View original document

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

L

3l . "S

*
1




1

SOUTHWEST MISSOURI STATE
COLLEGE LIBRARY
5 -

d e p o s it o r y

COpv

MAY 2 5 1972




BLS
Handbook
of Methods
for Surveys
and Studies
BULLETIN 1711

U. S. DEPARTMENT OF LABOR
J.D. Hodgson, Secretary
BUREAU OF LABOR STATISTICS
Geoffrey H. Moore, Commissioner
1971

For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 - Price $2.00
Stock Number 2901-0659




Preface

The Bureau of Labor Statistics is essentially a factfinding agency, the functions
of which are the collection, tabulation, evaluation, and publication of a wide range
of economic and statistical information. A part of the BLS’ responsibility is
to make available to the users of its data information on the scope of its pro­
grams and the methods employed in the surveys and studies that it conducts.
The users of BLS information encompass a broad spectrum of the American
society— administrators of businesses, large and small; workers; union officials;
academicians; technicians; government policymakers and administrators at Fed­
eral, State, and local levels; and others o f the general public.
The methods employed by the Bureau in collecting, analyzing, and presenting its
data are often highly technical and complex. A description of these methods at a
level of technical detail which would meet the requirements of all users might bore
the technician or bewilder the nontechnical user.
The Handbook of Methods is designed to serve the broad middle range of users.
The technician may wish to seek more detailed treatment of his field of interest,
and the casual user may find simpler explanations more helpful. Sources of addi­
tional information, some more technical and some more popular, are listed under
“ Technical References” at the end of most chapters.
The chapters for the Handbook were prepared in the program offices of the Bu­
reau and published by the Office of Publications, Division of Special Publications,
under the direction of Tommy C. Ishee with the assistance of Scott Wirtzman.




hi




Contents
Page

Introduction ______________________________ __ ___________________

___

1

Chapter:
1. Labor force, employment, and unemployment__________________
2. Employment, hours, and earnings ____________________________
3. Job vacancies and labor turnover_________________

7
17
35

Current Employment Analysis

Manpower Structure and Trends
4.
5.
6.
7.

Employment of scientific and technical personnel_______________
Occupational outlook _________________________________________
Projections of the labor fo r c e __________________________________
Industry-occupational matrix __________________________________

43
47
53
55

Prices and living conditions
8.
9.
10.
11.
12.
13.

Consumer expenditures and incom e____________________________
Family budgets________________________________________________
Consumer prices ______________________________________________
Wholesale prices __________
Industry-sector indexes _______________________________________
Spot market prices ____________________________________________

59
69
75
97
113
119

Wages and industrial relations
14.
15.
16.
17.
18.
19.
20.
21.
22.

Occupational pay and supplementary benefits __________________
Employee earnings and hours frequency distributions__________
Union wage rates --------Current wage developments ___________________________________
Employer expenditures for employee compensation _____________
Work stoppages___________ _________________ ____ ___ _______—
Collective bargaining agreements -------------- ---------------- ------------Union membership __________________________________ __ ______
Annual earnings and employment patterns of private nonagricultural workers _________________ ,------------------------- ---------------23. Measuring collective bargaining settlements ___________________
24. Wage chronologies and salary trend reports ------------------------------

123
137
147
153
161
171
181
187
195
201
209

Productivity and technology
25.
26.
27.
28.

Output per man-hour: Private sector______ ___________________
213
Output per man-hour measures: Industries_________________ -__
219
Technological change _____ ________________________ __ ______ ______________
Construction labor requirements ______________ ___ ____________
235

227

Occupational safety and health
29. Occupational safety and health statistics ___ ___ _______________




239

v

Contents— Continued
P ag *

Economic Trends and Labor Conditions
30. Foreign labor conditions, international comparisons, and trade
research __
31. Economic growth studies _____________________________________
Appendix: A. The BLS seasonal factor method _______
B. Industrial classification________________________________________
C. Geographic classification ______________________________________

TI




241
245
247
255
257

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

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 U.S.
Commissioner of Labor, he issued his first an­
nual report in March 1886, Carroll Wright
established the policy of explaining his sta­
tistical methods to his readers and of seeking
to avoid misinterpretation of the figures pre­
sented. 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
capable of varied application— some even are of doubt­
ful 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 indicate that condition of a locality, state, or
country when more goods have been produced than are
sufficient to meet the ordinary demand . . . .*

In the same report there are statements on
testing the validity of figures (p. 141), prob­
lems of nonresponse (p. 90), and restrictions
on coverage (headnotes to tables). Warnings
as to inadequacies of available information
occur frequently. During the 86 years 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 ex­
plained 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 ob­




served. Furthermore, whereas one might ex­
pect to breed a certain amount of doubt about
a statistical survey by revealing its lack of
perfection, frankness about unavoidable de­
fects 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 ex­
plained.
The Committee on Government Statistics and
Information Services emphasized 30 years ago
that the Central Statistical Board “ should urge
on each collecting agency the importance of
publication 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.” 1
2
Full understanding of the statistical series
and studies of the Bureau of Labor Statistics
is not to be gained solely from detailed descrip­
tions of them, but also from appraisal of the
philosophy and approach of the Bureau and of
the manner in which it functions.

Background

The history of the Federal Bureau of Labor
extends 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
1 See Industrial Depressions, The First Annual Report of the
Commissioner of Labor, March 1886, pp. 15-16.
2 Committee on Government Statistics and Information
Services, Government Statistics, April 1937, p. 53. Recommen­
dation #16 of the Committee states: “ Continued criticism and
analysis should be made of (a) statistical definitions, specifi­
cations, and classification; (b) coverage of supposedly com­
plete 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).

1

2

BLS HANDBOOK OF METHODS

Labor. From the Bureau’s beginnings in the
administration of President Arthur until it
became a part of a cabinet department under
President Wilson, it accumulated nearly 3 dec­
ades of experience in collecting, interpreting,
and presenting facts crucial to the welfare of
workers. Details of early Bureau history and
of developments of later years are to be found
elsewhere.3 However, in describing the various
statistical programs4 in this volume, some of
the events which led to the development of
particular statistical measures are recounted.
Against this historical background emerges the
philosophy and posture of the Bureau of Labor
Statistics as the impartial observer and in­
terpreter of trends important to the welfare
of workers. Voluntary reporting and the pre­
serving of the confidential nature of reported
data are important characteristics of BLS
programs.

Voluntary Reporting and
Confidentiality

In the 86-year history of the Bureau’s opera­
tion 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 sup­

plied the desired information, 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 infor­
mation. The response has been remarkable 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 per­
son supplying the information. The fact that
Bureau employees pledge themselves to protect
these data is less important than that they have
a deep understanding 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 suc­
cess of the survey and that he may find the
survey results useful in his own pursuits. The
policy of not identifying the respondent is im­
plemented 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 pur­
poses 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 success­
fully resisted.5 Another 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 sta­
tistical agency.0
While it cannot be proved that these policies
result in more reliable statistics, Bureau Com­
missioners and their staffs over the years have
been convinced from their experience that it is
so. It is notable that some other Federal agen­
cies (especially the Bureau of the Census),
well-equipped with authority to compel the
submittal of certain reports, rarely if ever in­
voke this power. Rather, 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 prescribed by law,7
has always relied upon voluntary cooperation
of respondents in collecting information.
3 See, for example, the Secretary of Labor’s First Annual
Report, 1913, 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. 172-173, 187, 206-207, and 230-234 (1949-63).
4 For a handy reference to BLS programs, showing their
principal characteristics, see U.S. Department of Labor,
Bureau of Labor Statistics, Major BLS Programs (issued
annually).
5 For example, see Norweigan 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 District of Columbia. In all of these
cases, the courts sustained the policy of protecting the con­
fidentiality of information given voluntarily and in confidence
to an agency of the Federal Government.
0 See Supreme Court of the United States, St. Regis Paper
Company, Petitioner, v. United States, No. 47, October term,
1961.
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.

INTRODUCTION

BLS Role, Staff, and Organization

Among Federal agencies collecting and issu­
ing 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 public, and especially the admin­
istrative and executive agencies, for infor­
mation on economic and social trends and
situations. While the data serve some admin­
istrative purposes, they are free from the
constraints which sometimes result from the
close ties normally existing between operations
and operational statistics. BLS 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 pos­
sible limitations be well understood.
The Bureau plays a larger and more signifi­
cant role than merely publishing generalpurpose statistics. Its activities frequently in­
fluence, and sometimes are crucial to, the de­
termining and shaping of public policy. The
Bureau’s experts have the keen understanding
of economic and social forces which results
from intensive and continuous involvement in
factfinding and in the painstaking analysis of
data.
Staff

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 eco­
nomics 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 possible in written
and oral form. How successfully this can be
accomplished depends greatly upon the com­
petence of the analysts and of their supporting
personnel.
In BLS, analytical and statistical work is
performed by economists, statisticians, and
s Bureau of the Budget, Statistical Services of the United
States Government (revised edition), 1963, p. 7.




3

mathematical-statisticians with the aid of an
experienced corps of programers, systems
analysts, and other professionals, as well as
statistical clerks. For analytical work, eco­
nomists at even the lowest grade level must
meet Civil Service Commission requirements
roughly equivalent to a college major in eco­
nomics. There are comparable requirements
for other professionals. The greatest effort is
made to locate the best of graduating seniors,
Masters, Ph.D.’s, and those with research ex­
perience, in the colleges, State agencies, busi­
ness 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 execu­
tive professionals in obtaining the best results
from their staffs.
In training staff, a special effort is made to
impart 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.
Organization

The statistical programs of the Bureau were
developed, for the most part, independently of
each other, taking on characteristics suited to
the requirements of the subject under observa­
tion. As a result, the Bureau was organized
according to subject-matter areas, an arrange­
ment which has proved efficient and has been
continued over the years. Expertise in tech­
niques, economic analysis, and other staff ac­
tivities across subject-matter lines were added
to provide better utilization of the Bureau’s
resources.
As the Bureau’s collection activities in­
creased, regional offices were established in
1943 to provide administration of the field pro­
grams 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 cooperating organizations.
An important aspect of the work of the re­
gional staffs has been the function of explain­
ing the concepts and techniques which we
utilize in compiling our statistics.

4

BLS HANDBOOK OF METHODS

Special recognition of the competence of the
Bureau in the field of statistics was given by
the Secretary of Labor in 1955 when he dele­
gated to the Commissioner of Labor Statistics
the responsibility for continuously reviewing
all of the statistical programs of the Depart­
ment of Labor and of making recommendations
for their improvement.

Consultation and Advice on Statistical
Programs

A statistical program too much detached
from the uses of its data may fail in its prin­
cipal mission. To avoid sterility and stagnation,
the Bureau continuously invites advice and
ideas from users and experts in business, labor,
and academic organizations and individual
members of the public. Over the years, the
advice the Commissioner of Labor Statistics
has received on policy and technical matters
from responsible parties, relating to the col­
lection and analysis of our statistics, has
usually been sound and therefore very helpful.
Of course, decisions on statistical policy have
always been the final responsibility of the
Commissioner.
In order to keep in touch with the current
and anticipated needs of business and labor
groups and to seek advice on technical prob­
lems, the Commissioner first established stand­
ing research advisory committees in 1947.
These groups, now called the Business Re­
search Advisory Council and the Labor Re­
search Advisory Council, serve in an advisory
capacity with respect to technical problems,
consult on Bureau programs, and provide per­
spectives 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 subjectmatter fields on a more specialized basis. Com­
mittee 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 reso­
lutions or recommendations on matters re­
garded as appropriate for such action, but such



resolutions are merely advisory. Members of
the councils and the subcommittees serve in
their individual capacities, not as representa­
tives of their organizations.
The members of the Labor Research Advisory
Council are designated by the Commissioner of
Labor Statistics under authorization by the
Secretary of Labor, from nominations by the
Director of Research, AFL-CIO. All research
directors of international unions represented in
the AFL-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 direc­
tion to the advisory activities of trade union
research directors in relation to the Bureau.
The members of the Business Research Ad­
visory Council are designated by the Commis­
sioner under authorization of the Secretary of
Labor, after consultation with the National
Association of Manufacturers, the U.S. Cham­
ber of Commerce, and other organizations
broadly representative 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 scien­
tists, educators, and others, either in their
individual capacities or as members of profes­
sional organizations. This is most likely to oc­
cur when a conceptual or theoretical question
arises which is considered fundamental to the
work of the Bureau in a specialized field, and
where professional acceptance of the Bureau’s
work in that field may be reinforced by the
findings of an independent analyst.
It is a fundamental objective of the Bureau
that its statistical practice be built soundly
upon established statistical theory. The objec­
tive can be realized only if BLS practitioners
are trained in statistics and if they keep their
knowledge up to date. For this reason the Bu­
reau encourages their participation in activities
of the professional societies, their efforts to
improve their education in statistics, and con­
tinuing contacts with other experts in their
disciplines. The efforts of other statistical or­
ganizations, public and private, are studied
unremittingly so that BLS may reap benefits
from the experiences of others.

INTRODUCTION

Standard Definitions

Where related statistics cut across program
lines or across Government bureaus, the Bu­
reau of Labor Statistics cooperates to the
maximum extent possible in the Office of Man­
agement and Budget’s (formerly, the Bureau of
the Budget) effort to obtain adherence to stand­
ard definitions of terms for maximum com­
parability. The use of the definition of estab­
lishment is a case in point.
The business establishment has been found
to be the most satisfactory source of data for
most industrial statistics because it is the first
level of business organization for which com­
plete records— i.e., production, employment,
purchases, sales, wages, inventories, etc.— are
generally maintained. 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 pro­
duction of a related group of products or
services. In BLS programs in which it is ap­




5

plicable, the standardized definition of the
establishment is utilized.9
The Bureau also follows the Office of Man­
agement and Budget’s definition of “ production
and related workers,” 1 1and uses the “ standard
0
payroll period.” 1 The reader is referred to
1
appendixes B and C for descriptions of the
standards followed with respect to industrial
classification and geographic classification.
9 See appendix B.
1 Standard Definitions of Types of Workers, Bureau of the
0
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, mainte­
nance, repair, janitorial, watchman services, product develop­
ment, auxiliary production for plant’s own use
(e.g.,
powerplant) and record-keeping and other services closely
associated with the above production operations. Excluded
are supervisory employees (above the working foremen
level) and their clerical staffs.”
1 Standard Definition of Payroll Periods for Employment
1
Reports, Bureau of the Budget, March 28, 1952. "In order to
maintain a coordinated system of employment reports and
to reduce the reporting burden on respondents, requests made
to employing 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 containing the 12th of the month.”




Current Employment Analysis

Chapter 1. Labor Force, Employment, and Unemployment
Background

Each month, the Bureau analyzes and pub­
lishes statistics on the labor force, featuring
information on employment, unemployment,
and nonparticipation, classified by a variety of
demographic, social, and economic charac­
teristics. These statistics are derived from the
Current Population Survey, (CPS), which is
conducted by the Bureau of the Census. The
survey is based on a probability sample of
households, representative of the civilian noninstitutional population of the United States.
Concepts of the labor force, employment,
and unemployment were introduced in the lat­
ter stages of the depression of the 1930’s,
chiefly to arrive at more objective measure­
ments of unemployment and employment than
were previously available. Before the 1930’s,
aside from attempts in some of the decennial
censuses, no direct measurements were made
of the number of jobless persons. The develop­
ment of mass unemployment in the early 1930’s
increased the need for statistics, and widely
conflicting estimates based on a variety of in­
direct techniques began to appear. Dissatisfied
with these methods, many research groups, as
well as State and municipal governments, be­
gan 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 individual. In most
of the surveys, the unemployed were defined as
those who were not working but were “ willing
and able to work.” This concept, however, did
not meet the standards of objectivity that many
technicians felt were necessary to measure
either the level of unemployment at a point in
time or changes over periods of time. The cri­




terion “ willing and able to work,” when applied
in specific situations, appeared to be too intan­
gible and too dependent upon the interpretation
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 in­
dividual depends principally on his actual ac­
tivity within a designated time period, i.e., was
he working, looking for work, or engaged in
other activities? These concepts were adopted
for the national sample survey of households
initiated by the Works Progress Administra­
tion (W PA) in 1940.
Originally termed the Monthly Report on
Unemployment when the WPA was responsible
for the collection of labor force statistics, the
household survey was transferred to the Bu­
reau of the Census late in 1943. Its name was
changed at that time to the Monthly Report on
the Labor Force (M RLF). The survey title
was changed once more— in 1948— to the pres­
ent “ Current Population Survey” 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 publish­
ing the CPS labor force data was transferred
to the BLS, although the Census Bureau has
continued to collect and tabulate the statistics.

Description o f Survey

The CPS provides statistics on the civilian
noninstitutional population 16 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 nonin-

7

8

BLS HANDBOOK OF METHODS

stitutional population. Persons under 16 years
of age are excluded from coverage in the sur­
vey because child labor laws, compulsory school
attendance, and general social custom prevent
most of these children in the United States
from working. The institutional population,
which is excluded from coverage, consists of
inmates of penal and mental institutions, tuber­
culosis sanitariums, and homes for the aged,
infirm, and needy.1
The CPS is collected each month from a
probability sample of approximately 47,000
households. Participation in the survey is on a
purely voluntary basis. Respondents are as­
sured that all information obtained is com­
pletely confidential and is used only for the
purpose of estimating national totals.
The time period covered in the monthly sur­
vey 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 obtained is “ current” and the time
reference not so short that the occurrence of
holidays or other accidental events causes ex­
tremely erratic fluctuations in the information
obtained. A calendar week fulfills these condi­
tions as well as being a convenient and easily
defined period of time. Since July 1955 the
week containing the 12th day of the month has
been used. The actual survey is conducted dur­
ing the following week, which is the week con­
taining the 19th day of the month.
Concepts

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 familyoperated enterprise and (2) all those who were
not working but who had jobs or businesses
1 For a fuller explanation of the Current Population Survey,
see Concepts and Methods Used in Manpower Statistics from
the Current Population Survey, U.S. Department of Com­
merce and U.S. Department of Labor (1967), Report 313.




from which they were temporarily absent
because of illness, bad weather, vacation, labormanagement dispute, or various personal rea­
sons. Each employed person is counted only
once. Those who held more than one job are
counted in the job at which they worked the
greatest number of hours during the survey
week. Excluded 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, were looking for work, and were avail­
able for work in the reference period. Those
who had made efforts to find work within the
preceding 4-week period— such as by register­
ing at a public 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 in­
cluded 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 ex­
cept that they were temporarily ill.
Duration of unemployment represents the
length of time (through the current survey
week) during which persons classified as un­
employed had been continuously looking for
work. For persons on layoff, duration of un­
employment represents the number of full
weeks since the termination of their most re­
cent employment. A period of 2 weeks or more
during which a person was employed or ceased
looking for work is considered to break the
continuity of the present period of seeking
work. Average duration is an arithmetic mean
computed from a distribution by single weeks
of unemployment.
Labor Force. The civilian labor force comprises
the total of all civilians classified as employed

LABOR FORCE, EMPLOYMENT, AND UNEMPLOYMENT

and unemployed. 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
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 retired persons, individuals re­
ported as too old or temporarily unable to work
and the voluntarily idle. It also includes sea­
sonal workers for whom the survey week fell in
an “ off” season and who were not reported as
looking for work and persons who did not look
for work because they believed either that no
jobs were available in the area or that no jobs
were available for which they could qualify.
Persons doing only incidental unpaid family
work (less than 15 hours in the specified week)
also are classified as not in the labor force.
The category “ not in labor force— in school”
includes persons attending school during the
survey 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 who were not working or
looking for work) are classified as not in the
labor force.
For persons not in the labor force, questions
are asked about previous work experience, in­
tentions to seek work, desire for a job at the
time of interview, and reasons for not looking
for work. The questions for persons not in the
labor force are asked only in those households
that are new entrants to the sample and in
those that are reentering the sample after 8
months’ absence. (See Sampling.)

Sampling
The Survey Design

The CPS sample is located in 461 sample
areas comprising 923 counties and independent
cities with coverage in every State and the
District of Columbia. In all, about 55,000 hous­



9

ing units and other living quarters are desig­
nated for the sample at any time, of which
about 47,000, containing about 105,000 persons
16 years and over, are occupied by households
eligible for interview. The remainder are units
that are vacant, converted to nonresidential
use, containing persons who reside elsewhere,
or ineligible for other reasons. Of the occupied
units eligible for enumeration, about 3 to 5
percent are not interviewed in a given month
because the residents are not found at home
after repeated calls, are temporarily absent, or
are unavailable for other reasons.2
Selection of Sample Areas. The entire area of
the United States consisting of 3,000 counties
and independent cities is divided into 1,931
primary sampling units. With some minor ex­
ceptions, a primary sampling unit (PSU) con­
sists of a county or a number of contiguous
counties. Each of the 237 standard metropoli­
tan statistical areas (SMSA’s )3 in existence at
the time of the 1970 Census constitutes a
separate PSU. By combining counties to form
PSU’s, greater heterogeneity is accomplished.
Moreover, another important consideration is
to have the PSU sufficiently compact in area so
that, with a small sample spread throughout,
it can be canvassed efficiently 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
occupations and industries.
The PSU’s are grouped into 461 strata.
Among these PSU’s, 146 of the largest SMSA’s
(including all having over 250,000 inhabitants)
and 10 other areas (not SMSA’s) are separate
strata representing themselves. In general,
however, a stratum consists of a set of PSU’s
as much alike as possible in various charac­
teristics such as geography, population density,
rate of growth in the 1960-70 decade, propor­
tion of Negroes and other minority races, prin­
cipal industry, and type of agriculture.
2 For a more detailed description of the sample design and
other technical phases of the CPS program, see U.S. Depart­
ment of Commerce, Bureau of the Census, Technical Paper
No. 7, “The Current Population Survey— A Report on Meth­
odology,” (1963).
3 See appendix C.

