34; concepts, 34; data sources, 35; collection
methods, 35; sampling, 38; estimating procedures, 38;
seasonal adjustments, 39; presentation, 30; uses, and
limitations, 39; questionnaire, 36
Laboratories, industrial research, 43
Laciskey, John F., 140
Lamale, Helen H., 56,61,87, 88
Larsgaard, Olga A., 88
Latimer, Murray Webb, 145
Layoffs, definition, 34
Level of living, 63, 65
Living standards, 63,65
Lockouts. See Work Stoppages
Longshoremen’s and Harbor Workers’ Compensation Act,
McGraw-Hill Company, 189
Macaulay, Frederick R., 228
Machine tools, survey of numerically controlled, 189
Mack, Louise J., 88
Man-hours, establishment, estimates of, 176
Man-hours, construction, definition, 194
Man-hours, indexes, 183
Man-hours, labor force, estimates of, 176
Man-hours, nonproduction workers, 177,184
Manpower Development and Training Act (MDTA), 49,
Manpower requirements, 45. See also Construction labor
Marimont, Martin L., 110,196
Marine cargo handling, 210
Maritime Safety. See Longshoremen’s and Harbor
Workers’ Compensation Act
Mark, Jerome A., 186
Market Prices, Spot. See Spot Market Prices
Mendelssohn, Rudolph C., 25
Metropolitan areas. See Standard Metropolitan Statisti­
cal Areas
Metropolitan Life Insurance Company, 34
Mines, U.S. Bureau of, 96,189,199, 221
Moss, Bennett R., 105, 110
MRLF. See Labor Force, Employment and Unemploy­



Murphy, Kathryn R., 62
Myers, Robert J., I l l
National Academy of Sciences, 43
National Bureau of Economic Research, 69, 95, 106, 179,
National Catholic Welfare Conference, 24
National Health Survey, 65
National Industrial Conference Board, 61
National Labor Relations Board, 145
National Mediation Board, 166
National Planning Association, 48, 221
National Recovery Act, 114
National Research Council, 43
National Research Project, 180
National Safety Council, 205, 207
National Science Foundation, U.S., 41,47,189
National Survey of Professional, Administrative, Techni­
cal and Clerical Pay, 115
National Vocational Guidance Association, 45
Occupational matrix (or structure). See Industry-Oc­
cupational Matrix
Occupational Outlook Program, 45; historical background,
45; uses, 49; description, 46; data sources, 46; analysis,
47; Handbook, 45; Quarterly, 45; presentation, 49
Occupational Pay and Supplementary Benefits, 114; his­
torical background, 114; description, 114; concepts, 116;
survey methods, 117; questionnaires, 117; sampling,
121; estimating procedures, 122; analysis and presenta­
tion, 123; uses and limitations, 123; reliability of sur­
veys, 124; technical references, 125
Occupational projections, 47
Office of Business Economics, U.S. Department of Com­
merce, 107,175,195, 221
Office of Emergency Planning, 52
Onanian, Edward D., 164
Organization for Economic Cooperation and Development
(OECD), 186, 228
Orshansky, Mollie, 67
Output, definitions, 175,181
Output, value of, deflated, 183
Output per man-hour m easures: Industries, 180; histori­
cal background, 180; description, 180; concepts, 180;
methods and sources, 181; benchmark indexes, 182;
comparability of output and man-hours data, 184; presentaton, 185; uses and limitations, 185; technical refer­
ences, 186
Output per man-hour measures: Private economy, 175;
background and description, 1 7 5 ; data sources and esti­
mating procedures, 1 7 5 ; analysis and presentation, 1 7 7 ;
uses, 1 7 7 ; limitations, 1 7 7 ; technical references, 1 7 9 ;
concepts, 175
Overtime man-hours, 23, 28
Paschell, William, 174
PATCP. See National Survey of Professional, Adminis­
trative, Technical and Clerical Pay
Peterson, Florence, 164