10

BLS HANDBOOK OF METHODS

Except for the 156 areas mentioned, each of
which is a complete stratum, the strata are es­
tablished so that their sizes in terms of 1970
population are approximately equal. From each
stratum a single PSU is selected to represent
the entire stratum. In the 156 strata in which
there is only a single PSU (the 146 SMSA’s
and 10 special cases), the single PSU auto­
matically falls in the sample. When the stratum
has more than one PSU, the sample PSU is
selected in a random manner in such a way
that its probability of selection is proportionate
to its 1970 population. For example, within a
stratum the chance that a PSU having a popu­
lation of 50,000 would be selected is twice that
for a unit having a population of 25,000.
Selection of Sample Households. The sample
design calls for a sampling ratio which de­
pends on the predetermined total sample size.
For 1971-73, it is 1 household for every 1,300
households in each stratum. The sampling ratio
is modified slightly by month, as the size of the
sample is held relatively constant despite the
overall growth of the population. 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 1970 Census. In a sample area
which was one-tenth of the stratum, the withinPSU sampling ratio that results is 1 in 130.0
thereby achieving the desired ratio of 1 in
1,300 for the stratum. For each PSU that is a
stratum representing only itself, the sampling
ratio is 1 in 1,300 regardless of the size of the
PSU.
With each of the 461 sample PSU’s, the num­
ber of households to be enumerated each month
is determined by the application of the withinPSU sampling ratio rather than through the
assignment 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 changing distribution of the popu­
lation and avoids the distortion which would
result from the application 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 1970 Census and contain, on the average,
about 250 households. These are selected sys­
tematically 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 proportionate to its 1970 popu­
lation.
The next step is to select a cluster of approxi­
mately 4 households to be enumerated within
each designated ED. This selection is made
wherever possible, from the list of addresses
for the ED compiled during the 1970 Census
or, if the addresses are incomplete or inade­
quate, by area sampling methods. The address
lists are used in about two-thirds of the cases,
primarily in urban areas, and area sampling is
applied in the remainder. An effort is made to
have all units at one single address included
within the same segment. This makes it rela­
tively easy for the interviewer to cover all units
designated for the sample.
The list sample is supplemented by a selec­
tion of the appropriate proportion of units
newly constructed in the PSU since the Census
date. The addresses of these units are obtained
mainly from records of building permits main­
tained by the offices responsible for issuing
permits in that area. A special procedure for
updating parts of the census lists also is fol­
lowed to reflect either units missed in the
Census or new construction in areas where
there is no adequate system of building per­
mits. In those enumeration districts where area
sampling 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 12 housing units or other living quar­
ters. For each subdivided enumeration district,
one segment is designated for the sample; the
probability of selection is proportionate to the
estimated “ size” of the segment. An inter­
viewer does not conduct interviews at all hous­
ing units in the segment, however; she is in­
stead given a sampling pattern so that, in gen­
eral, one-third of the units are enumerated. The

LABOR FORCE, EMPLOYMENT, AND UNEMPLOYMENT

remaining housing units in the segment are
then available for future samples.
Rotation of Sample. Part of the sample is
changed each month. A primary reason for ro­
tating the sample is to avoid the problems of
uncooperativeness which arise when a constant
panel is interviewed indefinitely. Another rea­
son for replacing households is to reduce the
cumulative effect of biases in response that
sometimes occur when the same persons are
interviewed indefinitely. To accomplish this
rotation of the sample on a gradual basis, maps
and other materials for several samples are
prepared simultaneously. 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 following 8
months, and then returns for the same 4 calen­
dar 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; the last
eighth are in for the eighth time, the fourth
month of the second 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 pro­
cedure provides a substantial amount of monthto-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 in­
quiry.

Collection Methods

Each month, during the calendar week con­
taining the 19th day, interviewers contact some
responsible person in each of the sample house­
holds in the CPS. At the time of the first
enumeration of a household, the interviewer
prepares a roster of the household members,
including their personal characteristics (date
of birth, sex, race, marital status, educational



11

attainment, veteran status, etc.) and their re­
lationship to the household head. This roster
is brought up to date at each subsequent inter­
view to take account of new or departed resi­
dents, changes in marital status, and similar
items. The information on personal charac­
teristics thus is available each month for iden­
tification purposes and for cross-classification
with economic characteristics of the sample
population.
At each monthly visit, a questionnaire is
completed for each household member 16 years
of age and over. The interviewer asks a series
of standard questions on economic activity
during the preceding week, the calendar week
containing the 12th day of the month, called
the “ survey week.” The primary purpose of
these questions is to classify the sample popu­
lation 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 ob­
tained on hours worked during the survey
week, together with a description of the curren'F job. For those temporarily away from
their jobs, the enumerator records their reason
for not working during the survey week,
whether or not they were paid for their time
off, and whether they usually work full or part
time. For the unemployed, he records (1)
methods used to find work, (2) the reason the
unemployed persons had started to look for
work, (3) the length of time they had been
looking for work, (4) whether they were seek­
ing full- or part-time work, and (5) a descrip­
tion 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 some­
thing else— is recorded. In addition, for all
households in the incoming or returning rota­
tion group, questions on the work history, rea­
sons for nonparticipation, and job seeking
intentions of individuals not in the labor force
are asked.
The questionnaires containing the informa­

12

BLS HANDBOOK OF METHODS

tion obtained for each person in the sample
are subjected to a field edit by clerks in each
of the 12 regional offices of the Census Bureau.
The field edit serves to catch omissions, incon­
sistencies, illegible entries, and errors at the
point where correction is still possible. Many
of the error corrections made in the field edit
prevent delays in further processing of the
questionnaires in Washington.
After the field edit, the questionnaires are
forwarded to the Washington office of the
Census Bureau. All of the questionnaires are
received in Washington by the end of the week
after enumeration. The raw data are trans­
ferred to computer tape and checked for com­
pleteness and consistency.
The interviewers on the CPS 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 re­
cruited and have either direct or home study
training each month, before the survey. More­
over, through editing of their completed
questionnaires, repeated observation during
enumeration, and a systematic reinterview of
part of their assignments by the field super­
visory staff, the work of the interviewers is
kept under control and errors or deficiencies
are brought directly to their attention.

Estimating Methods

To increase the reliability of the labor force
statistics derived from the sample, the esti­
mation procedure uses two stages of ratio
estimates and a “ composite estimate” . Achieve­
ment of this rather complicated procedure is
made rapidly and automatically because of the
availability of high-speed electronic digital
computers. The principal steps involved are as
follows.
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 be­
cause of absence, impassable roads, refusals, or



other reasons. This adjustment is made sep­
arately by groups of PSU’s and, within these,
for each color (white or Negro and other races)
and residence group of households (urban,
rural nonfarm, rural fa rm ). The adjustment is
made separately within each pair of rotation
groups (the incoming pair, the two continuing
pairs, and the outgoing pair).
Ratio Estimates. The distribution of the popu­
lation selected for the sample may differ some­
what, 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 correlated closely with la­
bor force participation and other principal
measurements made from the sample. There­
fore, some of the sample estimates can be
improved substantially when, by appropriate
weighting of the original returns, the sample
population is brought as closely into agreement
as possible with the known distribution of these
characteristics in the entire population. Such
weighting is accomplished through two stages
of ratio estimates:
1.
First stage. The first stage of ratio esti­
mates takes into account differences in the
distribution by color and residence of the popu­
lation estimates from the sample PSU’s and
that of the total population in each of the four
major regions of the country. However, in­
dependent distributions of the total population
by residence, cross-classified by color, are not
available on a current basis. Therefore, using
1970 Census data, estimated population totals
by color and residence for a given region are
computed from population counts for PSU’s in
the CPS sample. Ratios then are computed
between these estimates (based on sample
PSU’s) and the actual population totals for
the region as shown by the 1970 Census. In
deriving these ratios, PSU’s that made up en­
tire strata and were selected with certainty
(usually referred to as “ self-representing”
PSU’s) are excluded from the computations,
since they represent only themselves. In tabula­
tions of the monthly results from the CPS, the
weights for all sample households from nonself-representing PSU’s in a given region are

LABOR FORCE, EMPLOYMENT, AND UNEMPLOYMENT

multiplied 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
between 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 forward the most recent census
data (1970) to take account of subsequent
aging of the population, mortality, and migra­
tion between the United States and other coun­
tries. The CPS sample returns (taking into
account the weights determined after the first
stage of ratio estimates) in effect are used to
determine only the distribution within a given
age-color-sex group by employment status and
various other characteristics. In developing
statistics, these sample distributions are multi­
plied by the ratio of the independent population
estimate to the sample estimate for the appro­
priate age-color-sex group.
Composite Estimate. 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 for any particular item. The first esti­
mate is the result of the two stages of ratio
estimates described above. The second estimate
consists of the composite estimate for the pre­
ceding 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). Although the
weights for the two components of such a com­
posite estimate do not necessarily have to
be equal, in this instance the weights used for
combining these two estimates are each onehalf. 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



13

described; for some items the reduction is sub­
stantial. The resultant gains in reliability are
greatest in estimates of month-to-month
change, although gains also are obtained for
estimates of level in a given month, change
from year to year, or change over other inter­
vals of time.

Presentation and Uses

The CPS provides a large amount of detail
on the economic and social characteristics of
the population of the United States. It is the
source of monthly estimates of total employ­
ment, both farm and nonfarm; of nonfarm selfemployed persons, domestics, and unpaid help­
ers in nonfarm family enterprises; and of total
unemployment, whether or not covered by un­
employment insurance. It is a comprehensive
source of information on the personal charac­
teristics such as age, sex, race, educational
attainment, and marital status of the total la­
bor force and of the employed, the unemployed,
and those not in the labor force.
It provides distributions of workers by the
numbers of hours worked, as distinguished
from aggregate or average hours for an in­
dustry, permitting separate analyses of parttime workers, workers on overtime, etc. The
survey is a comprehensive current source of
information on the occupation of workers,
whether teachers, stenographers, carpenters,
laborers, etc. It also provides 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 information on their skills and past
work experience, if any.
Each month, a significant amount of basic
information about the labor force is analyzed
and published in Employment and Earnings.
The tables in this report provide information

14

BLS HANDBOOK OF METHODS

on the labor force, employment, and unemploy­
ment by a number of characteristics, such as
age, sex, color, marital status, industry, and
occupation. Approximately 150 of the most im­
portant estimates from the CPS are presented
each month on a seasonally adjusted basis.
These estimates are adjusted by the BLS Sea­
sonal Factor Method, which is described in the
appendix on seasonal adjustment.
The CPS is used also for a program of spe­
cial inquiries to obtain detailed information
from particular segments, or for particular
characteristics of the population and labor
force. Approximately 8 to 10 such special sur­
veys are made each year. The inquiries are
repeated annually in the same month for some
topics, including the earnings and total incomes
of individuals and families (published by the
Census Bureau), the extent of work experience
of the population during the calendar year, the
extent of overtime work at premium pay, usual
weekly earnings of wage and salary workers,
and the prevalence of multiple jobholding.
They also include marital and family charac­
teristics of workers, the employment of schoolage youth, the employment of recent high
school graduates, the educational attainment
of workers, and the employment situation in
poverty areas of our major cities. Surveys have
been made periodically on subjects such as job
mobility, and length of time on current job. In
addition, surveys are published in very great
detail on the characteristics of the unemployed
and persons not in the labor force.
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 sur­
vey requires the respondent to be the person
about whom the questions are asked.
Information obtained through the supple­
mental questions is combined with data in the
regular schedule to provide tabulations of all
the desired personal and economic charac­
teristics of the persons in the special survey.
Reports on these special surveys are first pub­
lished in the Monthly Labor Review. Reprints
of the articles, together with technical notes
and additional tables, are published as Special
Labor Force Reports.



Limitations

Geographic. The CPS 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 Errors 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 extent that varying the word­
ing of the question causes differences in re­
sponse, errors or lack of uniformity in the
statistics may result. Similarly, the data are
limited by the adequacy of the information pos­
sessed by the respondent and the willingness to
report accurately.
The estimates from the survey are subject to
various other types of errors beyond those al­
ready mentioned. Some of these are:
1. Nonresponse. About 3 to 5 percent of oc­
cupied 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 ac­
count for noninterviews, they still represent a
possible source of bias. Similarly, for a rela­
tively few households, some of the information
is omitted because of lack of knowledge on the
part of the respondent or because the inter­
viewer forgot to ask certain questions or record
the answers. In processing the completed ques­

LABOR FORCE, EMPLOYMENT, AND UNEMPLOYMENT

tionnaires, entries usually are supplied for
omitted items on the basis of the distribution
in these items for persons of similar charac­
teristics.
2. Independent Population Estimates. The
independent population estimates used in the
estimation procedure may also provide a source
of error, although on balance their use sub­
stantially improves the statistical reliability
of many of the important figures. (See discus­
sion under “ Ratio Estimates,” p. 12.) Errors
may arise in the independent population esti­
mates because of underenumeration 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 com­
ponents of population change (mortality, im­
migration, etc.) since that date.
3. Processing errors. Although there is a
quality 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. However, the
net error arising from processing is probably
fairly negligible.
Measuring the Accuracy of Results. Modern
sampling theory provides methods for estimat­
ing the range of errors due to sampling where,
as in the case of the CPS sample, the prob­
ability of selection of each member of the popu­
lation is known. Methods also are available for
determining the effect of response variability
in the CPS. A measure of sampling variability
indicates the range of difference that may be
expected because only a sample of the popula­
tion is surveyed. A measure of response varia­
bility indicates the range of difference that
may be expected as a result of compensating
types of errors arising from practices of differ­
ent interviewers and the replies of respond­

ents; these would tend to cancel out in an
enumeration 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 computations, however, do not, in­
corporate the effect of response bias, that is,
any systematic errors of response— for ex­
ample, 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 varia­
bility combined are provided in Employment
and Earnings and in other reports based on CPS
data, thus permitting the user to take this
factor into account in interpreting the data. In
general, the smaller figures and small differ­
ences between figures are subject to relatively
large variation and should be interpreted with
caution. The availability of the high-speed elec­
tronic computer makes possible considerably
more detailed estimates than were possible
earlier.
Estimation of response bias is one of the
most difficult aspects of survey and census
work. Systematic studies on this subject are
now an integral part of the CPS,4 but in many
instances available techniques are not suffi­
ciently precise to provide satisfactory estimates
of response biases. Considerable experimenta­
tion is in progress with the aim of developing
more precise measurements and improving the
overall accuracy of the series.*
* For a summary of these studies, see Bureau of the Census,
Technical Paper No. 6, “ The Current Population Survey Re­
interview Program— Some Notes and Discussion” (March
1963).

Technical References
Number

1.

President’s Committee to Appraise Employment and Unemployment Statistics, Measuring
Employment and Unemployment (1962).
A review of all Federal statistical series on employment and unemployment and a compari­
son 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 rec-




15

16

BLS HANDBOOK OF METHODS

2.

3.

4.

5.

ommendations for further research and improvements, several of which were inaugurated in
January 1967.
U.S. Department of Commerce, Bureau of the Census, Technical Paper No. 6, “ The Current
Population Survey Reinterview Program— Some Notes and Discussion” (1963).
A summary of procedures and results of the Current Population Survey Reinterview Pro­
gram from 1955 through 1961 and some interpretations and comparative results from other
studies.
U.S. Department of Commerce, Bureau of the Census, Technical Paper No. 7, “ The Current
Population 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.
U.S. Department of Labor, Bureau of Labor Statistics and U.S. Department of Commerce,
Bureau of the Census, Concepts and Methods Used in Manpower Statistics from the Current
Population Survey, BLS Report 313 and Current Population Reports, Series P-23, No. 22
(1967).
A concise description of the methodology in obtaining labor force information from sample
households. Labor force concepts and definitions are set forth. The adequacy of labor force
data and quality controls are discussed, and major improvements in the Current Population
Survey are listed chronologically.
U.S. Department of Labor, Bureau of Labor Statistics, How the Government Measures Un­
employment, Report 312 (1967).
A short nontechnical discussion of the concepts and methods used in obtaining labor force
statistics from the Current Population Survey.




Jo h n E . B regger

Chapter 2. Employment, Hours, and Earnings
Background

The first monthly studies of employment and
payrolls by the Bureau of Labor Statistics
(BLS) began in October 1915 and covered four
manufacturing industries. Before that year, the
principal sources of employment data in the
United States were the census surveys— the
decennial Census of Population, and beginning
in 1899, the quinquennial Census of Manufac­
tures. No regular employment data had been
compiled between the Census dates.
By November 1916, the BLS program had
been expanded to cover 13 manufacturing in­
dustries, and this number remained unchanged
until 1922. The depression of 1921 directed at­
tention to the importance of current employ­
ment statistics, and in 1922 Congress granted
additional funds to provide for program expan­
sion. By June 1923, the number of manufactur­
ing industries covered by the monthly employ­
ment survey had increased to 52. In 1928,
concern over increasing unemployment induced
Congress to provide additional appropriations
for the program. In the next 4 years, 38 manu­
facturing and 15 nonmanufacturing industries
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 Com­
mittee on Employment Statistics to study the
need for expanded data in this field. The Com­
mittee made its report in the spring of 1931
with a number of recommendations for exten­
sion 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 substantial in­
crease in the appropriation for the program.
In January 1933, average hourly earnings and
average weekly hours for the first time were
published for all manufacturing, for 90 manu­




facturing industries, and for 14 nonmanufac­
turing categories.
During the Great Depression when mass un­
employment threatened to become a continuing
aspect of American life, there was much contro­
versy among various authorities concerning
the actual number of the unemployed. These
discussions pointed up the fact that no reliable
measures of either unemployment or employ­
ment existed. In the early years of the Roose­
velt administration, the Secretary of Labor
frequently referred to the value of the Bureau’s
employment estimates as an indirect measure
of unemployment. This interest stimulated
efforts to develop comprehensive estimates of
total wage-and-salary employment in nonagricultural industries, and in 1936, the Bureau
first published such a figure.
The preparation of these estimates of overall
employment totals on a monthly basis was con­
tingent on the development of benchmark data.
It was recognized, even in the 1920’s, that
:month-to-month employment trends derived
from a sample of establishments might be fairly
accurate for short periods, but over long inter­
vals of time the series would not represent the
true movement of employment, unless they
were adjusted periodically to reasonably com­
plete counts of employment, called benchmarks.
The first such adjustment was made in 1935,
when the Bureau’s employment series in manu­
facturing were adjusted to totals from the Cen­
sus of Manufactures for 1923, 1925, 1927, 1929,
and 1931. These series were subsequently ad­
justed to the successive biennial Censuses of
Manufacturers, through that of 1939. For non­
manufacturing industries, benchmarks were
developed from various sources, including the
Censuses of Business taken at intervals from
1929 on.
From 1915 to the beginning of World War II,
interest in employment statistics for States and
areas was growing constantly. Even before the
Bureau of Labor Statistics entered the field in
17

18

BLS HANDBOOK OF METHODS

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 II in 1939, followed
by the entry of the United States after the as­
sault on Pearl Harbor in December 1941, placed
additional demands upon the Bureau’s Employ­
ment Statistics program. The added responsi­
bilities pointed up the need for greater uni­
formity among the various programs of estab­
lishment statistics on employment and related
subjects which were being prepared by the
BLS, the Bureau of the Census, and the agen­
cies 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 employ­
ment collected primarily for administrative
purposes by the newly organized social insur­
ance programs. Tabulations of such materials
became available about 1940 from the unem­
ployment insurance program and they soon be­
came the preferred sources of benchmark data.
They covered several industrial 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 sta­
tistics program has been a fully integrated



Federal-State project which provides employ­
ment, hours, and earnings information on a
national, State, and area basis in considerable
industrial detail. This cooperative program has
as its formal base of authority a Congressional
act of July 7, 1930 (ch. 873, 46 Stat. 1019; 29
U.S.C. 2). In 1971 cooperative arrangements
were in effect within 44 States and the District
of Columbia and with 6 State labor depart­
ments.

Description o f the Survey

The Bureau of Labor Statistics cooperates in
collecting data each month on employment,
hours, and earnings from a sample of establish­
ments in all nonagricultural activities including
government. In 1970, this sample included over
165,000 reporting units. From these data a
large number of series on employment, hours,
and earnings in considerable industry detail
are prepared and published monthly for the
United States as a whole, for each of the 50
States and the District of Columbia, and for
most of the metropolitan areas. The data in­
clude series on total employment, production or
nonsupervisory worker employment, women
employed, average hourly earnings, average
weekly hours, and average weekly overtime
hours (in manufacturing). For many series,
seasonally adjusted data also are published.
Concepts

An establishment is defined as an economic
unit which produces goods or services, such as a
factory, mine, or store. It is generally at a single
physical location and it is engaged predomi­
nantly in one type of economic activity. Where a
single physical location encompasses two or
more distinct and separate activities these are
treated as separate establishments, provided
that separate payroll records are available and
certain other criteria are met.1 In the collection
of data on employment, payrolls, and man­
hours, the BLS usually requests separate re­
1 Standard Industrial Classification Manual (Bureau of the
Budget, Executive Office of the President, 1967), p. x.

EMPLOYMENT, HOURS, AND EARNINGS

ports by establishment. However, when a com­
pany has more than one establishment engaged
in the same activity in a geographic area, these
establishments may be covered by a combined
report.
Industry employment statistics published by
BLS and the cooperating State agencies repre­
sent the total number of persons employed
either full-time or part-time in nonagricultural
establishments 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 personnel who worked on, or received
pay for, the last day of the month, plus inter­
mittent employees who worked any time during
the month (e.g., Christmas temporary em­
ployees of the postal service).
Employed persons include both permanent
and temporary employees and those who are
working either full- or part-time. Workers on
an establishment 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 dur­
ing the pay period are counted in each establish­
ment which reports them, whether the duplica­
tion 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 pro­
prietors, 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
employment statistics refer to civilian em­
ployees only.
The figure which includes all persons who
meet these specifications is designated “ all em­
ployees.” 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



19

of employees are designated production work­
ers, construction workers, or nonsupervisory
workers, depending upon the industry.
In manufacturing industries, data on em­
ployment, man-hours, and payrolls are collected
for 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 establish­
ment. Among the exclusions from this category
are persons in executive and managerial posi­
tions, and persons engaged in activities such as
accounting, sales, advertising, routine office
work, professional and technical functions, and
force account construction.2 Production work­
ers in mining are defined in a similar manner.
A more detailed description of the classes of
employees included in the production and non­
production workers categories in mining and
manufacturing is shown on the facsimile of
the BLS 790 A schedule on page 31 of this
bulletin.
In contract construction, the term construc­
tion 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 ordi­
narily performed by members of construction
trades. Exclusions from this category include
executive and managerial personnel, profes­
sional and technical employees, and routine
office workers.
Data on the employment, man-hours, and
payrolls of nonsupervisory workers are col­
lected from establishments in the transporta­
tion, communication, and public utility indus­
tries, 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 positions. (See fac­
simile of BLS 790 E, the reporting form for
wholesale and retail trade, p. 32.)
The series on hours and earnings is based on
reports of gross payroll and corresponding paid
- 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 employees.

20

BLS HANDBOOK OF METHODS

man-hours for production workers.3 To derive
*
these series, BLS collects the following data:
(See facsimile of BLS 790 A on p. 31).
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 work­
ers before deductions for old-age and unemploy­
ment insurance, group insurance, withholding
tax, bonds, and union dues. The payroll figures
also include pay for overtime, shift premiums,
holidays, 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., retrocative pay). 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 included unless they are paid
for at more than the straight-time rate. Hours
for which only shift differential, hazard, incen­
tive, or similar types of premiums were paid
are excluded from overtime hours.
Overtime hours data are collected only from
establishments engaged in manufacturing in­
dustries. For government organizations and
private educational institutions, payrolls col­
lected relate to all employees. Manhour data
are not collected.
3 Unless otherwise indicated, the references in this chapter
to production workers apply likewise to the construction
worker and nonsupervisory categories.




Industrial Glassification

All national, State, and area employment,
hours, and earnings series data are classified
in accordance with the Standard Industrial
Classification Manual, Office of Management
and Budget, 1967. (See appendix D of this
bulletin for a detailed description of this sys­
tem.)
Reporting establishments are classified into
significant economic groups on the basis of
major product or activity as determined by
the establishments’ percent of total sales or
receipts for the previous calendar year. This
information is collected once each year on an
“ Industry Class Supplement” to the monthly
report form. (See p. 33 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 under the indus­
try of the most important product or activity,
based on the percentages reported.
Time Period

Employment, hours, and earnings are meas­
ured for the pay period including the 12th of
the month, which is standard for all Federal
agencies collecting employment data on an
establishment basis.

Data Sources

Sample Data

Each month the State agencies cooperating
with the Bureau collect data on employment,
payrolls, and man-hours from a sample of es­
tablishments. The respondent extracts these
figures from his payroll records. These data
are readily available as the employers must
maintain such records for a variety of tax and
accounting purposes. A response analysis sur­
vey of the reporting practices of a scientifically
selected sample of reporting establishments in
manufacturing industries showed that the re­
ports were made out almost exclusively from
their payroll records. The survey also showed

EMPLOYMENT, HOURS, AND EARNINGS

that while a number of employers did not re­
port precisely the data requested on the sched­
ule 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 in­
correct reporting were quite insignificant.4
Participation in the industry employment
statistics program is entirely voluntary on the
part of the reporters. However, in many in­
dustries, particularly in manufacturing, em­
ployers 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
reasonably complete count of employment. The
estimates are adjusted periodically, annually if
possible, to new benchmark levels. Since 1939,
the basic sources of benchmark information for
“ all employees” have been periodic tabulations
of employment data by industry and, beginning
with 1959, by size of establishment. These are
compiled by State employment security agen­
cies, according to uniform procedures specified
by the Manpower Administration of the U.S.
Department of Labor, from reports of estab­
lishments covered under State unemployment
insurance laws. The State employment security
agencies receive quarterly reports, from each
employer subject to the laws, showing total
employment in each month of the quarter, and
total quarterly wages for all employees.5 If the
employer has more than 50 employees and op­
erates 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 indus4 Young, Dudley E. and Goldstein, Sidney, “ The BLS Em­
ployment Series and Manufacturing Reporting Practices,”
Monthly Labor Review, November 1957, pp. 1367-1371.
r The State employment security agencies are required to
>
submit tabulations of these reports to the Manpower A d ­
ministration each quarter. These tabulations are due in the
Washington office of MA by the middle of the sixth month
after the end of the quarter of reference. For example, the
first quarter tabulation, which provides the basis of the BLS
benchmarks, is due on September 15. Review and editing of
these tabulations and preparation of national summaries
from them requires several months additional work on the
part of both BLS and MA before the benchmark is completed.




21

tries. 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 assignments of industry classifications, so
there is a high degree of uniformity in this
respect between the benchmark and sample
data.
In 1970, unemployment insurance data ac­
counted for three-fourths of the total bench­
mark. For the group of establishments exempt
from State unemployment insurance laws be­
cause of their small size0 and for certain
classes of nonprofit institutions, the data used
are those provided by the national old-age in­
surance program administered by the Social
Security Administration of the U.S. Depart­
ment of Health, Education, and Welfare.
For industries not covered by either of the
two programs, benchmarks are compiled from
a number of special sources. The most im­
portant of these are the Interstate Commerce
Commission (interstate railroads), the Ameri­
can Hospital Association (private nonprofit
hospitals), the Office of Education in the U.S.
Department of Health, Education, and Welfare,
and the National Catholic Education Associa­
tion (private schools, colleges, and universi­
ties), the U.S. Civil Service Commission
(Federal Government), and the Governments
Division 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
reporting. Wherever possible, employment for
the standard midmonth pay period for March is
used as the benchmark.
1 In 1970 the unemployment insurance laws of 21 States and
1
the District of Columbia covered all employers of 1 or more
workers, 3 States covered employers of 3 workers or more,
and the remaining 26 States covered employers of 4 workers
or more. Beginning in 1972, all State unemployment insurance
laws should cover employers of 1 or more workers.
7 For a more detailed description of the benchmark, see
Armknecht, Paul A., “BLS Establishment Estimates Revised to
March 1970 Benchmark Levels,” Employment and Earnings,
September 1971, pp. 18-30.

22

BLS HANDBOOK OF METHODS

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 Manpower Administration,
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 sam­
ple of establishments in the States each month.
A “ shuttle” schedule is used (BLS 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 pro­
vided 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 in machine readable
form, to the Washington, D.C., office of the
Bureau of Labor Statistics, where they are
used to prepare estimates at the national level.
The shuttle schedule has been used in this
program since 1930, but there have been sub­
stantial changes in its design and in the data
collected over the period. All aspects of the
schedule, its format, the wording of the re­
quested items and definitions, and the concepts
embodied therein have been subjected to a con­
tinuing and intensive review over the entire
period, not only by the staff of BLS and of the
State agencies, but also by other government
agencies and by numerous persons in private
business and labor organizations. The report
forms are not exactly alike for every industry,
but most of them request data on total employ­
ment, number of women employees, number of
production workers (in manufacturing and
mining), construction workers (in contract
construction industries), or nonsupervisory
workers (in other nonmanufacturing indus­
tries), and, for these workers, data on payroll,
paid man-hours, and for manufacturing, over­
time man-hours. The schedule contains detailed
instructions and definitions for the reports.
There are several variants designed to meet
the specific problems of different industries.



(See facsimiles of BLS 790 A and, BLS 790 E,
pp. 31 and 32.)
The technical characateristics of the shuttle
schedule are particularly important in main­
taining continuity and consistency in reporting
from month to month. The design exhibits
automatically 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 sched­
ule also has operational advantages; for ex­
ample, 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 consistent within themselves, with the data
reported by the establishment in earlier
months, and with those reported by other es­
tablishments in their industry. This editing
process is carried out in accordance with a de­
tailed manual of instructions prepared by the
Bureau of Labor Statistics. When the reports
are sent to Washington, they are screened by
use of an electronic computer to detect process­
ing errors and reporting errors which may
have escaped the first editing. Questionable
cases discovered at any stage of the editing
process are returned, if necessary, to the re­
spondent for review and correction.8

Sam pling
Sampling is used by BLS in its industry
employment statistics program for collecting
data in most industries, since full coverage
would be prohibitively costly and time consum­
ing. The sampling plan for the program must:
(a) provide for the preparation of reliable
monthly estimates of employment, hours of
work, and weekly and hourly earnings which
can be published promptly and regularly; (b)
through a single general system, yield con­
siderable industry detail for metropolitan
areas, States, and the Nation; and (c) be ap8 Mendelssohn, Rudolph C. "Machine Methods in Employ­
ment Statistics,” Monthly Labor Review, May 1955, pp. 567-569.

EMPLOYMENT, HOURS, AND EARNINGS

propriate for the existing framework of op­
erating procedures, administrative practices,
resource availability, and other institutional
characateristics of the program.
In developing the sample design, the universe
of establishments was stratified first by in­
dustry and within each industry by size of
establishment in terms of employment, using
six standard size classes. Within each industry,
an optimum allocation design was obtained by
sampling with probability proportionate to
average size of establishment 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 estab­
lishment in each cell. In practice, this is equiva­
lent to distributing the predetermined total
number of establishments required in the sam­
ple among the cells on the basis of the ratio
of employment in each cell to total employment
in the industry. Within each stratum, the sam­
ple members are selected at random.
Under this type of design, large establish­
ments fall into the sample with certainty. In
nearly all industries, establishments with 250
or more employees are included in the sample
with certainty and in many industries the cut­
off is lower. The sizes of the samples for the
various industries were determined empirically
on the basis of experience and of cost consid­
erations. In a manufacturing industry in which
a high proportion of total employment was
concentrated in a relatively few large establish­
ments, a high percentage of total employment
had to be included in the sample. Consequently,
the sample design for such industries provides
for a complete census of the larger establish­
ments with only a few chosen from among the
smaller establishments, or none at all if the
concentration of employment in the larger es­
tablishments is great enough. On the other
hand, in an industry where a large proportion
of total employment is in small establishments,
the sample design calls for inclusion of all large
establishments, and also for a substantial num­
ber of the smaller establishments. Many in­
dustries 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



23

resources, it is necessary to accept samples in
these divisions with a smaller proportion of
universe employment than is the case for most
manufacturing industries. Since individual es­
tablishments in these nonmanufacturing indus­
tries generally show less fluctuation from
regular cyclical or seasonal patterns than es­
tablishments in manufacturing industries,
these smaller samples (in terms of employ­
ment) generally produce reliable estimates.
This sample design, although aimed pri­
marily at meeting the needs of the national
program, provides 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 preparation
of State and area estimates.9

Estimating Procedures
Employment

The “ all-employee” estimates by industry are
based on reasonably complete employment
counts or “ benchmarks.” To obtain employment
estimates for the individual estimating cells,
the following 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 rea­
sonably 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 es­
tablishments which reported in both months.
3. Beginning with the benchmark month, the
estimate for each month is obtained by multi­
plying the estimate for the previous month by
the link relative for the current month.
Application of the estimating procedure in
preparing a series is illustrated by the follow­
ing example: Assume that total employment
9 For the national sample, additional reports needed for
State and area samples are added to those required by the
national design.

24

BLS HANDBOOK OF METHODS

for a given series was 50,000 in July. The re­
porting sample, composed of 60 establishments,
had 25,000 employees in July and 26,000 in
August, a 4-percent increase. To derive the
August estimate, the change for identical es­
tablishments reported in the July-August sam­
ple is applied to the July estimate:
5 0 ,0 0 0 x -6>° - ( o r 1.04) =52,000
25,000
This procedure for estimating current em­
ployment is known as the benchmark and
link-relative technique. It is an efficient tech­
nique, taking advantage of a reliable complete
count of employment and of the high correla­
tion between levels of employment in successive
months in identical establishments.
In addition to estimates of total employment
by industry, the Bureau publishes data on
production, construction, or nonsupervisory
worker employment. For this purpose, the
sample ratio for the current month of produc­
tion workers to total employment is used. For
example, the 60 sample firms which had 26,000
employees in August, reported an August pro­
duction-worker figure of 19,500 resulting in a
19,500
ratio of
or .750. Using this ratio, pro26,000
duction-worker employment in August is esti­
mated to be 39,000 (52,000 multiplied by .750=
39,000). A similar ratio method is used to esti­
mate the number of women employed.1
0
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 bench­
mark month is March. The employment esti­
mates which had been published previously for
that month are compared with the new bench­
1 If permanent changes in the composition of the sample
0
take place, the “ production-workers, all-employee” ratios and
the "w om en” ratios calculated from the sample are modified
by a wedging technique described in this chapter under
"Hours and Earnings.” (See p. 25.)




mark data. The amount of adjustment in the
published employment information is indicated
by this comparison. The all-employee series,
for months between the current and the last
preceding benchmark, are adjusted by wedging
or tapering out the difference between the cur­
rent benchmark and the estimate for the bench­
mark 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 re­
visions made since 1966, is presented in table 1.
Table 1. Nonagricultural employment estimates, by
industry division, as a percentage of the benchmark
for recent years
Industry division

1966

1967

1968

1969

1970

__________
Total
Mining ...
____ ....
Contract construction ............
Manufacturing ____________
Transportation and public
utilities
_ ____ ___
_
Wholesale and retail trade ....
Finance, insurance, and
real estate __ _______ _____ _
Services ___ __ ___ ____________
Government _________________

99.9
100.5
99.7
99.4

100.0
99.5
101.6
99.5

100.4
101.7
99.5
99.8

99.8
101.5
99.0
99.8

100.0
100.0
100.1
100.1

99.7
100.1

99.8
100.7

100.7
100.3

100.4
100.0

100.0
100.1

99.5
100.3
100.0

100.2
99.8
100.0

99.2
99.2
102.8

100.0
99.1
100.1

100.3
99.6
100.2

Hours and earnings

Independent benchmarks are not available
for the hours and earnings series. Conse­
quently, the levels shown are derived from the
BLS reporting sample.
Since 1959, when benchmark data stratified
by employment size became available through
the employment security system, estimates of
employment, 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 esti­
mates shows that the six size classes as de­
scribed 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

EMPLOYMENT, HOURS, AND EARNINGS

size is needed.1 Whenever a new national
1
benchmark becomes available, national esti­
mates of average weekly hours and average
hourly earnings using eight size strata and
four regional strata (Northeast, North Cen­
tral, 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. If this is the case, a change is intro­
duced into the estimating structure at the time
of the next benchmark revision.

25

Note is then taken of the published estimate
of average hourly earnings for the previous
month, say V0. Because the panel of establish­
ments reporting in the sample is not absolutely
fixed from month to month, there may be differ­
ences between V0 and U0.1 A final figure for
1
2
the current month is obtained by making use
of both pieces of information; the estimate is
F1==(0.9 Vo + 0.1 U0) - h ( U 1- U 0)

The procedure, reflected in this last equation,
accepts the advantage of continuity from the
use of the matched sample, and at the same
time tapers or wedges the published estimate
a.
Average weekly hours and gross hourly
toward the level of the latest sample average.
earnings. To obtain average weekly hours for
The same procedure is used to adjust the
an individual estimating cell, the sum of the
production-worker all-employee-ratio and the
man-hour totals reported by the plants classi­
fied in that cell is divided by the total number
Table 2. Number of industries for which “Primary”
series are published under the BLS Industry Employ­
of production workers reported for the same
ment Statistics Program— employment, hours, and
establishments. Similarly, in computing aver­
earnings, January 1971
age hourly earnings, the reported payroll total
is divided by the reported man-hour total.
Hours
Pro­
All
Average
and
The first ratio estimates of average hourly
duction Women
Industry division
em­
earn­
overtime
hours
ployees workers 1
ings 2
earnings and average weekly hours are modi­
417
362
402
364
150
Total _____
fied at the estimating cell level by a wedging
12
11
11
9
Mining ______ __
Contract
technique designed to compensate for changes
11
11
construction ....
278
11
11
277
277
146
Manufacturing ..
278
in the sample arising mainly from the volun­
Transportation
and public
tary character of the reporting.
25
16
21
19
utilities ______
32
36
35
32
Trade __________
Finance,
For example, a first estimate of average
insurance, and
9
15
15
9
real estate .....
hourly earnings for the current month, Ult is
23
5
22
5
Services ............
15
4
Government ___
obtained from aggregates from a matched
__
1
1
1
1
Total private ___
Total nonsample of establishments reporting in the cur­
1
1
1
agriculture _
_
rent and previous month. Similarly an estimate
1 Production workers in manufacturing and mining; con­
of average hourly earnings, U0 for the previ­
,
struction workers in contract construction; nonsupervisory
workers, all other divisions.
ous month is calculated from the same matched
- Average hourly earnings, average weekly hours, and
sample. Hence, Ui-U0 is a measure of absolute
average weekly earnings.
change between the 2 months.
ratios for women with regard to changes in the
composition of the sample between successive
1 These combinations of size classes have been made be­
1
cause of operational economies. In particular, the preliminary
months.
estimates are based on less than full samples, and if the
Weekly hours and hourly earnings for indus­
estimation of preliminary estimates was attempted using the
full stratification pattern, there would be a number of cells
tries and groups above the basic estimating
for which there were no samples. Experiments and tests over
cell level are weighted averages of the figures
several years indicate that estimates of hours and earnings
prepared from the BLS sample using a maximum of three
for component cells and industries. The average
size strata generally do not differ significantly from those
weekly hours for each estimating cell are
computed with four size strata or more.
12 If the difference between the estimate and the average
multiplied by the corresponding estimate of
computed from the sample (V (— U0), is too great, the sample
1
production-worker employment, to derive ag­
average is accepted once it has been established that the
difference is due to a permanent change in the composition of
gregate man-hours. Payroll aggregates are the
the sample, and the series is regarded as discontinuous at
product of aggregate man-hours and average
that point. In general, a difference greater than 3 percent is
considered as defining a discontinuity or “ break.”
hourly earnings. Payroll and man-hour aggre


26

BLS HANDBOOK OF METHODS

gates for industry groups and divisions repre­
sent the sum of aggregates for component
industries.
Average weekly hours for industry groups
are obtained by dividing the man-hour aggre­
gates by the corresponding production-worker
employment estimates. Average hourly earn­
ings for groups are computed by dividing the
payroll aggregates by the man-hour aggre­
gates. This method is equivalent to weighting
weekly hours by estimated universe productionworker employment and hourly earnings by
estimated universe man-hours.
For all levels, from individual estimating
cells to major industry divisions, average
weekly earnings are computed by multiplying
average hourly earnings by average weekly
hours.
b. Overtime man-hours. To obtain average
weekly overtime hours in manufacturing in­
dustries, the sum of the overtime man-hours
reported is divided by the number of produc­
tion workers in the same establishments.
c. Spendable average weekly earnings. Be­
fore the majority of workers in lower income
brackets were subject to Federal income and
social security taxes, gross average weekly
earnings were a satisfactory measure of trends
in weekly earnings available for spending.
After Federal income taxes began to affect the
spendable earnings of an appreciable number
of workers, a method was developed for ap­
proximating spendable earnings by deducting
these taxes from gross earnings.1
3
The amount of individual income tax liability
depends on the number of dependents sup­
ported by a worker as well as on the level of
his gross income. Spendable earnings for work­
ers by major industry division 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 1967
dollars, to give an approximate measure of
changes in “ real” gross and spendable weekly
1 Utter, Carol, “The Splendable Earnings Series: A Technical
:5
Note on its Calculation,” Employment and Earnings and
Monthly Report on the Labor Force, February 1969, pp. 6-13.




earnings, that is, in purchasing power since
that base period. This series is computed by
dividing the weekly earnings average (in cur­
rent dollars) by the BLS Consumer Price Index
for the same month.
d. Average hourly earnings, excluding over­
time, in manufacturing are computed by divid­
ing the total production-worker payroll for the
industry group by the sum of total productionworker man-hours and one-half of total over­
time man-hours, which is equivalent to pay­
rolls divided by straight-time man-hours. This
method excludes overtime earnings at 11/2
times the straight-time rates; no further ad­
justment is made for other premium payment
provisions.
e. Indexes of aggregate weekly man-hours
and payrolls are prepared by dividing the cur­
rent month’s aggregates by the average for
1967.
Reliability of estimates

Although the relatively large size of the BLS
establishment sample assures a high degree of
accuracy, the estimates derived from it may
differ from the figures that would be obtained
if it were possible to take a complete census
using the same schedules and procedures. As
discussed previously a link relative technique
is used to estimate employment. This requires
the use of the previous month’s estimate as
the base in computing the current month’s esti­
mate. Thus, small sampling and response
errors may cumulate over several months. To
remove this accumulated error, the estimates
are adjusted annually to new benchmarks. In
addition to taking account of sampling and
response errors, the benchmark revision ad­
justs the estimates for changes in the indus­
trial classification of individual establishments
(resulting from changes in their product which
are not reflected in the levels of estimates until
the data are adjusted to new benchmarks). In
fact, at the more detailed industry levels, par­
ticularly within manufacturing, changes in
classification are the major cause of benchmark
adjustments. Another cause of differences, gen­
erally minor, arises from improvements in the
quality of the benchmark data.

EMPLOYMENT, HOURS, AND EARNINGS

One measure of the reliability of the employ­
ment estimates for individual industries is the
root-mean-square error (RM SE). The measure
is the standard deviation adjusted for the bias
in estimates.
RMSE—/(Standard Deviation)2-/(B ia s )2
If the bias is small, the chances are about 2 out
of 3 that an estimate from the sample would
differ from its benchmark by less than the rootmean-square error. The chances are about 19
out of 20 that the difference would be less than
twice the root-mean-square error.
The hours and earnings estimates for cells
are not subject to benchmark revisions, al­
though the broader groupings may be affected
slightly by changes in employment weights.
The hours and earnings estimated, however,
are subject to sampling errors which may be
expressed as relative errors of the estimates.
(A relative error is a standard error expressed
as a percent of the estimate.) Measures of rootmean-square errors for employment estimates
and relative errors for hours and earnings esti­
mates are provided in the “ Technical Note” of
Employment and Earnings.
Seasonally adjusted series

Many economic statistics, including employ­
ment and average weekly hours, reflect a
regularly recurring seasonal movement which
can be measured on the basis of past experi­
ence. By eliminating that part of the change
which can be ascribed to usual seasonal varia­
tion, it is possible to observe the cyclical and
other nonseasonal movements in these series.
Seasonally adjusted series are published reg­
ularly for selected employment hours, and
earnings series.
The seasonal adjustment method used for
these series is an adaptation of the standard
ratio-to-moving average method, with a pro­
vision for “ moving” adjustment factors to take
account of changing seasonal patterns. A de­
tailed description of the method is given in
appendix A of this bulletin.
The seasonally adjusted series on gross aver­
age weekly hours, average overtime hours and
average hourly earnings are computed by ap­
plying factors directly to the corresponding



27

unadjusted series, but seasonally adjusted em­
ployment totals for all employees and produc­
tion workers by industry divisions are obtained
by summing the seasonally adjusted data for
component industries. Selected seasonally ad­
justed series also are prepared for aggregate
weekly man-hours.

Presentation and Uses

At the national level, the program produces
each month a total of over 2,600 separate pub­
lished series. Tables 2, 3, and 4 provide a sum­
mary of the detail which is published currently.
Table 2 describes the “ primary” series pro­
duced 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 ap­
plication of special adjustments, while table 4
lists the seasonally adjusted series by type and
industry division.
In addition to the series published on a cur­
rent monthly basis, a single annual figure for
employment in March of each year (based on
benchmarks) is published for a number of in­
dustries for which monthly estimates do not
currently meet established standards for pub­
lication. In 1971, following revision to the 1970
benchmark, data for 239 such industries were
published.
In April 1971, employment and hours and
earnings statistics were available for 50 States,
the District of Columbia, and 211 areas. Ap­
proximately 8,200 employment series and hours
and earnings series for about 3,400 industries
were published for these States and areas by
the State agencies. The employment series
usually covered total nonagricultural employ­
ment, major industry divisions (e.g., contract
construction, manufacturing), and major in­
dustry groups (e.g., textile mill products,
transportation equipment, retail trade) for
each State and area. Additional industry detail
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
earnings appear in several BLS publications.