Population projections, 48, 50, 220
President’s Advisory Committee on Labor-Management
Policy, 187
President’s committee on the Cost of Living, 69
President’s Committee to Appraise Employment and Un­
employment Statistics, 14
Price Statistics Review Committee, 69,106
Prices and Living Conditions. See Consumer expendi­
tures; standard budgets; consumer prices; wholesale
prices; industry-sector indexes; spot market prices
Prochnow, Lloyd A., 186
Production indexes, 106
Production workers, definition of, 17
Productivity. See Output per man-hour; technological
change; construction labor requirements; economic
growth studies, foreign labor conditions and interna­
tional comparisons
Professions, survey of, 41
Projections. See labor force, population, etc.
Psychologists, surveys of, 41
Public Contracts Act, 114
Public Health Service, U.S., 51
Public Housing Administration, U.S., 64
Public Works Administration (PW A), U.S., 194
Publications, BLS. See individual subjects, especially
technical references at ends of Chapters
Quits, definition, 34
Raff, Morton S., 228
“Real” earnings, 29
Research studies. See individual subjects
Roosevelt, Franklin D., President, 45
Ross, Arthur M., 163
Rothman, Abe, 228
Rothwell, Doris P., 72, 82, 84, 88,90
Ruben, George, 144
Safety, industrial. See work injuries
Sampling. See individual subjects
Schecter, Samuel, 24
Schilling, Jane Metzger, 163
Schiro, Bruno, 87
Scientific and technical personnel surveys, 41; historical
background, 41; description, 41; data sources, 41; col­
lection methods, 42; sampling, 42; estimating proce­
dures, 43; analysis and presentation, 44; uses and
limitations, 44; technical references, 44; State and local
governments, 41
Searle, Allan D., 91,104,105
Seasonal Adjustments. See B L S Seasonal Factor Method
Seasonal Factors—CPI, 82; WPI, 103
Separations, definition, 34
Shelton, William C., 219
Shi skin, Julius, 222, 228
Siegel, Irving H., 182,186
Simonds, R. H., 207
Social Security Administration, U.S., 24, 68,78
“ Specialization ratio”, definition, 182


Spot market prices, daily index of, 111; historical back­
ground, 111; description, 111; data sources, 111; collec­
tion methods, 111; selection of products, 112; estimat­
ing procedures, 112; analysis and presentation, 112;
uses, 112; limitations, 113; specifications, 113
Stallings, Carlyle P., 90
Stambler, Howard, 45,49
Standard budgets, 63; historical background, 63; descrip­
tion, 63; data sources, 64; analysis and presentation,
65; uses, 66; limitations, 66; technical references, 67
Standard Consolidated Areas. See Standard Metropoli­
tan Statistical Areas
Standard Industrial Classification (SIC ), 229; use in
scientific and technical personnel surveys, 42; industrysector indexes, 107; employment, earnings and hours,
23, 126; work injuries, 199; occupational pay, 114;
maritime work injuries, 210; current wage develop­
ments, 142; labor turnover, 35; work stoppages, 160
Standard International Trade Classification, 93
Standard Metropolitan Statistical Areas, 32, 55, 70, 75,
114,126,162, 231
Standard time period for employment reporting, 5,23
Stanford Research Institute, 48
State agencies, cooperation with, 15,34,203
State Economic Areas, 231
Steffes, Robert B., I ll, 230,231
Stewart, Maxine, 46
Stigler, George J., 62
Stotz, Margaret S., 67
Strasser, Arnold, 157
Strikes. See Work Stoppages
Supplementary Benefits. See Occupational Pay and Sup­
plementary Benefits; Employer Expenditures for Sup­
plementary Compensation Practices
Tables of Working Life, 47
Taft-Hartley Act, 163,165
Technical Committee on Area Definitions, 231
Technicians, surveys of, 41
Technological Change, 187; historical background for
studies, 187; description of studies, 187; data sources
and collection methods, 188; analysis and interpreta­
tion, 190; definition, 190; impact on productivity, 190;
impact on employment, 190; impact on occupations, 191;
adjustments to, 191; uses and limitations of studies,
191; technical references, 192; industry trends, 187
Trade Associations, 183, 198. See also names of Associa­
Troy, Leo, 174
Turnover, Labor. See Labor Turnover
Unemployment, definition of, 7
Unemployment Insurance Program: Benchmarks, 16, 23,
42; Listings, 127,142,147
Unemployment surveys. See Labor Force, Employment,
and Unemployment
Union agreements. See collective bargaining agreements,
union wage scales, current wage developments
Union membership, 169; historical background, 169; data
sources and collection, 169; presentation, 173; uses and