28

BLS HANDBOOK OF METHODS

Table 3. Number of industries for which special series
are published under the BLS Industry Employment
Statistics Program— employment, hours, and earnings,
January 1971

Industry division

Total
private ..
Mining, construction and
manufacturing
Mining _____ ____
Contract construction
Manufacturing ..
Transportation
and public
utilities ______
Trade ..... ........... .
Finance, insurance and real
estate ________
Services _____

Index of Index o f
aggreaggregate
gate
weekly
weekly
manpayhours
rolls

Spendable
average
weekly
earnings 1

Gross
weekly
earnings
(1 9 6 7
dollars)

1

Average
hourly
earnings
( excluding
overtim e)

1

1
1

1

1

1

1
24

I
1

1
1

1
1

1
1

1
1

1
1

1
1

23

1 In current and 1967 dollars.

The summary data are first published each
month in a press release which contains pre­
liminary national estimates of nonagricultural
employment, weekly hours, and gross average
weekly and hourly earnings in the preceding
month, for major industry categories. The re­
lease also includes seasonally adjusted data on
employment, average weekly hours, and aver­
age overtime hours. The preliminary estimates
are based on tabulations of data for less than
the full sample to permit early release of fig­
ures. This release normally is issued 3 weeks
after the week of reference for the data. The
press release also includes a brief analysis of
current trends in employment, hours, and earn­
ings, pointing up current developments as com­
pared with those for the previous month and
the same month in the preceding years.
National estimates in the detail described in
tables 2, 3, and 4 are published in the monthly
report Employment and Earnings. This pub­
lication is issued about 5 weeks after the week
of reference. Employment data for total non­
agricultural employment and for the major
industry divisions, as well as hours and earn­
ings for all manufacturing, are published for
States and areas in Employment and Earnings
1 month later than those for the Nation. Special
articles analyze long-term economic movements
or describe technical developments in the pro­



gram. Many of the national series are repub­
lished in the Monthly Labor Review with data
shown for each series for the most recent 13
months.
Following each benchmark revision, an his­
torical volume called Employment and Earn­
ings, United States is published. This provides
historical data, monthly and annual averages,
from the beginning date of each series, in a
few instances as far back as 1909. A companion
volume, Employment and Earnings, States and
Areas, provides historical data (annual aver­
ages) 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. De­
tailed industry rates are available monthly in
releases published 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 Governors of the Federal Reserve System
and the Council of Economic Advisors repub­
lish all or part of the data. They are also reg­
ularly republished in summary form or for
specific industries in many trade association
journals, the labor press, and in general refer­
ence works.
These series are used by labor unions, busi­
ness firms, universities, trade associations,
private research organizations, and many gov­
ernment agencies. Research workers in labor
unions and industry, as well as others responsi­
ble 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 eco­
nomic 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 economic data to yield measures of
productivity.
Analysts study employment trends to detect

29

EMPLOYMENT, HOURS, AND EARNINGS

Table 4. Number of seasonally adjusted series published under the BLS Industry Employment Statistics Pro­
gram— employment, hours, and earnings, January 1971
Seasonally adjusted series
Industry division
A ll
em­
ployees
Total nonagriculture ___________ ______
Total Drivate _____________________
Mining, construction, and manufacturing _______
Mining
Contract construction .... ........ ..............................
Manufacturing
Transportation and public utilities __________
Trade ___________________
_______
Finance, insurance, and real estate .....
Services
Government ______ ___ ___ ______ ____ _ __ . . .

Pro­
duction
workers

1
1
1

1
1
1

24

24

1
3

1
1
1

Employment

Total employment in nonagricultural estab­
lishments from the “ payroll” survey is not
directly comparable with the Bureau’s esti­
mates of the number of persons employed in
nonagricultural industries, obtained from the
monthly “ household” survey.1 The “ payroll”
5
1 Nelson, Darrell, “BLS Earnings Statistics for Use in
4
Escalation Agreements,” Employment and Earnings and
Monthly Report on the Labor Force, March 1968, pp. 18-21.
See chapter 1 for a description of this survey.

—

Average
over­
time
hours

_

....

1

1

1
1
1
1
1
1
1

3

1
1
1

Man
hour
index
....

1
1
1
1

Limitations

Average
weekly
hours

1
1

changes in industrial structure, and to observe
growth and decline proclivities of individual
industries. They also are used in the Bureau’s
Occupational 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 increase or a lower­
ing of the settlement price depending on the
movement of average hourly earnings in a
selected industry.1 Wide need has been demon­
4
strated by both labor and business for industry
series on hourly earnings and weekly hours, to
provide a basis for labor-management negotia­
tions. They not only furnish current and his­
torical information on a given industry but
provide comparative data on related industries.




Index
Average Average
weekly
hourly
of
earn­
employ­
earn­
ings
ment
ings

1
1

1
1
1

24

24

3

—

—

1
3
1
1

series excludes unpaid family workers, domes­
tic servants in private homes, proprietors and
other self-employed persons, 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 accord­
ing to his single major activity. Certain per­
sons on unpaid leave are counted as employed
under the “ household” survey, but are not in­
cluded in the employment count derived from
the “ payroll” series. In addition to these differ­
ences in concept and scope, the surveys employ
different collection and estimating techniques.
Therefore, although each survey measures
changes in employment, 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 provides a great deal
of information on the demographic charac­
teristics (e.g., sex, age, race) of the labor force.
The survey is not well suited to providing de­
tailed information on the industrial and the
geographic distribution of employment. The
establishment survey, while providing limited
information 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 em­
ployment figures. The payroll and household
surveys therefore may be regarded as comple­
mentary.
Employment estimates derived by the Bu-

30

BLS HANDBOOK OF METHODS

reau of the Census from its quinquennial
censuses and from the annual sample surveys
of manufacturing establishments may differ
from BLS employment statistics. The most im­
portant reason for difference stems from the
degree to which multiproduct establishments
file separate or combined reports in one survey
but not the other, which m^y result in different
industrial classification of employment. There
is also a significant difference at the more de­
tailed industry levels, since Census classifies
auxiliary units and central and district ad­
ministrative units on the basis of the most
appropriate 2-digit major group, while BLS
codes these units to the most appropriate
4-digit industry. For broad categories, how­
ever, the two surveys do show similar levels
and trends.

Hours and earnings

The workweek information relates to aver­
age hours paid for, which differ from scheduled
hours or hours worked. Average weekly hours
reflect the effects of such factors as absentee­
ism, 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. Earnings for those
employees not covered under the production
worker and nonsupervisory-employee cate­
gories are, of course, not reflected in the
estimates.
The series on spendable weekly earnings
measure the net earnings of workers who earn
the average gross weekly earnings, have the
specified number of dependents, and take the
standard deductions for Federal income tax
purposes. Spendable 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 earnings.1 They
6
do not take into account payroll deductions for
such purposes as State income taxes, union
dues, or group insurance, and they do not re­
flect such factors as total family income or tax
deductions above the standard amount.
The “ real” earnings data (those expressed in
1967 dollars), resulting from the adjustment
of gross and spendable average weekly earn­
ings 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 unemployment.
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, lateshift work, and premium overtime rates other
than at time and one-half.
The ultimate goal of the program is to pro­
vide current estimates of employment, hours,
and earnings for all nonagricultural industries
in the Nation as a whole, and also for all sig­
nificant industries in all States and all Stan­
dard Metropolitan Statistical Areas, as defined
by the Office of Management and 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 di­
rected toward strengthening the sample so that
series for employment, hours, and earnings for
additional industries may be published, and
also toward developing series for additional
standard metropolitan areas.
1 For a complete analysis of the difference between spend­
6
able and disposable earnings, see Paul Schwab, "T w o Mea­
sures of Purchasing Power Contrasted,” Monthly Labor Re­
view, April 1971.

BLS Codes

Form BLS 790 A
S ta te

R e p o rt No.

In d .

L O C A T IO N O F E S T A B L IS H M E N T S C O V E R E D IN TH IS R E P O R T

(Num ber of establishments)

(C ity)

(County)

(State)

..................................................
..................................................
B e fo re en terin g d ata see exp la n a tion s on oth er side
P A Y P E R IO D

ALL E M P L O Y E E S

The Bureau o f Labor S ta tistics and the State agencies coop erating
in its statistical p rogram s will hold all in form a tion furnished by
the respondent in strict confidence.

P R O D U C T IO N A N D R E L A T E D W O R K E R S

YOUR

From—

Through

(Both date incluiive)
(1)

(2 )

(3)

Knter the number of
days worked plus paid
holidays and paid va­
cation days for majorit y
of p rod u ction
w orkers.
( N ea rest
<lay)

Vi

During
the
entire pay
period
(4)

During the 7
consecutiveday period
which in­
cludes the
12th
(5)

NUM BER
DO
NOT
USE

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

Knter in these columns the number of production and re­
lated workers who worked during or received pay for
any part of the period reported, the p a y earned (before
deductions), and all hours worked or paid for. Include
pay and man-hours for overtime, sick leave, holidays,
and vacations

L/P
Both sexes

(« )

Women
only
(8)

T o ta l
P r o d u c t io n W orker
P ayroll

N um ber or
P r o d u c t io n
W orkers
(»)

(7)

1970
D ec.
1971
J an ..

T o tal
P r o d u c t io n W orker
M a n -H o u r s

COM M ENTS

Enter in column 13 the main factors responsible for significant month-to-month
changes in employment, average hours worked (col. 11+col. 9), average hourly
earnings (ool. 10+col. 11), etc., as indicated by this report. Examples are:
Wage rate increase
More bueineet

DO
NOT
USE

( Omit centi)

(Omit fractioni)

Expl.
Code

(10)

(11)

Strike
Fire

Overtime
Weather

If any general wage-rate changes (not Individual changes for length of
service, merit, or promotion) have occurred slnoe last month’s report note
amount of increase or decrease (as + 2% , — M ), the effective date of the
change, and the approximate number of production workers affected.
(13)

(12)

S

F eb -..
M a n ..

A p r...

BLS HANDBOOK OF METHODS

Y ear
AND
M onth

Knter beginning
anti e n d i n g
dates of pay
period w h i c h
includes
th e
12th of t h e
month

M a y ..

>r

'

June .
J u ly .»

A ug...

7

S ept..
O c t ...
N o v ..
D e c ...

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




(Position)

CO

Form BLS 790 E

BLS Code.
......................1
........................ . . J .............................. II......................
S ta te
| R ep ort N o. i
Ind.
L O C A T IO N

OF

E S T A B L IS H M E N T S ) C O V E R E D

( N u m b e r o f e s t a b li s h m e n t s )

OJ

to

Before entering data see
explanations on other side

(C ity )

IN T H I S

REPORT

(C o u n ty )

(S td tn

T he Bureau o f Labor S tatistics and the State agencies co o p e r­
ating in its statistical p rogra m s will hold all in form ation fu r­
nished by the respondent in strict confidence.

PAY

Y ear
AND
M onth

K n tei
b e g in n in g
and e n d i n g
dates
of
pay
l»e r io d
w h ic h
in c lu d e s
th e
1 2 th
of
th e
m o n th

D u r in g
th e
e n t ir e p a y
p e r io d

( Hoth date: in c lu s iv e )

(2)

(1)

D u r in g th e
7 c o n s e c u t iv e
d a y p e r io d
w h ic h
in c lu d e s
th e 1 2 th

(4)

T h rou gh

(5)

v3)

N O N S U P E R V IS O R Y

ALL E M P L O Y E E S
NUM BER
DO
N O T
USE

L /P

I n c lu d e a ll p e r s o n s w h o w o r k e d
d u r i n g o r r e c e iv e d p a y for a n y
p a r t o f p e r io d r e g a rd le ss o f
t y p e o f w o r k p e r fo r m e d

B o t h sexes

W o m e n o n ly

K ilter in th ese c o lu m n s the* n u m b e r o f n o n s u p e r v is o r y e m ­
p lo y e e s w h o w o r k e d d u r i n g o r r e c e iv e d p ay fo r a n y p a r t
o f th e p e r io d r e p o r t e d , th e p a y e a r n e d ( b e f o r e d e d u c ­
t io n s b u t e x c l u d i n g c o m m is s io n s ,) a n d a ll h o u rs w o r k e d
o r p a id fo r .
I n c lu d e p a y a n d m a n -h o u r s f o r o v e r t im e ,
s ic k le a v e , h o lid a y s , a n d v a c a t io n s

N um ber or
N on su per­
v is o r y

E m ployees

(6)

(8)

(?)

EM PLOYEES

\ON8U PERU ISORY KM PLOY EE
P ayroll
( E x clu d in g c o m ­
m is s io n s r e jio r te d
in c o l . 1 0 A )
( O m it c e n t s )

T o ta l
X o n s u p e r v is o r y Kmployee
M a n -H ours
( O m it f r a c t i o n s )

A m ount of
c o m m is s io n s
(o m it c e n ts)

(10)

(11)

( 10A )

(»)

D O N O T USE

C o m m is s io n s o r N o n s u p e r v is o r y
E m ployees

P e r io d in w h ic h
earn ed
PR
|T hrough
1
( Hoth d a tes in c lu s iv e )
(1 0 B )
1 (1 0 C )

H

E xpl
code

(111)

(12)

F rom -

(10*)

1970
$

$

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

1971

M a r ..
Apr
......................................
M ay

June .
J u ly..
Aug
<
N('pt-.

D e c ...

YOUR COMMENTS ON CHANGES IN EMPLOYMENT, PAYROLL. OR WAGE RATES
Enter in colum n 13 the main factors responsible for significant m onth-to-m onth changes in the report above.

Examples are:

W a g e r a te

increase, more

b u s in e s s , fir e , te m p o r a r y s u m m e r h elp ,

o v e r tim e , s tr ik e , w ea th er .

If any G E N E R A L W A G E -R A T E C H A N G E S (not individual changes for length of service, merit, or p rom otion ) have occurred since last m onth’s report, note the am ount of increase or
decrease (as + 2 % , — 5^), the effective date of the change, and the approxim ate number o f nonsupervisory em ployees affected.
_________________________________________________________________
(13)

(13)

(13)

1971
Sept.

1971

1971

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

O c t......................................

Feb
July
* R r.............................................. ....................................................................

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




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

A ug.........

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

N o v ...............................................................
D ec.

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

(Position)

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

EM P LO YM EN T , HOURS, A N D E A R N IN G S

From

P E R IO D
K ilte r th e n u m b e r o f d a y s
w o r k e d p lu s p a id h o l i ­
d a y s a n d p a id v a c a t io n
d a y s f o r m a jo r i t y o f
n o n s u p e r v is o r y
e m p lo y e e s .
( X ca rest
\
(l a y )

BLS HANDBOOK OF METHODS

33

BLS 790 Industry Class Supplement

Budget Bureau No. 44-R745

Approval expires January 31, 1972

STATEMENT OF PRODUCTS

MANUFACTURING

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

U.S. DEPARTMENT OF LABOR
BUREAU OF LABOR STATISTICS
WASHINGTON, D.C. 20212

r

SAMPLE COPY

The Bureau of Labor Statistics and the State agencies cooperating
in its statistical programs will hold all information furnished by
the respondent in strict confidence.

L
BLS
CODES

J

State

Report No.

Ind.

Empl.

Mo.

Yr.

L O C A T IO N ________________________________________________________
(City)
(County)
(State)

UI
CODE

Proposed Ind.

Ind.

Mo.

Yr.

This report will be used to insure the proper industrial classification of your regular MONTHLY REPORT ON EMPLOYMENT,
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 1970.
Describe your processes or goods produced in your own words, making the distinctions requested on the list of special characteristics
provided on the enclosed sheet. This list is not complete but represents the kind of information which should be reported.
P rin cipa l P roducts

or

P ercent o r
T otal S ales
V alue or
R eceipts
D uring
1970

A ctivities D uring

19 7 0
(L is t item s separately)

P rin cipa l M aterials U sed
( F or each product listed in colum n ( a ) )

(b)
gg
M
j

(a )

1A. Manufacturing (Specify below)

W a s M aterial
U sed
P roduced
I n T h is
Est a bu sh m e n t ?
(d )

(c )
Yes

?

i

No

^
0
%
%
%
%
%
IB. Nonmanufacturing (Specify below)

%
,

3.

Is the establishment primarily engaged in
performing services for other units of the

%

company?

%

Yes □

No □

If “ Yes,” indicate nature of activity of this establishment:

%

□

Central administrative office

%

□

Research, development, or testing

□
£

. ' ----- - J

~_________

L ,.

IIS

Combined Total

2. Is this establishment part of a multiunit company ?
□ Yes □ No

100%
4.

Storage (warehouse)

□

Other (Specify: power plant, etc.)

Space for Your Comments.

If “Yes,” enter name and location of controlling company.

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




☆

U.S. GOVERNMENT PRINTING OPF1CE: 1970— 0 - 3 9 8 - 1 6 4

(P osition )

34

EMPLOYMENT, HOURS, AND EARNINGS

Technical References
Number

1. Armknecht, Paul A . Jr., “ BLS Establishment Estimates Revised to March 1970 Benchmark
Levels,” Employment and Earnings, September 1971, pp. 18-30.
2. “ The Calculation and Uses of the Spendable Earnings Series,” Monthly Labor Review, April
1966, pp. 405-409.
3. Mendelssohn, Rudolph C. “ Machine Methods in Employment Statistics,” Monthly Labor Re­
view, May 1955, pp. 567-569.
A description of the use of electronic data-processing equipment in the preparation of
employment statistics, with particular reference to screening employers’ reports for errors.
4.

President’s Committee to Appraise Employment and Unemployment Statistics, Measuring Em ­
ployment and Unemployment, 1962.
A comprehensive review and critique of the methods and concepts used by various Fed­
eral Government programs providing statistics on employment, unemployment, and the
labor force in the United States.

5. Nelson, Darrell R., “ BLS Earnings Statistics for Use in Escalator Agreements,” Employment
and Earnings and Monthly Report on the Labor Force, March 1968, pp. 18-21.
A discussion of the availability, uses, and limitations of BLS average hourly earnings
series in escalation agreements.
6.

Schechter, Samuel, “ The 1959 Benchmarks for the BLS Payroll Employment Statistics,”
Monthly Labor Review, December 1962, pp. 1385-1392.
A detailed description of the sources and construction of BLS employment benchmarks.

7.

Schwab, Paul M., “ Two Measures of Purchasing Power Contrasted,” Monthly Labor Review,
April 1971, pp. 3-14.
An examination of the divergent trends in real net spendable earnings and real per
capita disposable income.

8. Utter, Carol M., “ The Spendable Earnings Series: A Technical Note on its Calculation,” Em ­
ployment and Earnings and Monthly Report on the Labor Force, February 1969, pp. 11-21.
9. Weinberg, Edgar, “ BLS Earnings Series as Applied to Price Escalation,” Monthly Labor
Review, July 1952, pp. 57-59.
A discussion of the use of BLS average hourly earnings series in escalation clauses in
contracts.
10. Wymer, John P., “ The Revised and Expanded Program of Current Payroll Employment Sta­
tistics,” Employment and Earnings, November 1961, pp. iv-vii.
A description of the impact of a major benchmark adjustment and of important tech­
nical innovations on the industry employment statistics series.
11. Young, Dudley E. and Sidney Goldstein, “ The BLS Employment Series and Manufacturing
Reporting Practices,” Monthly Labor Review, November 1957, pp. 1367-71.
A discussion of the findings in a survey analyzing the response patterns of manufactur­
ing establishments cooperating in the industry employment statistics program.




Pa

u l

A . A

r m k n e c h t

, Jr .

Chapter 3. Job Vacancies and Labor Turnover
Background

Job vacancies, as used by the Bureau of La­
bor Statistics, refers to the number of unfilled
wage and salary jobs available to workers at
the end of a month. Labor turnover refers to
the gross movement of wage and salary work­
ers into and out of employment status with
respect to individual establishments during the
month.
The current job vacancy-labor turnover pro­
gram began in January 1969 with the addition
of questions about job vacancies to the labor
turnover questionnaire. The labor turnover
program, however, has been in existence for
many years.
In January 1926, the Metropolitan Life
Insurance Co. began the collection of labor
turnover data from a small sample of manu­
facturing establishments. By February 1927,
the sample included 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 man­
agers 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 sam­
ple 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
security agencies affiliated with the Bureau of
Employment Security (now the U.S. Train­




ing and Employment Service) had collected
labor turnover information for use in job
market analysis and as a guide for the op­
erations of the State employment services.
Cooperative arrangements between these agen­
cies 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
extended to cover all 50 States and the District
of Columbia.
Experimental programs to determine the
feasibility of collecting job vacancy informa­
tion were conducted in 1965 and 1966. The
1965 study was carried out in 16 metropolitan
areas and the 1966 study was conducted in 14
metropolitan areas and three States. Following
the successful completion of these programs,
cooperative arrangements for the collection of
job vacancy data were set up between the BLS
and 48 of the State employment services,
through the Manpower Administration.
In December 1970, these agencies published
about 8,000 labor turnover series in manu­
facturing and mining industries for State and
areas, and about 130 job vacancy area series in
manufacturing industries. These rates were
based on a sample of approximately 32,000
reports in manufacturing and about 1,200 in
mining.
Description o f the Survey

Labor turnover actions are divided into two
broad groups: accessions or additions to em­
ployment, and separations or terminations of
employment. These two broad groups are fur­
ther divided; accessions into new hires and
other accessions, and separations into quits,
discharges, layoffs, and other separations. La­
bor turnover is expressed in the BLS series as
a monthly rate per 100 employees. Separate
rates are computed for each of the component
items.
35

36

BLS HANDBOOK OF METHODS

The primary difference between types of
separations is whether action is initiated by
the employee or employer, i.e., whether it is
voluntary on the employee’s part or involun­
tary. Voluntary actions— quits— are initiated
by the employee for an almost unlimited va­
riety 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 em­
ployee; these actions may arise from economic
causes such as business conditions, physiolog­
ical reasons such as aging, or performance
reasons such as incompetence.
Job vacancy data are collected both for total
job vacancies and for vacancies which have
continued unfilled for 30 days or more. Data on
the occupations for which the vacancies exist
and the wage rates offered for them are also col­
lected quarterly. Vacancy data are expressed
as a monthly rate equal to job vacancies divided
by the sum of vacancies and employment, with
the quotient multiplied by 100.
Concepts

Separations are terminations of employment
of persons who have quit or been taken off the
rolls for reasons such as layoff, discharge, re­
tirement, 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 initi­
ated by employees for any reason except re­
tirement, 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 unauthorized absences
which, on the last day of the month, have lasted
more than 7 consecutive calendar days.
Layoffs are suspensions from pay status
(lasting or expected to last more than 7 con­
secutive calendar days), initiated by the em­
ployer without prejudice to the worker, for
reasons such as lack of orders, model change­



over, termination of seasonal or temporary
employment, inventory-taking, introduction of
labor saving devices, plant breakdown, or
shortage of materials.
Discharges are terminations of employment
initiated by the employer for such reasons as
incompetence, violation of rules, dishonesty,
laziness, absenteeism, insubordination, failure
to pass probationary period, etc.
Other separations include terminations of
employment for military duty lasting or ex­
pected to last more than 30 days, retirement,
death, permanent disability, failure to meet the
physical standards required, and transfers of
employees to another 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 1959).
New hires are permanent and temporary
additions to the employment roll of persons
who have never before been employed by the
establishment, and former employees rehired
although not specifically recalled by the em­
ployer. This category excludes 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
employment roll other than new hires.
Job vacancies are defined as vacant jobs
which are immediately available for filling, and
for which the firm is actively trying to find or
recruit workers from outside the firm.
“ Actively trying to find or recruit” means
that the establishment is engaged in current
efforts to fill the job vacancies.
Long-term job vacancies are those current
vacancies which have continued unfilled for 30
days or more.
The reporting establishment is also asked to
indicate the number of openings with future
starting dates for which the firm is actively
trying to recruit workers from outside the
firm. Job openings with future starting dates
may exist for such reasons as: Job unavailable
until expected separation of present incumbent

JOB VACANCIES AND LABOR TURNOVER

occurs; work will not start until some future
date; new branch to be opened in the future;
or anticipated increase in business.
Industry Classification

The classification system used for compiling
and publishing rates is that described in the
1967 Standard Industrial Classification Manual
issued by the Office of Management and Bud­
get. (See appendix B of this bulletin for a de­
tailed description of this system.)
Reporting establishments are classified on
the basis of major product or activity as de­
termined by annual sales data for the previous
calendar year. Most establishments in the job
vacancy-labor turnover sample also report em­
ployment, hours, and earnings under the Bu­
reau’s industry employment statistics program,
and are assigned the same industry classifica­
tion in both programs. Further discussion of
industry classification in the two programs is
given under the heading, Industrial Classifica­
tion in chapter 2 of this bulletin.
Occupational Classification

Occupational classifications are made in
accordance with those established in the Dic­
tionary of Occupational Titles, Third Edition,
U.S. Department of Labor, 1965. These classi­
fications are the same as those used in State
employment service operations.

Data Sources

Each month cooperating State employment
security agencies collect data on the number of
job vacancies and on labor turnover actions
from a sample of establishments drawn from
a list of those subject to State unemployment
insurance programs. In nonmanufacturing,
supplemental sources are also used to obtain
lists of establishments which are not covered
by Unemployment Insurance laws. (See chap­
ter 2, p. 17 of this bulletin.) The respondent ex­
tracts the figures largely from his personnel
records, though some smaller establishments
which do not maintain special personnel rec­
ords utilize their payroll records in making out



37

the reports. Response analysis surveys, which
analyzed the reporting practices of a scien­
tifically selected sample of the establishments
in the job vacancy-labor turnover panel,
showed that while some employers did not re­
port the figures for all items precisely as re­
quested on the schedule, the effect of these
deviations on the published data appeared to
be quite insignificant, particularly for the
broader classes, such as total accessions, total
separations, and total job vacancies.

Collection Methods

Job vacancy and labor turnover data are
collected primarily at the State level by em­
ployment security agencies from cooperating
employers via the medium of a mailed “ shuttle”
schedule, U.S. Department of Labor form 1219.
(See pp. 38 and 39 for a facsimile of this sched­
ule.) The same form is returned to the re­
spondent each month of the year for the entry
of current data. The respondent reports the
total number of job vacancies, the number of
actions for each turnover item during the cal­
endar month and total employment. These em­
ployment figures, which are the bases used to
compute the rates, represent the number of per­
sons who worked or received pay for any part
of the pay period (usually 1 week) which in­
cludes the 12th of the month.
Information on the occupations for which
job vacancies exist and the pay rate being
offered for them is collected quarterly on a
supplemental schedule, form DL 1219A.
The State agency uses the information pro­
vided on the schedule to develop job vacancy
and labor turnover rates for the States and for
metropolitan areas, and forwards the data to
Washington, where they are used by the Bu­
reau of Labor Statistics to prepare rates at
the national level.

Sampling

Sampling is used by BLS for collecting data
in its job vacancy-labor turnover statistics pro-

38

BLS HANDBOOK OF METHODS

D L 1219

Budget Bureau No. 44— R1064.
Approval expires January 31, 1972.
State

U.S. DEPARTMENT OF LABOR

Report No.

Ind.

Area

Local Off.

MONTHLY REPORT ON JOB OPENINGS
AND LABOR TURNOVER

B U R EA U O F L A B O R S T A T IST IC S
and
THE M A N P O W E R A D M IN IS T R A T IO N
W A SH IN G T O N , D.C. 20212

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

CHANGE NAME AND MAILING ADDRESS IF INCORRECT— INCLUDE ZIP CODE

r

“ I
The Bureau of Labor Statistics, the Manpower Administra­
tion, and the State agencies cooperating in their statistical
programs will hold all information furnished by the respond­
ent in strict confidence.

L

LOCATION ...................

J

Before entering data see explanations on other side

(C ity)

(County)

II. EMPLOYMENT
(one pay period)

I. LABOR TURNOVER DURING CALENDAR MONTH
YEAR
AND
MONTH

PERIOD COVERED BY
LABOR TURN OVER
(Cols. 4 through 11)
Preferably one calendar
month
From—
(B oth dates

(1 )
1970
Dec.

Through
inclusive)

(2 )

(3)

ALL E M P L O Y E E S

TOTAL NUMBER

S E P A R A T IO N S (during calendar month)
Total
Separations
(Sum o f cols.
5 through 8)
(4)

(State)

A C C E S S I O N S (during calendar month)

Quits

Discharges

Layoffs

Other
separations

( 5)

(6)

(7 )

(8 )

Total
Accessions
(Sum o f cols.
10 and 11)
(9)

New
hires

Other
Accessions

(10)

(11)

who worked dur­
ing or received pay
for any part of the
pay pericd which
include^ the 12th
o f the month
(12)

1971
Jan.
Feb.
M a r.
Apr.
M ay
June
July
A u g.
Sept.
Oct.
N ov.
Dec.

III. JOB OPENINGS for which workers from outside the firm
were actively being sought as of close of last business day of
most recent month. (If “ NONE,” enter “0,” do not leave blank.)
YEAR

AND
M TH
ON

C R E T JOB OPENINGS
URN
OPENINGS W
ITH
(unoccupied and ready for imme­
FTR
UUE
STA TIN
R G
diate filling)
DTS
AE
Openings included in

DO
N
OT
UE
S

col. 13 continuing
unfilled for 30 days
or more
(14)

Expl.
Code
(16)

Number o f current
Job Openings
(13)

(not included in
column 13)
(15)

IV. YOUR COMMENTS
Enter main factors responsible for any significant month-tomonth changes in SECTIONS I, II, and III. Examples are:
More business
Strike
Fire

Weather
Temporary summer
help
Seasonal Increase
(17)

1970
Dec.
1971
Jan.
Feb.
M a r.
A p r.
M ay
June
July
A u g.
Sept.
Oct.
N ov.
Dec.

( Person to be addressed if questions arise regarding this rep ort)




(Position)

39

JOB VACANCIES AND LABOR TURNOVER

INSTRUCTIONS FOR COMPLETING THIS FORM
I.

LABOR TURNOVER

12th of the calendar month for which labor turnover data are
reported.

PERIO COVERED.—Information on labor turnover, cols. 4 through
D
11, is requested for the most recent entire calendar month speci­
fied in column 1 or, if this is not possible, for a period which
most closely covers that calendar month. In either case enter in
cols. 2 and 3 the beginning and ending dates for the monthly
period for which turnover data are reported.

SEPARATIONS (A L E PLO E S)
L M YE
Column 4. TO
TAL SEPARATIONS DU G CALEN
RIN
DAR MONTH.,- Enter
in column 4 the sum of columns 5 thru 8.
Column 5. QUITS.— quit is a termination of employment ini­
A
tiated by the employee for any reason except to retire, to trans­
fer to another establishment of the same firm, or for service in
the Armed Forces. Include a person who fails to report after be­
ing hired (if previously counted as an accession) and an unau­
thorized absence if on the last day of the month the person has
been absent more than 7 consecutive calendar days.
Column 6. DISCH
ARGE.— discharge is a termination of employ­
A
ment initiated by the employer for such reasons as incompetence,
violation of rules, dishonesty, laziness, absenteeism, insubordina­
tion, failure to pass probationary period, etc. Inability to meet
organization’s physical standards should be reported in other
separations, col. 8.
Column 7. LAYOFFS.— layoff is a suspension from pay status
A
(lasting or expected to last more than 7 consecutive calendar
days without pay) initiated by the employer without prejudice
to the worker for such reasons a s: lack or orders, model change­
over, termination of seasonal or temporary employment, inven­
tory-taking, introduction of labor saving devices, plant break­
down, shortage of materials, etc.; include temporarily fur­
loughed employees and employees placed on unpaid vacations.
Column 8. O E SEPARATIONS.—
TH R
This group should include only
terminations of employment for military duty lasting or ex­
pected to last more than 30 calendar days, retirement, death,
permanent disability, failure to meet the physical standards re­
quired, and transfers of employees to another establishment of
the company. N O T E : If you include any other types of separa­
tions in this column, mention the number and type under Com­
ments. Employees involved in labor-management disputes should
not be counted as separations.

ACCESSIONS (A E PLO E S)
LL M Y E
Column 9. TO
TAL ACCESSIONS DU G CALEN
RIN
DAR MONTH.-An ac­
cession is any permanent or temporary addition to the employ­
ment roll whether of new or former employees, or transfers from
another establishment of the company. Enter in column 9 the
sum of cols. 10 and 11.
Column 10. N
EW H E New hires are temporary or permanent
IR S.—
additions to the employment roll of (1) anyone who has never
before been employed in this establishment, or (2) former em­
ployees you did not call back. Former employees you did call
back should be included in total accessions and “other acces­
sions,” but not in “ new hires.” Persons transferred from other
establishments of this company should be reported in “ other
accessions.”
Column 11. O E ACCESSIONS.—
TH R
Include all additions to the em­
ployment roll other than new hires. This includes all employees
called back to work by the employer from a layoff as defined for
col. 7, transfers from other establishments of the company, and
former employees returning from military leave or other ab­
sences without pay who have been counted as separations. Em­
ployees involved in labor-management disputes should not be
counted as accessions when they return to work.

II.

EMPLOYMENT

PERIO COVERED.—Employment information, col. 12, is requested
D
for one pay period (preferably one week) which includes the




☆

Column 12. A L EM YEE Enter the total number of persons
L
PLO S.—
(both sexes) on the payrolls of the establishment (s) covered in
this report who worked full- or part-time or received pay for
any part of the pay period (preferably one week). Include
salaried officers of corporations, executives and their staffs, and
employees engaged in a force-account construction but exclude
proprietors, members of unincorporated firms, and unpaid
family workers. Include persons on vacations and sick leave for
which they received pay directly from your firm for the pay
period covered 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 pay
period covered.
N O TE : If the number differs from the “ All Employees” total
reported on the Employment, Payroll, and Hours form, explain
under Comments.

III.

JOB OPENINGS

PERIO COVERED.—
D
Job openings information, columns 13, 14 and
15, is requested as of close of last business day (or nearest pos­
sible day) of the most recent monthly period for which labor
turnover data are reported.
Column 13. N M
U BER O C R E T JOB OPENINGS (VACANCIES).-Enter
F URN
the number of current job openings in your establishment. IF
“ N O N E ,” E NTER “0.” A current job opening is an existing
vacant job in your establishment that is immediately available
for filling and for which your firm is actively trying to find or
recruit some worker from outside your firm (i.e., a “new”
worker— not a company employee).
Include such openings for all kinds of positions, classifications
and employment (full-time, part-time,' permanent, temporary,
seasonal), including those outstanding on orders with employ­
ment agencies and notifications to unions.
Exclude jobs to be filled by recall, transfer, promotion, demo­
tion or return from paid or unpaid leave; jobs unoccupied be­
cause of labor-management disputes; job openings for which
“ new” workers were already hired and scheduled to start work
later; and the “ openings with future starting dates” reported
in column 15.
“ Actively trying to find or recruit” means current efforts to fill
the job opening through orders listed with public or private em­
ployment agencies and school placement offices; notifications to
labor unions and professional organizations; “ help wanted” ad­
vertising (newspaper, posted notice, etc.); recruitment pro­
grams ; interview and selection of applicants.
Column 14. C R E T JOB OPENINGS CONTINUING U F L D FO 30
URN
N IL E
R
DAYS OR M
ORE.—
Enter the number of current job openings in­
cluded in the figure reported in column 13 which have continued
unfilled for 30 days or more. IF “ NO N E ,” E N TER “ 0.”
Column 15. OPENINGS W
ITH F T R STARTIN D TE Enter the
UUE
G A S.—
number of openings in your establishment for which your firm
is actively trying to find or recruit some worker from outside
your firm (i.e., a “new” w orker), but which relate to jobs that
r
are currently occupied or unavailable for immediate occupancy
by “ new” workers for such reasons as: job unavailable until ex­
pected separation of present incumbent occurs: work will not
start until some future date; new branch to be opened in future
or anticipated increase in business. IF “ N O N E ,” E N T ER “ 0.”
N O T E : OPENINGS W IT H FUTURE STARTING DATES
ARE NOT TO BE INCLUDED IN COLUMN 13.

IV. COMMENTS
Column 17. YOUR COM ENTS.—Enter the main factors responsible
M
for significant month-to-month changes in Labor Turnover <cols.
4 through 11), Employment (col. 12), and Openings (cols. 13
through 15). Some examples are listed in the heading of col­
umn 17.

U.S. G O V ER N M EN T PRINTING OFFICE: 1970-0-407-201

40

BLS HANDBOOK OF METHODS

gram, since full coverage would be prohibi­
tively costly and time consuming. The sampling
plan for the program must: (a) provide the
preparation of reliable monthly estimates of
job vacancy and 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 pro­
cedures, administrative practices, resource
availability, and other institutional charac­
teristics of the program.
In developing the sample design, the universe
of establishments was stratified first by in­
dustry 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 average size of employment
within each of the strata. The total size of
sample regarded as necessary to produce satis­
factory estimates of employment was dis­
tributed among the size cells on the basis of
average employment per establishment in each
cell. In practice, this is equivalent to distribut­
ing the predetermined total number of estab­
lishments 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 establish­
ments fall into the sample with certainty.
Establishments with 250 or more employees
are included in the sample with certainty, al­
though in some cases the cutoff is lower. The
sizes of the samples for various industries were
determined empirically on the basis of exper­
ience.
The sample design, although aimed primarily
at meeting the needs of the national program,
provides a technical framework within which
State and area sample designs can be de­
termined. Since, however, the rates for States
and areas are not generally prepared at the
same degree of industry detail as the national
1 For the national sample, additional reports needed for
State and area samples are added to those required by the
national design.




rates, the national design usually provides
sufficient reports for the preparation of State
and area rates.1

Estimating Procedures

Both job vacancy and labor turnover rates
are estimates of ratios. For individual indus­
tries, turnover rates are computed by dividing
the number of turnover actions of each type,
as reported by the sample establishments, by
the total number of employees reported by
those establishments. 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 January 11-17. The Jan­
uary quit rate for the industry is :
623
30,062 X 100- 2*
1
Turnover rates for industry groups are com­
puted by weighting the rates for the component
industries by the estimates of total employ­
ment, prepared by the BLS industry employ­
ment statistics program. These estimates,
which cover the pay period including the 12th
of the month, are described in chapter 2 of
this bulletin. Rates for “ all manufacturing”
and for the durable and nondurable goods sub­
divisions of manufacturing are weighted by
employment in the major industry groups.
Computation of job vacancy rates for indus­
try stratum also involves a weighting process.
The number of vacancies reported by the sam­
ple of establishments is weighted by the esti­
mates of total employment in that industry
stratum. The weighted number of vacancies is
then divided by the sum of employment in the
stratum plus vacancies; the quotient is then
multiplied by 100 to determine the vacancy
rate. Rates for major industry groups, for the
durable goods and nondurable goods sub­
divisions, and for total manufacturing are
computed by summing the weighted number of
vacancies in the component cells of industries
or subdivisions, then dividing by the sum of
vacancies and employment in that industry,

JOB VACANCIES AND LABOR TURNOVER

subdivision or division. Again, the quotient is
multiplied by 100 to obtain the rate.
As of 1971, size stratification was not used
in the preparation of job vacancy and labor
turnover rates. Tests were underway to de­
termine the effect on the rates of introducing
size stratification. Preliminary results of the
tests suggested that size stratification would
improve the labor turnover rates but have little
effect on the job vacancy rates. Upon comple­
tion of the tests, size stratification will be in­
troduced, if warranted.
Seasonally Adjusted Series

Many economic statistics, including labor
turnover 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 variation, it is possible to
observe the cyclical and other nonseasonal
movements in these series. Seasonally 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-to-moving average method, with a pro­
vision for “ moving” adjustment factors to take
account of changing seasonal patterns. A de­
tailed description of the basic method is given
in appendix A of this bulletin.
The 2 years of experience with job vacancy
data indicate the emergence of some probable
seasonal patterns. However, a minimum of 3
years of data are needed to develop estimates
of the seasonal factors.

Presentation

The BLS publishes, on a national basis,
monthly series of labor turnover rates for se­
lected industries. These series are currently
published for the manufacturing division, the
durable and nondurable goods subdivisions, 21
major industry groups in manufacturing, 191
individual manufacturing industries, and 7
categories in mining and communications.
Rates are available for all manufacturing from



41

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.
Rates for certain highly seasonal industries,
for example canning and preserving, are not
now published separately but are included in
the computation of rates for the major manu­
facturing groups. Before 1958, these industries
and the printing, publishing and allied indus­
tries major group were not included in the
rates for all manufacturing. The rates for all
manufacturing for years prior to 1958 were
revised, however, to reflect the influence of
these industries.
Monthly rates for total accessions, new hires,
total separations, quits, and layoffs are shown
for manufacturing and mining industries. Ex­
cept for the new hire rates, the same items are
published for the telephone and telegraph in­
dustries.
On a national basis, the BLS presently pub­
lishes monthly estimates of total and long-term
job vacancies for the manufacturing division,
the durable goods and nondurable goods sub­
divisions, and nine selected industry groups.
Rates are available from April 1969 for each
of these divisions and industry groups.
Preliminary job vacancy rates for the nine
selected major industry groups and turnover
rates for the 21 major industry groups in
manufacturing are published monthly in a
BLS press release about a month after the
reference month, and in the Monthly Labor
Review 3 months after the reference month.
Preliminary turnover rates for both detailed
industries and broad categories and prelimi­
nary job vacancy rates for the same categories
as are included in the press release are pub­
lished in Employment and Earnings about 2
months after the month of reference.
Both job vacancy and labor turnover rates
for all manufacturing for selected States and
metropolitan areas are published each month
in Employment and Earnings. More detailed
information is 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 annual volume

42

BLS HANDBOOK OF METHODS

called Employment and Earnings, United
States. New editions of this volume are pub­
lished annually, following each adjustment of
the Bureau’s industry employment statistics
series to new benchmark levels.

Uses and Limitations

The two major causes of change in both job
vacancy and labor turnover rates are industrial
expansion and contraction. In prosperous
times, job vacancy rates, quit rates, and new
hires are high because of job availability; in
periods of economic recession, high layoff rates
are coupled with low vacancy, quit, and acces­
sion rates. Turnover rates are, therefore, re­
garded as good economic indicators and are
widely used by economic analysts in both gov­
ernment and private industry. Together with
the turnover data and other labor force data,
such as unemployment rates, job vacancy data
are also expected to provide an indicator as to
the condition of the economy.
Labor turnover rates by industry are also
valuable for personnel planning and analysis.
Employers 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 opera­
tions and good labor-management relations. A
consideration of turnover is essential for sched­
uling production and for planning the orderly
recruitment and maintenance of an adequate
manpower supply. Labor turnover rates are
also widely used by State employment services
to plan and appraise their operations.
Job vacancy data should also prove useful to
State employment services. By identifying
emerging labor shortages, vacancy data will




allow for more intelligent planning of training
programs and should be useful in counselling
the unemployed and new entrants to the labor
market. The data may also make interarea re­
cruitment of workers possible.
The use of turnover or job vacancy rates to
interpret changes in the BLS monthly employ­
ment series is limited for the following rea­
sons: (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 job
vacancy and labor turnover rates, which are
computed as the arithmetic 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 unusual
or affected strongly by seasonal influences.2
— S

h e il a

C.

W

h it e

Bureau

2 Because they are liable to misinterpretation, the
does not prepare cumulative annual rates of labor turnover.
For example, an annual quit rate could be obtained b y divid­
ing the total number of quits during the year b y average
employment during the year. An approximation of this figure
can be obtained by cumulating the 12 monthly rates. Suppose
the annual rate thus obtained amounted to 50 per 100 em ­
ployees. This might seem to im ply that 50 percent of all
employees in January voluntarily left their jobs b y 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 information on the number o f em ­
ployees who remained with the establishment during the en­
tire year. Over short periods of time, labor turnover rates
probably include relatively little repetitive counting of em­
ployees w ho have held the same jobs, while over a period of
as long as a year there is considerable duplication.

Manpower Structure and Trends
Chapter 4. Employment of Scientific and Technical Personnel
Background

The growth in industrial employment of
scientists and engineers has been both a symp­
tom and a driving force in the rapid technolog­
ical progress of American industry. For a
number of years, a major element in this
growth has been the need of the Federal Gov­
ernment for increasingly complex and tech­
nologically advanced national defense material
and space related research and technology. The
surveys of scientific and technical personnel
provide a means for estimating current man­
power resources and a basis for evaluating
future requirements for scientific and engineer­
ing personnel.
The surveys of scientific and technical per­
sonnel were initiated by the Bureau of Labor
Statistics in 1959, under the sponsorship of
the National Science Foundation. They are
part of a comprehensive statistical program,
coordinated by the Foundation, designed to
yield estimates of the employment of scientists
and engineers in all sectors of the U.S. eco­
nomy. The Bureau of Labor Statistics con­
ducted surveys of scientific and technical
personnel in industry annually during the 1960’s
(except for 1965) and related surveys of State
government agencies less frequently. Since
1962, the Bureau of Labor Statistics has had
full responsibility for the conduct and publica­
tion of these surveys, which continue to be
planned in consultation with the National Sci­
ence Foundation. The last survey of scientific
and technical personnel was conducted in 1970.
In the future, estimates of scientific and tech­
nical personnel employed in private industry
will be developed as part of the Bureau’s Oc­
cupational Employment Statistics Program,
which will provide employment statistics on
all occupational fields.

cialty. Engineers and scientists are further
distributed according to whether they are en­
gaged primarily in research and development
activities, in management and administration,
in technical sales and service, in production
and operations, or in “ all other” functions.
Data are published separately for all major
industry groups, and in finer industry detail
for a selected number of major industry groups.
Beginning in 1966, data were developed by
geographic area.
State agencies are surveyed for information
on employment of engineers, scientists, tech­
nicians, economists, statisticians, psychologists,
social workers, and health professionals. Em­
ployment is tabulated by the various govern­
mental functions in which the workers are
engaged. Data also are published by State when
appropriate.
For each professional occupation, respond­
ents are asked to report the number of persons
whose current positions require knowledge
equivalent at least to that acquired through
completion of a 4-year college course with an
appropriate academic major, regardless of
whether they hold a college degree. These sur­
veys, thus, cover all persons actually working
in one of the designated occupations, specifi­
cally including those who do not hold an ap­
propriate degree or any degree at all, but
specifically excluding persons trained in the
occupation but currently employed in positions
not requiring the use of such training.
Technicians are defined as persons actually
engaged in technical work at a level which re­
quires knowledge of engineering, mathematical,
physical and life sciences, comparable to that
acquired through technical institutes, junior
colleges, or other formal post-high school train­
ing less extensive than 4-year college training,
or through equivalent on-the-job training or
experience.