limitations, 173; technical references, 174; question­
naire, 170
Union Wage Scales, 136; historical background, 136; de­
scription, 136; data sources and collection methods, 136;
sampling and estimating procedures, 137; indexes, 137;
analysis and presentation, 140; uses, 140; limitations,
140; questionnaire, 138
Unions, labor. See specific subjects, such as collective
bargaining agreements, Taft-Hartley Act, Work Stop­
pages, Union membership
Unit labor cost, 218; See also output per man-hour
United Nations Economic and Social Council, 106
United Nations Statistical Office, 62,186
U.S. Government Organizations and Agencies. See name
of organization
Vaccara, Beatrice N., 110,196
Veterans Administration, U.S., 45
Vital Statistics, Division of. See Public Health Service,
Voluntary Reporting, 2
Wage Boards, 114
Wage scales, union. See Union Wage Scales
Wage Stabilization Board, U.S., 141
Wages. See occupational p ay ; employee earnings; union
wage scales; current wage developments; industry em­
ployment, hours, and earnings; labor and nonlabor
Wages and Hours and Public Contracts Divisions, U.S.
Department of Labor, 135
Walker, Francis A., Gen., 1
Walker, James F., 196
War Industries Board, U.S., 114
War Labor Board, U.S., 145
Welfare and Pension Plans Disclosure Act, 145
Weinberg, Edgar, 31,193
Wholesale prices, 91; historical background, 91; descrip­
tion, 91; universe, 92; concepts, 91; classification, 93;
data sources, 95; collection methods, 95; weights, 94,
97; sampling, 96; calculation, 97; base period, 97; im­
putation procedures, 98; weekly index, 98; analysis and
presentation, 98; uses, 103; limitations, 103; technical
references, 104; quality change, 93; special indexes, 93;
questionnaires, 99
W h olesale P r ic e s , W ag es a n d T ra n sp o rta tio n (Aldrich
Report), 91
Wilkerson, Marvin, 70, 75,89
Williams, Faith M., 62
Wolman, Leo, 174
Work Injuries, Federal government, 200
Work Injuries, Frequency and Severity, 197; historical
background, 197; description, 198; standardization,
198; definitions, 198; measures, 198; rates, 198; data
sources, 200; collection methods, 200; sampling, 203;
estimating procedures, 204; analysis and presentation,
204; uses, 204; annual estimates, 205; limitations, 206;
technical references, 207; questionnaires, 201; accident
causes, 212



Work Injuries, Maritime, 208; historical background, 208;
description, 208; data sources, 209; estimating proce­
dures, 209; uses, 210; limitations, 210; technical refer­
ences, 211
Work Stoppages, 158; historical background, 158; de­
scription, 158; data sources, 159; collection methods,
159; estimating procedures, 100; analysis and inter­
pretation, 160; presentation, 162; uses and limitations,
163; technical references, 164; questionnaire, 161
Working life tables. See Tables of working life

Workmen’s Compensation Statistics, 204
Works Progress Administration, U.S., Standard budgets,
63; Monthly Report on Employment, 6; output per man­
hour, 180
Wymer, John P., 33, 40
Wright, Carroll D., 1,180
Yabroff, Bernard, 46
Young, Dudley E., 23
Zimmerman, Carle C., 62