Description o f Surveys

The surveys of scientific and technical per­
sonnel in industry gather data on the employ­
ment of engineers as a group, and on scientists
and technicians by major occupational spe­



Data Sources

Sources of occupational data reported by re­
spondents are personnel records and especially
for the small reporting units, personal knowl­
43

44

BLS HANDBOOK OF METHODS

edge of persons completing the reports. Dis­
cussions with a number of large respondents
indicate that their records typically 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.
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 unemploy­
ment insurance program. The survey of State
government agencies is based on a compilation
of data from all State government agencies
employing personnel in occupations covered by
the survey.

Collection Methods

Data are collected from respondents pri­
marily by mail, but personal visits are made to
many large employers, and to other respond­
ents who indicate particular difficulty in com­
pleting the questionnaires. These visits, which
limited resources have prevented from being
either numerous or frequent, are carried out by
senior staff members. Normally two mailings
follow and a sub-sample of residual nonre­
spondents are contacted further by telephone.
The response to these surveys has been very
encouraging. Respondents supplying usable in­
formation have constituted approximately 80
percent of the reports solicited in virtually
every year, and have never been below 75
percent.

Sampling

The basic sample for the industry survey was
drawn from lists of establishments reporting to
each of the 51 (State and D.C.) employment
security agencies for unemployment insurance
(UI) purposes, and was supplemented by a list
of interstate railroads and related companies
supplied by the Interstate Commerce Commis­
sion. Industry classification of establishments is



based on information available to the State
agencies.
Certain categories of establishments are
eliminated from the master list before the
sample is selected, either because a separate
survey of the given category is being made or
because the number of scientific and technical
personnel employed are believed to be negligi­
ble. The categories of organizations omitted are
those classified according to the Standard In­
dustrial Classification system in the following
major industry groups: 01 and 02— farms; 80
— medical and other health services (except
807, medical and dental laboratories, which is
included); 82— educational services; 84— mu­
seums, art galleries, and botanical and zoo­
logical gardens; 86— nonprofit membership
organizations; 88— private households; 89—
miscellaneous services (except 8911, engineer­
ing and architectural service, which is in­
cluded); 91 through 94— government; and 99
— nonclassifiable establishments.
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
establishments in the small size groups, mini­
mum size cutoffs are essential to the efficiency
of the survey. Since excluded establishments
employ very few scientists, engineers, or tech­
nicians, survey results are affected little by
these omissions.
Sample numbers are allocated among the
various industry-size strata according to the
principle of optimum allocation; expected re­
sponse rates by industry and by size are taken
into account 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 number of scientists and engineers for
all industries combined is about 5 percent. In
every covered industry, all establishments with
1,000 employees or more are included in the
sample. In other industry-size cells, the sam­
pling ratios range from 1 in 1 to 1 in 100. In
general, the larger the establishment 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.

EMPLOYMENT OF SCIENTIFIC AND TECHNICAL PERSONNEL

Since scientific and engineering employment
is concentrated to a significant degree in re­
search and development laboratories, not sep­
arately identifiable in the UI universe, the
probability sample was supplemented. The
essential rule for unbiased supplementation is
that the supplementary units must be drawn
independently of the probability sample; that
is, the chance that a unit is drawn in the prob­
ability sample must be independent of the
chance that the unit is selected as a supplement.
The initial supplementation was drawn from a
list of industrial research laboratories compiled
by the National Academy of Science— National
Research Council and from a list of small busi­
ness concerns interested in performing research
and development compiled by the Small Busi­
ness Administration. Beginning in 1966, when
a new sample was drawn, supplementation 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. Establishments
selected as supplements are tabulated as a sep­
arate cell within their industry and size class
with a weight of 1, regardless of whether they
are also members of the probability sample.
(See Estimating Procedures.)
The mailing list for 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 oc­
cupations. The agencies are identified from in­
formation in directories and other documents
furnished by the States.

Estimating Procedures

For the survey of scientific and technical
personnel in industry, 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 supplemental units.1
Estimates for supplemental cells are obtained
by summing the primary item employment for




45

the supplemental reports plus an estimate for
nonrespondent supplementary units.
Estimates for the survey of scientific and
technical personnel employed by State govern­
ment agencies are obtained by summing the
reported data. The response rate in this survey
is extraordinarily high— 96 to 98 percent— and
examination of the nonrespondents shows that
the number of scientists, engineers, and other
personnel employed by them is negligible.

Analysis, Interpretation, and
Presentation

A report on the findings of each survey is
published, usually within 2 years of the refer­
ence date of the survey. Each report consists of
an analytical interpretation of the findings,
and is supported by a statistical appendix con­
taining in tabular form all of the data that can
be meaningfully derived from the survey.

Uses and Limitations

Data from these surveys form the essential
statistical base (1) for evaluating the adequacy
of scientific and technical manpower resources
of the United States in light of current or
projected demands and (2) for determining the
rate of growth of these resources. They have
been used to evaluate the impact of new or en­
larged Federal programs calling for substantial
scientific and technical manpower. These data
also provide the bases for projections of future
manpower requirements in science and engi­
neering. For example, data are furnished for
occupational guidance counselors and others
who provide young people with information on
which to base a choice of career.
These estimates must be interpreted as ap­
proximations. All surveys are subject to possi1 Symbolically, P '= - ? £ i
.
v e ( M, where M is the cell universe total
employment, v p ( is the sum of the primary item employment
of the cell respondents,
t is the sum of total employment
of the cell repsondents, and P' is the estimate. M is adjusted
to prevent duplicate estimation for supplemental sample re­

porters.

46

BLS HANDBOOK OF METHODS

ble response and processing errors, although
these are reduced insofar as possible, through
checking procedures and through correspond­
ence with reporters whose data are internally

inconsistent or appear to involve misinterpreta­
tions of definitions or other instructions. In
addition, estimates derived from sample sur­
veys are limited by sampling error.

Technical References
Number

1. U.S. Department of Labor, Bureau of Labor Statistics. Scientific and Technical Personnel In
Industry, 1961-66 (Bulletin 1609, 1968).
2. -------------- Scientific and Technical Personnel in Industry, 1967 (Bulletin 1674, 1970).
3 . ------------- Employment of Scientific, Professional, and Technical Personnel In State Govern­
ments, January 196U (Bulletin 1557, 1967).
4 . ------------- “ Scientific and Professional Employment By State Governments” (Reprinted from the
August 1969 issue of the Monthly Labor Rview.)
5. -------------- Employment of Scientific and Technical Personnel in State Government Agencies
(BLS Bulletin 1412, 1964).
6.

National Science Foundation, Employment of Scientific and Technical Personnel inState Gov­
ernment Agencies, Report on a 1959 Survey. N SF 61-17 (1961).

7. -------------- Reports prepared for the National Science Foundation by the U.S. Department of
Labor, Bureau of Labor Statistics: National Science Foundation, Scientific and Technical
Personnel in American Industry, Report on a 1959 Survey. N SF 60-62 (1960).
8. _________ National Science Foundation, Scientific and Technical Personnel in Industry, 1960.
N SF 61-75 (1961).
9. _________ National Science Foundation, Scientific and Technical Personnel in Industry, 1961.
N SF 63-32 (1963).




— M ic h a e l F . Cr o w l e y

Chapter 5. Occupational Outlook
Background

The occupational outlook program originally
stemmed from a report of the Advisory Com­
mittee on Education appointed by President
Roosevelt. This Advisory Committee recom­
mended, in 1938, that an occupational outlook
service be set up in the Bureau of Labor Sta­
tistics to make studies and provide information
for use of individuals in choosing a career, and
for the use of those responsible for the plan­
ning of education and training programs. In
1941, the Occupational Outlook Service was
organized under a specific authorization by the
Congress. Although the first, preliminary stud­
ies 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 prep­
aration of occupational reports for use in guid­
ance. In mid-1946, a manual of occupational
outlook information was prepared for use in
the Veterans Administration (VA) counseling
and rehabilitation program.
In response to a resolution by the National
Vocational Guidance Association, calling upon
the Congress to authorize this type of informa­
tion 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 re­
leased in 1951.
Following the conclusion of the Korean hostil­
ities, there was a sharp increase in public
recognition of the key role of vocational guid­
ance in staffing essential occupations and
effectively utilizing the Nation’s manpower
resources. This resulted in the Congress in 1955
providng for the mainteance of the Occupa­
tional Outlook Handbook and its related publi­
cations on a regular, continuing, up-to-date
basis. In 1957, the third edition of the Occupa­
tional Outlook Handbook was published; also
in that year, the Occupational Outlook Quar­
terly was originated as a companion piece to




the Handbook. The 1957 Handbook was fol­
lowed in due course by the 1959, 1961, 1963-64,
1966-67, 1968-69, and 1970-71 editions of the
Handbook.

Description o f Program

Under the occupational outlook program, the
Bureau of Labor Statistics conducts research
in, and provides information on, future occupa­
tional and industry manpower requirements
and resources. It provides vocational guidance
information on expected employment oppor­
tunities 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 national training
authorities and policymakers for use in de­
veloping programs of education and training.
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 21/2 decades of industry and occupa­
tional research, the occupational outlook pro­
gram has systematically accumulated and
analyzed considerable manpower information
on such topics as employment trends for major
industries of the economy and for most major
occupations; employment effects of a great
many long-term programs of government agen­
cies, including those for defense, highways,
scientific research, space technology, 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
1 “ America’s Industrial and Occupational Manpower Re­
quirements, 1964-75,” prepared for the National Commission
on Technology, Automation, and Economic Progress, pub­
lished in Technology and the American Economy, Appendix
Vol. I, February 1966, and The U.S. Economy in 1980, (BLS
Bulletin 1673, 1970).

47

48

BLS HANDBOOK OF METHODS

year, hundreds of different, detailed occupa­
tional and industry statements are published in
the Occupational Outlook Handbook. In most of
these occupational outlook statements, informa­
tion is provided on: nature of work; places of
employment, education and training require­
ments; employment outlook for about 10 years
ahead, including, in most cases, estimates of
annual requirements for growth and replace­
ment needs; and earnings and working condi­
tions. In presenting outlook statements for
industries, information is included on nature
and location of each industry and other indus­
try characteristics, 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 poten­
tial supply of workers from many sources—
schools and other training institutions, trans­
fers 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 competition in a specific field
facing young people in the years ahead.
In addition to the overall and detailed indus­
try and occupational projections developed for
the Handbook and described earlier, special
manpower studies are prepared, as part of the
occupational outlook program, that provide in­
formation, narrower in scope and greater in
depth, on the changing industrial structure and
2 Employment and Changing Occupational Patterns in the
Railroad Industry, 1947-60 BLS Bulletin 1344 (1963); Em ploy­
ment Requirements and Changing Occupational Structure in
Civil Aviation BLS Bulletin 1367 (1964); and Employment
Outlook and Changing Occupational Structure in Electronics
Manufacturing BLS Bulletin 1363 (1963).
3 Maxine G. Stewart, “A New Look at Manpower Needs in
Teaching,” M onthly Labor Review, June 1964, pp. 639-644;
Technician Manpower: Requirements, Resources, and Train­
ing Needs BLS Bulletin 1512 (1966); Joe L. Russell, ‘‘Chang­
ing Patterns in Employment of Nonwhite Workers,” Monthly
Labor Review, May 1966, pp. 503-509; "Job Hopes High
for Record College Class,” U.S. Department of Labor, Press
Release, June 5, 1966 (USDL 7246), and College Educated
Workers, 1968-80, BLS Bulletin 1676 (1970).
4 Allan F. Salt, “Estimated Need for Skilled Workers,
1965-75.” M onthly Labor Review, April 1966, pp. 365-371;
Joseph F. Fulton, “Employment Impact of Changing Defense
Programs,” Monthly Labor Review, May 1964, pp. 508-516;
Bernard Yabroff, “ Trends and Outlook for Government Em­
ployment,” Monthly Labor Review, March 1965, pp. 285291, and Max Carey, “ The Crafts— Five Million Opportunities,”
Occupational Outlook Quarterly, Spring 1971, pp. 2-11.




occupational composition of American indus­
tries, such as railroads, civil aviation, and elec­
tronics.2 Other special studies provide more
technical information and quantitative projec­
tions of manpower requirements and resources
in specific occupations, industries, or for speci­
fic groups of workers— such as teachers, tech­
nicians and nonwhite workers— which include
consideration of the current and future de­
mand-supply relationships and their implica­
tions.3 Still others discuss only manpower
requirements trends and projections, especially
in those occupational groups where the supply
of workers is difficult to estimate, such as
skilled workers and workers in defense-related
employment.4

Sources o f Data

The projections and other manpower infor­
mation developed in the occupational outlook
program utilize a wide variety of data sources,
which vary mainly with the particular occupa­
tion or industry under examination. The fol­
lowing sections indicate some of the major
sources of statistical and other information
utilized in the program.
The basic statistics on current and past em­
ployment in occupations and industries have
been based mainly on Bureau of Labor Statis­
tics household data from the Monthly Report
on the Labor Force (MRLF) and establishment
data from Employment and Earnings. (A
single publication starting with the February
1966 issue). Use is made also of the scientific
and technical personnel surveys conducted by
the Bureau, which contain data on scientists,
engineers, and technicians. The decennial Cen­
sus of Population is utilized for data on most
occupations not covered by the limited detail
published in the MRLF and Censuses of Busi­
ness and Manufacturing are used to fill in in­
dustry detail. Information from the Civil
Service Commission is used for data on Federal
Government workers. These basic sources of
occupational and industry employment statis­
tics are augmented by data from Federal reg­
ulatory agencies, such as the Federal Aviation
Agency and Interstate Commerce Commission,

49

OCCUPATIONAL OUTLOOK

which collect industry and occupational sta­
tistics. In some cases, employment data are
obtained from unions, industries, trade associa­
tions, 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 par­
ticular 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 major sources of the
statistics used are Bureau of Labor Statistics
studies of productivity and technological de­
velopment, the Federal Reserve Board produc­
tion indexes, and the U.S. Department of
Commerce output data from the Annual Survey
of Manufactures and the Census of Manu­
factures. Industry associations and unions also
often provide similar types of data.
Estimates of the past and probable future
supply of workers utilize entirely different
sources of information. U.S. Office of Education
data on enrollments and degrees in high school,
post-secondary schools, and colleges and uni­
versities form a major component of the supply
estimates. Bureau of Apprenticeship and Train­
ing statistics on apprenticeship, as well as
information on company training programs,
provided through company reports and per­
sonal interviews, provide other inputs into the
supply estimates.
Special studies of various aspects of the sup­
ply of workers provide essential information
for the development of the estimates and
projections of supply. A few examples of these
types of statistical source materials are occu­
pational mobility studies (Bureau of Labor
Statistics); Tables of Working Life (Bureau
of Labor Statistics); followup studies of col­
lege graduates (National Science Foundation);
and many other specific types of studies, often
of a one-time nature. Earnings information,
which appears in many of the outlook publica­
tions, is drawn primarily from wage and earn­
ings surveys conducted by the Bureau of Labor
Statistics, supplemented with additional infor­
mation 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 sta­
tistics used is information obtained from (1)
personal interviews with employers or others
closely associated with an industry or occupa­
tion; (2) reports and interviews with profes­
sional or trade associations and licensing
agencies; (3) union publications and officials;
and (4) periodicals, trade journals, annual re­
ports, and so on.

Methods o f Analysis

The projections of requirements and re­
sources developed for the occupational outlook
program require varying methods of analysis,
usually because of differences in the factors
affecting a particular 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, comprehensive 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 factors may operate in the future.
Occupational employment is affected by a host
of factors. Technological change is the most
often discussed factor affecting occupational
employment, but occupational changes are also
influenced by other factors, such as growth in
population and its changing age distribution,
as in the case of teachers. Government policy—
relating, for example, to the magnitude of the
defense and space programs and to expendi­
tures 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 occupa­
tions, as for example, the substitution effect
resulting from shortages of engineers and their
replacement by technicians. Also influencing
occupational employment are changes in the

50

BLS HANDBOOK OF METHODS

total demand for the employing firm’s product;
changes in the level of income and distribution
of income among consumers, industry and gov­
ernment; and changing patterns of consump­
tion.
It is apparent, in view of this multitude of
factors, that no one technique can be used
successfully to project manpower requirements
in all occupation 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 (related 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 depend upon
education practices and available financing.
Projections of requirements for scientists,
engineers, and technicians require considera­
tion of different factors. These must take into
account such factors as the growing utilization
of technical personnel, the increasing tech­
nological complexity of industrial products and
processes, changes in levels of expenditures for
defense, and growing research and develop­
ment 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 popula­
tion and urbanization; for truckdrivers, to im­
proved equipment and highways, and for
competing methods of transportation.
For many occupations, the significant factors
influencing employment are the prospective
levels of demand for the products of the various
industries in which the occupation is found,
and the effect of these changes in demand on
employment in the industries. Among the gen­
eral 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 services, expected tech­
nological changes, output per man-hour, and
changes in hours of work. More specifically, in
projecting the activity or production level of
an individual industry, it is necessary first to
establish the nature of the demand for an in­



dustry’s products or services and the relation­
ship of this industry to the growth of the whole
economy. Obviously, an industry producing
products directly for consumers will have a dif­
ferent type of demand function than an indus­
try 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 output per capita. The total require­
ments for steel depend on the requirements
projected for each of the principal steel-using
industries, such as the automobile, construc­
tion, electrical appliances, machinery, and con­
tainers 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 output 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 output for each industry, and
by making assumptions as to changes in hours
of work.
Because of the tremendous amount of re­
sources necessary to make an extensive study
of each industry in the economy, a more global
type of analysis has been used to fill the gaps
and to provide an overall framework for the
occupational and industry projections. In mak­
ing the analysis for the products of each indus­
try, 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 production can
then be translated into requirements for work­
ers in total and by occupation.
The general approach in the development of
this industry framework is to begin with the
population 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 un­
employment, annual hours of work, and output
per man-hour. Multiple correlations are made
which take into account past employment

OCCUPATIONAL OUTLOOK

trends and relationships, and variables such as
unemployment, size of the Armed Forces, gross
national product, and population. By this tech­
nique, preliminary projections of manpower
requirements are developed for each industry
for which adequate data are available.
The results of the multiple regression analy­
sis are then used as the basis for further judg­
ment decisions as to the level of manpower
requirements in the projected period, par­
ticularly for those industries for which detailed
industry studies have been made. The consider­
able amount of information on individual in­
dustries developed in the occupational outlook
program over many years and through discus­
sion with representatives of industry and
unions provides essential background in mak­
ing these judgments. Analysis of trends and
projections for the economy as a whole, or for
indivdiual industries made by other groups,
such as the National Planning Association,
Stanford Research Institute, State and local
governments, and universities, also contribute
to these judgments. The adjusted overall in­
dustry projection framework is then utilized
as a basis for occupational requirements. These
rough occupational projections are then an­
alyzed and adjusted on the basis of the individ­
ual occupational studies previously described,
both the detail and the control totals. (See
description of Industry-Occupational Matrix,
chapter 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 information 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 occupations or because of retire­
ment or death, than are needed to staff new
positions created by growth of the field. In esti­
mating the total number of openings likely to
arise in an occupation, the occupational outlook
program analyzes studies of occupational mo­
bility among selected groups of workers, and
tables of working life.



51

These tables of working life, which are sim­
ilar to the actuarial tables of life expectancy
used by insurance companies, provide a basis
for assessing future rates of replacement re­
sulting from deaths and retirements, which are
in turn affected by differences in sex and aver­
age age of the workers in particular occupa­
tions. Where men comprise the great majority
of workers, estimated replacement rates for
death and retirement usually average between
1 and 3 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 em­
ployment to get married and assume family
responsibilities but who return to paid employ­
ment after marriage or raising a family. The
replacement rate among elementary school
teachers, for example, is estimated at 4.8 per­
cent 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 identifi­
able. Statistics on high school and college en­
rollments and graduations are the chief sources
of information on the potential supply of per­
sonnel in the professions and in occupations
requiring extensive formal education. Data on
numbers of apprentices and graduates of voca­
tional and technical training programs provide
some limited information on new entrants into
skilled trades. However, in many occupations
most new entrants are trained informally,
through on-the-job training or company train­
ing programs.
It is not enough to know, of course, how
many persons are being formally trained for
an occupation, since not all those completing
formal training 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 em­
ployment plans of college seniors, job place­
ments of college graduates, and jobs entered
after completion of MDTA and other types of
training. Limited data on transfers out of an

52

BLS HANDBOOK OF METHODS

occupation or re-entries into an occupation are
also utilized, although, in general, data on oc­
cupational mobility are available for only a few
occupations.
The estimates of the future demand in an
occupation is then related to estimates of the
future supply to develop the employment out­
look in that field and to provide information to
policy makers, educators, and others on the
implications of these relationships.

Presentation

The Occupational Outlook Handbook is the
major publication of the occupational outlook
program. Oriented toward vocational guidance,
the Handbook is a basic reference source, pub­
lished every other year, which includes com­
prehensive and non-technical job information
on approximately 800 occupations and 30 major




industries, covering the entire spectrum of
white-collar, blue-collar, and service occupa­
tions. An occupational outlook report series
provides reprints of individual statements from
the Handbook.
The Occupational Outlook Quarterly provides
a continuous flow of current occupational and
job information between editions of the Hand­
book, together with the most recent information
available on earnings, training requirements,
and other related topics. In addition to these
two publications, developed mainly for use in
vocational guidance, the occupational outlook
program conducts technical and detailed stud­
ies on specific occupations and industries in
order to furnish information to manpower ex­
perts, personnel departments, and others in­
terested in the more technical aspects of the
Nation’s future manpower needs.
— N

eal

H. R

o sen th al

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 pur­
poses. They provide a basis for establishing the
amount of employment growth the economy
must generate to maintain high levels of em­
ployment. They serve as the basis for one ap­
proach in setting goals for a general economic
growth rate consistent with full utilization of
human resources. Projections help to gain in­
sight into the characteristics and numbers of
workers who will be available for industry, and
to see what this implies for education, training,
and personnel policies. In addition, labor force
projections, together with population projec­
tions, are used to estimate demand for products,
develop marketing plans, and evaluate expan­
sion programs. The U.S. Department 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.

Method

Projections of the labor force as a whole and
of the separate age-sex groups are made for
quinquennial dates usually for about 15 years
ahead. The schedule for preparing the projec­
tions has been irregular owing, in part, to the
timing of new projections of the population of
working age. Labor force projections for 1975,
1980, and 1985 were published in May 1970.
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, projec­
tions have been made on the basic assumptions
that past trends in labor force participation
would continue into the future, and that the
economy would continue to expand and main­
tain high levels of employment opportunity
consistent with an unemployment rate of 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
patterns of the population. It also presupposes
that the trend toward increased school enroll­
ment beyond the high school level, which has
a direct bearing on labor force activity o f young
persons, would continue, 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 agesex group or subgroup that is expected to be
in the labor force, i.e., the labor force partici­
pation 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 projection. The procedure is
modified, as appropriate, to discount the tem­
porary effect of factors judged to be operative
for only short periods.
The population projections used in project­
ing 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 im­
migration and projecting 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 assumptions
with respect to birth rates. The uncertainty of
projecting birth rates does not directly affect
the level of the labor force projections 15 years
ahead, since everyone of working age (16 years
and over) at that future date has already been
born when the projections are made. However,
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
68

54

BLS HANDBOOK OF METHODS

seemed most reasonable on the basis of an in­
dependent evaluation of past trends in birth
rates. For recent labor force projections, series
“ C” of the population projections published in
the Census Bureau’s Current Population Re­
ports, P-25, No. 381 was chosen.
The overall size of the labor force is built up
by age and sex, not only because the composi­
tion is needed for many of the purposes noted
earlier, but also because the degree of labor
force participation 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 behavior of the labor force
participation rates and which affect particular
groups include school attendance, marital
status, birth rates, and the availability of social
security benefits and the expansion of private
pension plans. The method of projecting the
labor force participation rates for the various
age-sex groups takes into account the influence
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
future 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 spac­
ing. Similarly, projected marital status distri­
butions 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 respective future populations and the re­

sulting labor force summed to provide the total
labor force for each age-sex group and for all
ages.

Sources o f Data

The source of the basic historical data on
labor force participation rates by age and sex
used to project the labor force is the monthly
statistics on the labor force. These are pub­
lished by the Bureau of Labor Statistics and
are based on the Current Population Survey of
the Bureau of the Census. Historical data on
labor force activity by various categories
within several of the age-sex groups are ob­
tained from the recurring supplementary labor
force surveys also based on the Current Popu­
lation Survey. These include information from
the October surveys of the employment of
school-age youth and the March surveys of the
marital and family characateristics of workers.
The population projections are the latest
available projections made by the Bureau of the
Census and published in their Current Popula­
tion Reports, Series P-25. Projections of school
enrollment and marital status, by age, are based
on published and unpublished data of the Bu­
reau of the Census. Data used in projecting the
proportion of women in each age group who
will have children under age 5 years include
published and unpublished 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 re­
ports of the Bureau of the Census, Current
Population Reports, Series P-20, and data from
the decennial censuses of population.

Technical References
Number

1.

Travis, Sophia C., “ The U.S. Labor Force: Projections to 1985,” Monthly Labor Review,
May 1970, pp. 3-12, reprinted as Special Labor Force Report No. 119.
2. Cooper, Sophia, and Denis F. Johnston, “ Labor Force Projections for 1970-80,” Monthly La­
bor Review, February 1965, pp. 129-140, reprinted as Special Labor Force Report No. 49.
3. Population and Labor Force Projections for the U.S., 1960 to 1975 (BLS Bulletin 1242, 1959).




— So

p h ia

C ooper T

r a v is

Chapter 7. Industry-Occupational Matrix
Background

Sources o f Data

The Bureau of Labor Statistics has developed
a comprehensive set of data on the occupational
employment composition of all industry sectors
in the economy. Presently, industry-occupa­
tional matrices are available for 1960, 1967,
1970, 1975, and 1980. These data are set up to
form a matrix, or table, of 162 specific occupa­
tions plus groupings of occupations crossclassified with 116 industries. Thus, the occupa­
tional pattern of each industry is shown— i.e.,
the proportion of each occupation to total em­
ployment in an industry. Looked at another
way, the tabulation shows how total employment
in an occupation is distributed by industry.
Initially, work on the Industry-Occupational
Matrix grew out of concern by the Department
of Defense for anticipating the economic prob­
lems that might arise from various defense
programs. The first set of tables related to
1950 and were prepared by the Bureau as a
part of the interindustry 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 successors
continue to provide the basic information for
emergency manpower planning, now carried on
by the Office of Emergency Preparedness. In
recent years, a strong interest has developed in
determining manpower needs for other pur­
poses. The latter have included training new
workers, retraining workers displaced by auto­
mation, and providing information to high
school counselors and to students making career
decisions. The change has focused increased at­
tention on the need for estimates of numbers
presently employed in specific occupations and
the likely future employment requirements by
occupation. The Industry-Occupational Matrix
provides a systematic approach to developing
the desired information.

Data for the Industry-Occupational Matrices
are brought together from a wide variety of
sources. A major source for the development of
the 1960 matrix was the Occupation by Indus­
try report from the 1960 Census of Population.
The Current Population Surveys (CPS) are the
source for total employment, employment for
broad occupational groups, and for a few large,
specific occupations.1 Other sources of occupa­
tional employment data included the Bureau of
Labor Statistics annual surveys of occupational
wage rates in metropolitan areas and selected
industries ;2 regulatory agency statistics on em­
ployment by occupation in the telephone, rail­
road, and air transportation industries; U.S.
Civil Service Commission statistics on employ­
ment by occupation in the Federal Government *
,
statistics on selected professional occupations
based on licensing data and membership rec­
ords of professional societies; and surveys of
employers by the Bureau and other agencies to
obtain estimates of employment in a limited
number of highly important occupations such
as scientists, engineers,3 teachers, and police­
men.
Specific estimates from sources other than
the Census were incorporated into the cells of
the matrix for about 16 million workers, or
one-fourth of all those who were employed in
1960. The remaining details in the matrix were
derived by forcing 1960 population census esti­
mates for detailed cells (published in Occupa­
tion by Industry) into agreement with control
totals for occupational groups and industries
from sources other than the Census. The occu­
pational control totals were average annual
employment by occupational group taken from
the CPS. Most of the industry employment




1See chapter 1.
’ See chapter 14.
* See chapter 4.

65

56

BLS HANDBOOK OF METHODS

totals were based on BLS estimates of private
wage-and-salary workers adjusted to include
the self-employed, unpaid family workers, and
government workers, and to exclude the sec­
ondary jobs of dual job holders. Total employ­
ment in agriculture and private households was
based on CPS estimates. The adjustments of
the matrix to consistency with CPS estimates
of total employment and industry employment
estimates, derived as described above, brings
the matrix for 1960 into agreement with data
used as the basis for the Bureau’s projections
of total employment and occupational employ­
ment by industry. The Bureau’s occupational
projections are reflected in and developed in
part through matrix techniques. (See section
on analysis and uses.) The 1975 matrix 4 was
developed by examining a variety of historical
statistics on the changing occupational struc­
ture of industries and data from the 1950 and
1960 censuses and evaluating the factors likely
to influence changes in the future such as ex­
pected new technology, changes in product mix,
and the general organization of industries. A
similar procedure, together with use of infor­
mation for the 1960-67 period, was used in
preparing the 1980 matrix.
The 1960 matrix provided the base for the
1967 and 1970 matrices.5 Where available, oc­
cupational data from other sources, such as
those cited above, were incorporated into the
updated matrices as fixed cells. For the re­
maining cells, first approximations of the oc­
cupational patterns for 1967 were made by
interpolating between the patterns of the 1960
and the 1975 matrices. The resulting patterns
(in mining and manufacturing) were then
brought into consistency with data on produc­
tion worker trends available from the Bureau’s
Current Employment Statistics program. The
patterns were then applied to individual indus­
try employment controls and summed up to
arrive at occupational totals. These occupa­
tional control totals were then compared to
data from the CPS and other sources of infor-*
* For a detailed description of the procedures followed in
developing the 1975 matrix, see Occupational Employment
Patterns, 1960 and 1975 (BLS Bulletin 1599, 1968).
•For a more complete discussion of the methods used in
developing the 1967 and 1970 matrices, see Occupational
Employment Statistics 196^-67, BLS Bulletin 1643.




mation. When necessary, certain occupations
(except for fixed cells) were then forced on a
prorated basis to predetermined occupational
control levels. This iterative forcing procedure
was repeated until the internal matrix cells
were consistent with both the industry and the
occupational controls. A similar procedure was
followed in preparing the 1970 matrix with the
exception that the 1967 matrix provided the
basic data file of occupational ratios used for
the iterative forcing procedure. Thus, both
the 1967 and the 1970 industry-occupational
matrices were consistent with (a) national em­
ployment by industry, (b) broad occupational
employment levels from the CPS, (c) trends in
production (and nonproduction) worker em­
ployment by industry, (d) anticipated trends
in occupational structure within industries, and
(e) reliable estimates of detailed occupational
employment available from the CPS and other
sources.

Analysis and Uses

A basic objective of the project is to have
available a comprehensive set of data on in­
dustry-occupational 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 de­
tailed industries. Each industry utilizes a
unique combination of occupational skills, to­
gether with other factors of production, in its
efforts to achieve least cost for its output. Oc­
cupational patterns may be markedly different
from one industry to another. For example,
employment in the insurance industry is pri­
marily of white-collar workers such as insur­
ance agents, office clerical workers, actuaries,
and others. In contrast, the work force in
restaurants is largely made up of waiters,
waitresses, cooks, and owner-managers. Over
periods as short as a decade, the occupational
structure of many industries is relatively
stable. Consequently, if good information is
available on the occupational composition of
individual industries for a base period, it can

57

INDUSTRY-OCCUPATIONAL MATRIX

be used together with the available statistics
on changing employment in each industry to
develop estimates of current employment by
occupation for later periods. Further, if pro­
jections of output and employment are avail­
able by industry, the base period occupational
ratios applied to the industry employment pro­
jections will yield initial estimates of employ­
ment requirements by occupation for future
periods.
Although the occupational patterns of many
industries are relatively stable over periods of
5 to 10 years, it is clear that occupational pat­
terns change with the advance of technology
and changes in the supply of workers in each
occupation. Hence, information on how tech­




nology and labor supply are changing the oc­
cupational pattern in each industry is used to
modify the initial estimates. This improves the
estimates of current employment by occupation
and of future employment requirements by oc­
cupation, developed by applying base period
industry-occupational ratios to industry em­
ployment estimates. Changing technology and
other factors which affect skill requirements
are constantly being studied in order to esti­
mate the future occupational structure of each
matrix industry. The adjusted occupational
patterns are then used, together with projec­
tions of employment by industry, to prepare
estimates of future employment requirements
to 1980.
— R

ic h a r d

D

em psey




Prices and Living Conditions
Chapter 8. Consumer Expenditures and Income
Background
Consumer expenditure surveys* are specialized
1
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
members.
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 (C P I). (See chapter 10.)
The next major study, for 1934-36, 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
people.

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 CBS are used to refer to the
program of consumer expenditure surveys.
1 The State of Massachusetts 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. Bor 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 Q (1960).
*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 families; as a byproduct
of the reports required for the analysis of prices, certain data on
family Income, savings, and expenditures were obtained.
« In addition to these 7 major surveys, the Bureau has con­
ducted a number of expenditure surveys In cities selected for
specialized studies.

69

60

CONSUMER EXPENDITURES AND INCOME

All data were collected through the voluntary
cooperation of families.6 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, housefumishings,
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 housefumishings 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 BLS was responsible for collecting data from
all residents o f urban places. The BLS and
USDA shared this responsibility in the rural areas



of Standard Metropolitan Statistical Areas
(SMSA’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
“ 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 BLS Regional Offices1 and brought to
0
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.
Questionnaires

The detailed questionnaires used by the BLS
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 o f the family’s
living arrangements, income, spending, and
* See discussion of bureauwide policy on voluntary reporting
and confidentiality, under Introduction.
* This category Includes children temporarily away from home
at school or college.
TThe tabulations published in the CBS 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
research.
8 See appendix B for description of SMSA classification system.
* See discussion of sample design, p. 56.
“ The following description refers to procedures of the BLS,
but USDA procedures were similar.

BLS HANDBOOK OF METHODS

changes in sayings, consisted of 59 pages and
formed the basic framework of the survey. Sched­
ule C 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.1 The coding system for
1
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 items1 *
1
2
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.
Reported 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
11A 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 was not used In interviewing
farm families.
u 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 I9 6 0 , Philadelphia, Pa., University of Pennsylvania, Wharton
School of Finance, 1959, pp. 18-19.
1 Prior to editing the schedules for machine tabulation, the
1
schedules were reviewed in the Washington offices of the BLS
and USDA primarily to determine that entries conformed to the
survey concepts and methodology. This review was 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 was a
substantial lack of balance between the family’s reported total
receipts and disbursements.
1 This procedure involves the probability selection of a sample
4
“ 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 1960, pp. 1078-1083. (BLS Reprint 2352.)




61

ambiguous entries or to complete a record.1 On
8
the average, the interviewer spent 7 to 8 hours with
a family in a series of visits arranged at the
family’s convenience.

Sampling
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 o f 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 BLS utilized the
“ controlled selection” procedure.1 * The primary
4
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

62

CONSUMER EXPENDITURES AND INCOME
S u r v e y o p C o n s u m e r E x p e n d i t u r e s , 1960-61

(S u m m a r y o f S a m p l e S iz e a n d A v a i l a b i l i t y o p 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 ­
l a t io n S t r a t u m , a n d SMSA 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
Assignment
addresses
Total urban and rural—United States............................................................
Northeast........................... .......................................................................
North Central.............................................................................................
South.................................................................................................. ........
West............................................................................................................

Publications

Survey year

Usable
schedules

1960

1961

(*)
(*)
(*)
(*)
(*)

(*)
(*)
(*)
(*)
(*)

BLS report
number

Supplements

237-93
237-89
237-90
237r91
237-92

1,2,3
1,2,3
1,2,3
1,2,3
1,2,3

«17,283
4,127
5,187
5,100
12,869

113,728
3,228
4,092
4,180
12,228

Rural farm—United States___________________________________________
.................
.......................
Northeast
North Central .. ..
...
... .
South
West.. ........ ......
....
...
_
...

2,581
'214
980
1,181
'206

1,967
145
742
925
155

(*)
(•)
(*)
(*)
(*)

Northeast_______________________________________________________
North Central___________________________________________________
South___________________________________________________________
West
... ............ _ ............ .

2,497
' 453
702
999
343

2,285
'406
628
948
303

(*)
(*)
(*)
(*)
(*)

237-88
237-84
237-85
237-86
237-87

1,2,3
1,2,3
1,2,3
1,2,3
1,2,3

»12,205
3,460

«9,476
2,677

(*)
(*)

(*)
(*)

237-38
237-34

1,2,3
1,2,3

375
625
500
375
375

268
448
356
313
323

(*)
(*)
(*)
(*)
(*)

(*)
(*)
(*)
(*)
(*)

237-7 and 57
237-4 and 54
237-13 and 63
237-8 and 68
237-11 and 61

*1
*1
»1
»1
»1

250
250

199
176

(*)

(*)

237-18
237-68

•1
»1

160
160

135
151

(*)
(*)

237-14
237-64

«1
11

65
65
65
65
65
65

52
47
41
60
56
53

(*)
(*)
(*)

(*)
(*)
(*)

237-24
237-24
237-24
237-74
237-74
237-74

3,505

2,722

(*)

500
375.
375
375

371
294
290
319

(*)
(*)
(*)
(*)

250
250
250

173
180
189

(*)

160
160
160

125
126
130

(*)
(*)

65
65
65
65
65
65
65
65
65
65

49
55
65
61
48
43
61
50
45
58

(*)
(*)
(*)
(*)
(*)

2,920

2,307

375
375

313
323

250
250
250

198
178
201

(*)
W

160
160
160
160

110
106
112
135

(*)
(•)

Urban—United States......................................................................................
Northeast....................................................................................................
SMSA, population 1,400,000 and over
Boston, Mass...............................................................................
New York, N .Y ............................................................................
Northeastern New Jersey.............................................................
Philadelphia, Pa...........................................................................
Pittsburgh, Pa...............................................................................
SMSA, population 250,000 to 1,400,000
Hartford, Onnn

. ...................... . . _

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

SMSA, population 50,000 to 250,000
Portland, Maine
Lancaster, Pa .
Nonmetropolitan urban place, population 2,500 to 50,000

Athol, Mass
Millville, N.J
_____
Sonthhridge, Mass . .

______ _
............

North Central.............................................................................................
SMSA, population 1,400,000 and over
Chicago, 111...................................................................................
Cleveland, Ohio........................................ ...................... ..........
Detroit, Mich........ ................__........... ...................................__
St. Louis, Mo................................................................................
SMSA, population 250,000 to 1,400,000
Tndianapnlis, Tnd. _

SMSA, population 50.000 to 250,000
Cedar Rapids, Tnwa. .
.....................
Champaign-Urbane, T
U
GreenBay, W is ..._______________________________________
Nonmetropolitan urban place, population 2,500 to 50,000
Devils Lake, N.Dak ............ ..........
..........

Crnnlrstnn, Minn _
T^iganspnrt, Tnd .
Manhattan, Nans .
Menasha, Wis_. _

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

South...........................................................................................................
SMSA, population 1,400,000 and over
Baltimore, Md..............................................................................
Washington, D.C............ .......................................... - ................
SMSA, population 250,000 to 1,400,000

SMSA, population 50,000 to 250,000
Orlando, Fla_______
_______ _
_________
Raton Rouge La
....
Durham, N.C................... .............................. ................. ..........
See footnotes at end of table.




(2)
(2
)
(J
)
(2
)
(2
)

(»)
(2
)
(2
)
(2
)
(2
)

(*)
(*)
(*)
(4
)
(4
)
(4
)

(*)

237-35

(*)
(*)
(*)
(*)

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

»1
*1
*1
*1

(*)
(*)

237-10
237-60
237-67

11
•1
»1

(*)

237-17
237-23
237-73

»1
*l
»1

(»)
(*)
(*)
(*)
(*)

237-27
237-27
237-27
237-27
237-27
237-77
237-77
237-77
237-77
237-77

(*)

(*)

237-36

(*)
(*)

(*)
C)

237-16 and 66
237-3 and 53

*1
*1

<*)

237-6
237-20
237-56

»1
*1
*1

(*)
(*)

237-12
237-19
237-62
237-69

»1
»1
*1
*1

1,2,3

(4
)
(<)
(4
)
(4
)
(4
)
(4
)
(4
)
(4
)
(4
)
(4
)
1,2,3

63

BLS HANDBOOK OF METHODS
Su r v e y
(S u m m a r y

of

S a m p l e S iz e

of

C o n s u m e r E x p e n d it u r e s , 1 9 6 0-6 1 — Continued

a n d A v a il a b il it 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
o r O t h e r U r b a n P l a c e ) — Continued

R e g io n , P o p u ­

SMSA

Urban—United States—Continued
South—Continued
Nonmetropolitan urban place, population 2,500 to 50,000
Cleveland, Tenn___________________ _____________
Griffin, Ga______________________________________
McAllen, Tex____ ______________________________
Reserve, La_____________________________________
Union, S .C ...................................................................
Vicksburg, Miss________________________________
Florence, Ala___________________________________
Gainesville, Tex___________________ ____ _________
Mangum, Okla............................................................
Martinsville, Va________________________________
Okmulgee, Okla................................................. .........
Sebring, Fla............................................................ ....

Assignment
addresses

Publications

Survey year

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

1960

Usable
schedules

1961

BLS report
number

Supplements

65
65
65
65
65
65
65
65
65
65
65
65

43
61
38
64
50
55
54
56
50
55
48
57

12,320

>1,770

(»)

(*)

237-37

500
375

388
302

(*)
(*)

(*)
(*)

237-22 and 72
237-2 and 52

»1
»1

SMSA, population 250,000 to 1,400,000
Seattle, Wash..................................
Denver, Colo____ ______________
Honolulu, Hawaii..........................

250
250
250

209
(*)
.........
204 ............
(*)
215 ............
(*)

237-9
237-59
237-78

»1
»1

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

160

120

(*)

237-70

U

Nonmetropolitan urban place, population 2,500 to 50,000
Anchorage, Alaska....................................................... .
Gallup, N.Mex_________________________________
Klamath Falls, Oreg....................................................
Eureka, Calif................................................................
Orem, Utah..................................................................

275
65
65
65
65

134
58
44
42
54

(s)

237-29
237-26
237-26
237-76
237-76

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

(*)
M
H
<)
*
(*)

(')

8
(*)

237-25
237-25
237-25
237-25
237-25
237-25
237-75
237-75
237-75
237-75
237-75
237-75
1,2,3

«1

(‘
W
(*
)
(*
)
<
«
>

•Asterisk Indicates year of survey,
i Includes Anchorage, Alaska, which was surveyed for 1959.
* Published by the U.S. Department of Agriculture.
1 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.
* Survey for 1959.

N ote ; 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. H alf 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 BLS in the fall of 19601
6
or 1961. The BLS 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, BLS
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.1 In the third stage, subsamples
6
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
stratum.

u Because of the shortage of time, the BLS 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 Districts (ED’s).
M These classifications, which were on the basis of Census
definitions, made it possible for the BLS 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.




64

CONSUMER EXPENDITURES AND INCOME

In the first stage of the USDA’s design for the
rural sample in nonmetropolitan areas, counties
were grouped by State Economic Areas1 into 126
7
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.1
8
Usable schedules were tabulated for 13,728 con­
sumer units. The distribution ,o f 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.1 This information was sum­
9
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,



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,2 including Anchor­
0
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
1 See appendix B.
7
MThe address following each address In the master sample waa
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 at 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
State.
*° For a single year such as 1961, the city weights dlifered
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.

BLS HANDBOOK OF METHODS

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 o f
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 o f 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 o f 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
tenure.
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.




65

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 BLS 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 o f 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 o f 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
studies.2
1
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
n The Bureau has Initiated a series of special analytical re­
ports (BLS 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.

CONSUMER EXPENDITURES AND INCOME

66

come tax returns. Currently the Department of
Health, Education, and Welfare is studying the
data in connection with numerous welfare
analyses.
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
organizations.
Data obtained from a sample survey as complex
as the survey o f 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 BUS 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 o f urban and
rural families furnished usable schedules and
some of the nonrespondents supplied limited infor­
mation on family characteristics.2 Among the
2
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 BLS and USDA have
accumulated substantial knowledge about such
reporting errors and will continue research in
this field, these errors cannot be quantified
satisfactorily.
” This Information la 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
Num ber

1. Brady, Dorothy S. and Williams, Faith M ., "Advances in the Techniques of Measuring and Esti­
mating Consumer Expenditures," Journal of Farm Economics, 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 American Economic 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, Family Living Studies— A Symposium, 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
American Economic 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.




67

BLS HANDBOOK OF METHODS

Technical References—Continued
Number

5. Lamale, Helen H. Study of Consumer Expenditures, Incomes and Savings— Methodology of the Survey
of Consumer Expenditures in 1950. 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
1950.
6. Stigler, George J., “ The Early History of Empirical Studies of Consumer Behavior,” The Journal
of Political Economy, 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.
7. United Nations Statistical Office, Handbook of Household Surveys: A Practical Guide for Inquiries
on Levels of Living, 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.
8. U.S. Bureau of the Census, Historical Statistics of the United States— Colonial Times to 1957, (1960)
Chapter G. The same, Continuation to 1962 and Revisions, 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.
9. U.S. Department of Labor, How American Buying Habits 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.
10. Williams, Faith M. and Zimmerman, Carle C., Studies of Family Living in the United States and
Other Countries: An Analysis of Material and 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 Consumer
Behavior— Research on Consumer Reactions (Lincoln H. Clark, ed.) New York, Harper and
Brothers, 1958.
11. Zimmerman, Carle C., Consumption and Standards of Living, 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.




—K

athryn

R. M

urphy

Office of Prices and Living Conditions




Chapter 9. Family Budgets
Background
“ Standards of living” refer to the goals of
consumers and workers in their consumption of
goods and services, use of leisure time, and
conditions of work. Standard budgets, also de­
scribed as family budgets, measure the total
costs or amounts of income required to achieve
the levels and manner of living implicit in these
goals.1 Cost estimates are developed by trans­
lating the generalized concept of a living stan­
dard into a list of commodities and services
which can be priced. Thus, standard budgets are
normative, or benchmark, estimates of living
costs. They do not represent the ways in which
family incomes should be spent, or the ways
average families actually spend their incomes.
The first standard budgets prepared by the
Bureau of Labor Statistics were developed for
the specific purpose of evaluating living condi­
tions of cotton-mill workers in Fall River, Mass,
and in the South in 1908-09. These budgets
described two standards of living— a minimum,
including only bare necessities; and a fair stan­
dard, including some allowance for comfort. An­
other budget defining a standard of health and
decency was developed in 1919. In the late
1930’s, BLS cooperated with the Works Prog­
ress Administration in pricing two budgets: A
maintenance budget, described as above the
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 “standard” 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 conditions which they seek to attain or regain, or
which they regard as fitting and proper for themselves to
enjoy. 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 Tech­
nical Reference No. 3, listed on p. 73.
3 See Technical Reference No. 18. For description of a com­
panion budget for an elderly couple, see Technical Reference
No. 19.
4 See Technical Reference Nos. 7 and 10.
3 See Technical Reference Nos. 11, 14, 15, and 16.




minimum subsistence level but 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 concerning the requirements for nutrition­
ally 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, the “ require­
ments” were formulated primarily on the basis
of the personal judgment of the budget makers.1
2
In 1946, the Bureau compiled the City Work­
er’s Family Budget for a “modest but adequate”
standard of living. The procedures used stan­
dards of adequacy that reflected the judgments
of scientists and experts where these were avail­
able; for other components, they depended on
statistical analyses of consumer choices.3 The
same method, with some refinements, was used
in 1959, in the interim revisions of the City
Worker’s Family Budget and the Budget for a
Retired Couple.4 These procedures were used
again, with additional refinements, in the mid1960’s to develop the “ intermediate” budgets
(initially described as “ moderate” budgets) for
a four-person family and a retired couple. Sub­
sequently, the costs of the intermediate level
budgets were scaled downward and upward by
a variety of techniques to produce a “ lower”
and a “ higher” budget for each family type.5
Procedures for the intermediate budgets of the
1960’s and the scaling techniques are described
in the remainder of this chapter.

Description of the Budgets
All normative estimates of living costs must
be based on specific family situations. The con­
struction of a family budget, therefore, requires
a set of assumptions, i.e., specifications, which
must be formulated explicitly by the budget
maker at the outset. These relate to the age,
69

70

BLS HANDBOOK OF METHODS

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 position of the living standard in rela­
tion to the actual scale of consumption.
Family composition has a significant effect
on spending patterns, manner of living, and
family needs. The budgets ior a younger, fourperson family, specifies that the family consists
of an employed husband, age 38, who has a wife
not employed outside the home, and two chil­
dren, a girl of 8 and a boy of 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 budgets for a retired couple consists of a
husband and wife, age 65 or over, who are
assumed to be self-supporting, in reasonably
good health, and able to take care of themselves.
This unit, which has a markedly different pat­
tern of living and needs than the younger fam­
ily, has been the subject of special concern in
national policy formation over the last three
decades. Budget quantities and budget com­
ponent cost estimates for other family types
cannot be derived as fractions or multiples of
the quantities or cost estimates for food, shel­
ter, clothing, transportation, etc. for the fourperson family or the retired couple.6
Both types of families were assumed to live
in an urban area. Assumptions also were made
concerning the living arrangements and tenure
of the families; inventories of housefurnishings, household equipment, and clothing; means
of transportation; ownership of life insurance;
provisions for medical care; savings positions,
etc. In making these assumptions, the budget
8 Extensive analyses of consumption data dating back over
more than a century have provided a variety of measures of
general welfare; e.g., the relative adequacy of diets, the pro­
portion of income spent for various categories of goods, or the
proportion of income saved. These measures, either singly
or in combination, have been used as the basis for determin­
ing scales of equivalent income for families of different size.
One such scale is described in Technical Reference No. 12.
The scale is based on the assumption that families spending
the same proportion of income on food have attained equal
levels of living. Although the scale is useful in estimating
equivalent costs of goods and services, or net income re­
quirements after income taxes and occupational expenses, it
cannot be applied to individual items or major components of
budget costs.
7 For a discussion of the relativity of living standards, see
Technical Reference Nos. 5 and 6.
8 For a description of the Bureau’s surveys of consumer e x­
penditures, see chapter 8.




makers were guided by data on the prevalence
of ownership of particular types of assets in the
urban metropolitan population, and the avail­
ability of goods and services provided by gov­
ernments for collective consumption or provided
under collective bargaining agreements between
employers and unions.
All three budgets provide for the mainte­
nance of physical health and social well-being,
and participation in community activities.
Within this broad framework, different levels
of adequacy were obtained by varying the as­
sumptions concerning the manner of living and
by providing different quantities and qualities
of the necessary goods and services.
The content of the budgets is based on the
manner of living and consumer choices in the
1960’s. The lower budget differs from the inter­
mediate and higher budgets in several specifica­
tions: The family lives in rental housing with­
out air conditioning, (except for a proportion
of retired couples who may own their own
homes), relies heavily on public transportation,
supplemented, where necessary, by the use of
an older car, performs more services for itself,
and utilizes free recreation facilities in the com­
munity. Compared with the intermediate
budget, the life style in the higher budget, is
marked by more homeownership, high levels of
new-car ownership, more household appliances
and equipment, and more paid-for services. For
most items common to all budgets the quanti­
ties are greater and the qualities higher in the
intermediate than in the lower budget, and in
the higher than in the intermediate budget.7

Data Sources
Budget quantities and pricing specifications
were derived from two sources: (1) Scientific
or technical judgments concerning the require­
ments for physical health and social well-being;
and (2) analytical studies of the choices of
goods and services made by consumers in suc­
cessive income groups, as reported in the Bu­
reau’s surveys of consumer expenditures,8 to
determine by statistical procedures the income
class whose spending pattern would be used as
the “ norm” for a specified budget level.
Scientific standards for nutritionally ade­

71

FAMILY BUDGETS

quate diets for individuals in different sex-age
groups have been developed by the Food and
Nutrition Board of the National Research Coun­
cil, 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 Ameri­
can Public Health Association and the U.S.
Public Housing Administration were adopted
for the budgets. These standards relate to sleep­
ing space requirements, essential household
equipment (including plumbing facilities), ade­
quate utilities and heat, structural condition,
and neighborhood 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 cli­
mate. Estimates of electricity and utility serv­
ices 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 basis for
a standard for medical care, and a family mem­
bership in a group health insurance plan (Medi­
care for the retired couple) was specified.
Quantities of medical care services not covered
by insurance were derived from data on utiliza­
tion rates provided by the 1963-64 U.S. Na­
tional Health Survey and the 1960-61 Consumer
Expenditures Survey. Major medical provisions
were specified for the higher budget.
No generally accepted scientific standards
are available for other components of the budg­
ets (clothing, housefumishings, transportation,
personal care, household operation, reading,
recreation, tobacco, education, gifts and contri­
butions, and miscellaneous expenses). There­
fore, for most of these components a technique
was developed which relied on the choices of con­
sumers as the basis for a standard. Purchases
were examined at successive income levels to
determine the income level at which the rate of
8 For a description of the Consumer Price Index, see chapter
10.
1 Technical Reference Nos. 13, 15, and 16 provide additional
0
details on the methods, as well as lists of goods and services
priced. Dollar cost estimates are provided in these reports
and also in Technical Reference No. 9.




increase in quantities purchased began to de­
cline in relation to the rate of change in income,
i.e., the point of maximum elasticity. The aver­
age numbers and kinds of items purchased at
this income level became the quantities and
qualities specified for the intermediate level
budget. In general, income classes below and
above the classes used for the intermediate level
were specified as the source of quantities for
the lower and higher budgets, respectively.
For the transportation component, quantities
for the intermediate and higher budgets were
based on the average consumption pattern of
families of each budget type. For the lower
budget, average patterns of renter families
were used. Except for the higher budget where
costs include a car for all families, automobile
ownership was specified in inverse relationship
to the availability of mass public transporta­
tion. Mileage allowances were adjusted by the
use of automobiles for work.
In determining budget costs, levels of prices
paid for items are as important as the numbers
of items bought. Items in the intermediate
budget were priced in the types of stores and
professional and service establishments custom­
arily patronized by urban families. Prices, pric­
ing procedures, reporting stores and service
establishments, and price calculation methods
were those used by the BLS for the Consumer
Price Index,8 except that additional quotations
9
were obtained in some cases to calculate aver­
ages and different qualities were priced in other
cases to represent the intermediate budget
level. For some items in the lower and higher
budgets, special prices were collected directly
from stores and establishments. In the main,
however, prices for those two levels were esti­
mates in a variety of ways.1
0

Analysis and Presentation
In the methods described, a family budget
is the end result of a multitude of decisions by
the budget maker, based on standards formu­
lated by scientists or experts or on analyses of
data on consumption patterns from a variety of
sources. The budgets are not simply the prod­

72

BLS HANDBOOK OF METHODS

ucts of a survey of ways families at particular
income levels actually spend their money. The
judgment of the budget maker is involved in
selecting among the family types and manners
and levels of living to be represented; in deter­
mining the most appropriate sources of data to
be used in deriving budget quantities; and in
interpreting actual family consumption in terms
of norms or benchmarks. The appropriateness
of the operating assumptions can be evaluated
only by the budget users in relation to the pur­
poses to be served.1
1
Budget estimates may be analyzed in four
ways: (1) Costs are compared with income.
However, costs for a specific family type should
be compared only with average incomes, or in­
come distributions, for families of the same
type. This kind of analysis has been restricted,
therefore, by the availability of cost estimates
for only two family types. However, family
equivalence scales may be used to develop esti­
mates for comparable benchmark levels for
families of other types. (2) Budget costs in one
place are compared with costs in another, i.e.,
the budgets provide a basis for calculating an
index of locality differences in living costs. The
Bureau has provided this type of analysis in
conjunction with its published reports. (3)
Costs are compared over time to measure
changes in living standards. The sporadic char­
acter of the Bureau’s family budget research
program imposes serious limitations on this
type of analysis. Also the judgment factor in
developing budgets introduces a serious bias for
evaluating changes in the levels and living stan­
dards of families from decade to decade. (4)
Finally, budget estimates of different levels are
compared to provide a measure of the aggre­
gate addition to income required to raise con­
sumption to particular levels. The development
of budgets for three different levels facilitates
this type of analysis.

Uses and Limitations
Family budgets are used in economic research
to appraise the economic condition of the popu­
lation and to evaluate the need for, and the
effect of, specific laws and programs. For ex­



ample, normative living costs are used, to mea­
sure the extent to which social security or un­
employment insurance benefits provide income
sufficient to purchase the manner and content of
living used to define a specified budget level; to
estimate aggregate costs of consumer goods as
a basis for developing public policies; or to pre­
pare estimates of the number of families living
below the specified budget level. Budgets also
provide benchmarks for administrative deter­
minations, as required by a number of existing
laws or policies of social, welfare, and educa­
tional agencies; e.g., to establish criteria of
eligibility for public assistance, public housing,
support services for individuals in job develop­
ment programs, subsidized medical or mental
health, guidance services, or college scholar­
ship aid.
In addition to their primary use as tools in
evaluating income adequacy, family budgets are
used 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.
Locality indexes based on the BLS budgets
reflect differences in costs of established resi­
dents in a community. Rental costs, for ex­
ample, are based on the averages for occupied
dwellings and are not a valid measure of the
costs of vacant units available to new residents.
Similarly, the costs of maintaining a home pur­
chased 7 years ago, while an appropriate mea­
sure for an established, budget-type family,
does not provide information on the relative
costs of purchasing homes in current markets.
The cost of food reflects not only differences in
price levels but also, and more important, differ­
ences in regional preference patterns in the
choice of food to meet nutritional standards.*2
1
The indexes, therefore, are more appropriate as
research tools in analyses of the relationship
between income and costs of established resi­
dents in different locations than as measures of
differences in costs for families moving from
one location to another.
u For a discussion of the uses of family budgets, see Tech­
nical Reference Nos. 2, 4, 8, and 17.
1 For a measure of the effect on food costs of price-level
2
differences versus regional differences in the choice of foods,
see Technical Reference No. 1.

FAMILY BUDGETS

Technical References
1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

Brackett, Jean C., “ Intercity Differences in Family Food Budget Costs,” Monthly Labor Re­
view, 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.
Brackett, Jean C., “ New BLS Budgets Provide Yardsticks for Measuring Family Living
Costs,” Monthly Labor Review, April 1969, pp. 3-16.
Provides a nontechnical description of the concepts and procedures used to develop the
budgets for a four-person family at three levels of living, a summary of the spring 1967 cost
estimates and locality indexes, and a discussion of the appropriate uses of the budgets.
Clorety, Joseph A., “ Consumption Statistics: A Technical Comment,” How American Buying
Habits Change, Chapter X , 1959, pp. 217-242.
Presents a representative cross-section of budgets compiled in this country during the
20th century. Shows average dollar cost figures for the total and for the major components of
each budget.
Hawes, Mary H., “ Measuring Retired Couples’ Living Costs in Urban Areas,” Monthly Labor
Review, November 1969, pp. 3-16.
Includes estimates of the costs of budgets for three levels of living for a retired couple in
spring 1967 and budget-based locality indexes. Describes uses of these budgets as tools in de­
termining eligibility for various programs and in helping older couples to evaluate their own
spending habits.
Lamale, Helen H., “ Changes in Concepts of Income Adequacy Over the Last Century,” Ameri­
can Economic Review, 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.
Lamale, Helen H., “ Poverty: The Word and the Reality,” Monthly Labor Review, July 1965,
pp. 822-827.
Discusses the standard budget approach to the evaluation of income adequacy for different
family types and in different geographical locations and estimation of the extent of poverty in
the United States.
Lamale, Helen H. and Stotz, Margaret S., “ The Interim City Worker’s Family Budget,”
Monthly Labor Review, August 1960, pp. 785-808.
Estimates 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. Includes the detailed list of the goods and services used to define the living
standard for the 1950’s, and describes the way this list was developed and priced.
Orshansky, Mollie, “ Budget for an Elderly Couple: Interim Revision by the Bureau of Labor
Statistics,” Social Security Bulletin, December 1960, pp. 26-36.
A summary report on “ The BLS Interim Budget for a Retired Couple.” (See Reference
No. 10.) Includes a discussion of various conceptual problems encountered in developing norm­
ative living costs estimates for a retired couple, and some limitations of this particular bud­
get.
Ruiz, Elizabeth, “ Spring 1970 Cost Estimates for Urban Family Budgets,” Monthly Labor
Review, January 1971, pp. 59-61.
Summarizes annual costs and comparative cost indexes at three levels of living for fam ­
ilies of four persons in Spring 1970. Equivalence scale values are applied to four-person
family consumption costs to obtain comparable estimates for other family types.
Stotz, Margaret S., “ The BLS Interim Budget for a Retired Couple,” Monthly Labor 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. Includes the detailed list of the goods and services used to define the living
standards for 1950’s; and describes how this representative list was developed and priced.
U.S. Department of Labor, Bureau of Labor Statistics, “ City Worker’s Family Budget for a
Moderate Living Standard, autumn 1966” (Bulletin 1570-1, 1967).
Describes changes in this budget over the last two decades, and gives autumn 1966 costs




73

74

BLS HANDBOOK OF METHODS

Technical References—Continued

12.

13.

14.

15.

16.

17.

18.

19.

for urban United States and costs and comparative indexes for 39 metropolitan areas, and
4 nonmetropolitan regions.
U.S. Department of Labor, Bureau of Labor Statistics, “ Revised Equivalence Scale for Esti­
mating Equivalent Incomes or Budget Costs by Family Type” (Bulletin 1570-2, 1968). /
Includes scale values for selected family type which can be used to approximate total costs
of consumption for the three budget levels. Also includes a summary and discussion of the
status of research on family equivalence scales.
U.S. Department of Labor, Bureau of Labor Statistics, “ City Worker’s Family Budget Pric­
ing, Procedures, Specifications, and Average Prices, Autumn 1966” (Bulletin 1570-3, 1968).
Reports on pricing methodology used in the intermediate budget and includes U.S. urban
average prices and averages for five metropolitan areas for selected items priced for the
budget.
U.S. Department of Labor, Bureau of Labor Statistics, “ Retired Couple’s Budget for a Mod­
erate Living Standard, Autumn 1966” (Bulletin 1570-4, 1968).
Describes changes in this budget over the last two decades, and gives autumn 1966 costs
for urban United States and costs and comparative indexes for 39 metropolitan areas and
four nonmetropolitan regions.
U.S. Department of Labor, Bureau of Labor Statistics, “ Three Standards of Living for an
Urban Family of Four Persons, Spring 1967” (Bulletin 1570-5, 1969).
Describes budgets for a four-person family at three levels of living. Explains in detail the
concepts, procedures, data sources, and estimating methods, and provides lists of goods and
services priced. Includes spring 1967 costs and locality indexes.
U.S. Department of Labor, Bureau of Labor Statistics, “ Three Budgets for a Retired Couple
in Urban Areas of the United States, 1967-68” (Bulletin 1570-6, 1970).
Describes budgets for a retired couple at three levels of living. Explains in detail the con­
cepts, procedures, data sources, and estimating methods, and provides lists of goods and serv­
ices paid. Includes spring 1967 costs and locality indexes. (A supplement to this Bulletin pro­
vides costs and indexes for 1969-70.
U.S. Department of Labor, Bureau of Labor Statistics, “ Report of the Advisory Committee
on Standard Budget Research,” June 1963.
Recommendations formulated by a committee of experts with extensive experience in us­
ing standard budgets on the needs for various types of budgets, general concepts of the stan­
dards of living to be described by the budgets, and methodological and other problems asso­
ciated with estimating and publishing budget costs. Includes a selected bibliography on the
major uses of standard budgets.
U.S. Department of Labor, Bureau of Labor Statistics, “ Workers’ Budgets in the United
States: City Families and Single Persons, 1946 and 1947,” (Bulletin 927, 1948).
Concepts, definitions, and techniques used in developing the original City Worker’s Fam­
ily 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.
“ A Budget for an Elderly Couple,” Social Security Bulletin, 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 Worker’s Budget. See Reference No. 18.)




— Je

a n

C. B

r ack ett

Chapter 10. Consumer Prices
Background
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.1 This ad­
2
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
5
and Labor conducted a detailed examination of the
index in 1951.® 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 (BLS
Bulletin 1039, 1951).
3 See Consumer Prices in the United States 195S-58 (BLS 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, 8 2 /1 , Subcommittee Report No. 2, Washington, 1951.
• Oovernment Price Statistics— Hearings before the Subcom­
mittee on Economic Statistics, Joint Economic Committee, Con­
gress of the United States, 8 7 /1 , Part 1, Washington, Janu­
ary 24, 1961.

75

76

BLS HANDBOOK OF METHODS

Description of the Index
Concept and Scope.7 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
services.
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 su6h durable goods as automo­
biles, refrigerators, etc.
In the 1964 revision, the index coverage was
extended to include single consumer units1 * in
0
addition to families of two or more.1 The average
1
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)1 in 66 Standard Metropol­
2
itan Statistical Areas (SMSA’s) and smaller cities
7 See article by Sidney A. Jaflfe, “ The Statistical Structure of
the Revised CPI” , Monthly Labor Review, August 1964, pp. 9 1 6 924.
* 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 workers; service workers, ex­
cept private household, such as waitresses, practical nurses, e tc .;
sales workers; and laborers, except farm and mine. It excludes
professional, technical, and kindred workers, such as engineers
and teachers; farmers and farm managers; managers, officials
and proprietors, except fa rm ; private household workers; 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 was obtained in a wage-earner and clericalworker occupation and at least one family member was a full­
time earner (i.e., worked 37 weeks or more during the survey
year).
•For a more detailed discussion, see “ Taxes and the Con­
sumers’ Price Index,” Monthly Labor Review, January 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.
u From January 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.
7 The Survey of Consumer Expenditures is discussed in
3
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)

CONSUMER PRICES
T able

1.

77

C it i e s , P o p u l a t io n W e ig h t s , a n d P r ic in g S c h e d u l e f o b t h e R e v is e d C o n s u m e r P r ic e I n d e x
Pricing schedule *
Other items
City and size stratum

Population
weight >

Food*

Schedule*
Samples
M

1

A. Standard metropolitan statistical areas of 1,400,000 or more in 1960:
Baltimore, Md.................... ................................... ........................................
Boston, Mass......... ........... ...............................................................................
Chicago-Northwestern Indianai..................................................................
4
*
Cleveland, Ohio................................................................... ...........................

1.402
1.930
5.552
1.325

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

X

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

2.895
5.017
12.577
2.703

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

1.565
1.428
2.372
1.255

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

B. Standard metropolitan statistical areas of 260,000 to 1,399,999 in 1960: *
Atlanta, Qa............................................................................ ..........................
Buffalo, N .Y .....................................................................................................
Cincinnati, Ohio-Ky... .............................................................. ...................
Dallas, Tex.......................................................................................................
Dayton, Ohio...................................................................................................

2.934
2.347
.740
2.934
1.096

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

Denver, Colo.....................................................................................................
Hartford, Conn.................... ............... ..................................... ....................
Honolulu, Hawaii...........................................................................................
Houston, Tex.......................................................................................... ......
Indianapolis, In d ................................... ............................................... ........

1.838
2.348
.354
.999
1.095

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

Kansas City, Mo.-Kans................................................................................
Milwaukee, Wis............................................................................................
Minneapolis-St. Paul, Minn____________ ______________________ ______

.710
.850
1.042

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

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

2.933
.672
1.837
1.096

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

X

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

1.250
1.323
1.250
1.284
1.284

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

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

X

Durham, N .C....................................................... ............ ............... .............
Green Bay, Wis...................................... .......................................................
Lancaster, Pa........... ..................... ........................................................... ......
Orlando, Fla................................................... ................................................
Portland, Maine...............................................................................................

1.250
1.284
1.803
1.250
1.803

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

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

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

.065
1.352
1.352
1.352
1.227

1.2
1
2
1
1

1.2
1
2
1
1

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

1.171
1.338
1.352
1.226
1.227

2
1
2
1
2

2
1
2
1
2

X

McAllen, Tex....................................................................................................
Millville, N.J.....................................................................................................
Niles, Mich........................................................................................................
Orem, U tah .......................................................... ..................................... .
Southbridge, Mass.................................................. .................................. ......

1.227
1.171
1.351
1.339
1.170

2
2
1
2
1

2
2
1
2
1

X
X

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

1.227
1.226

1
2

1
2

3

X
X
X
X

Pittsburgh, Pa........................................................... .....................................
St. Louis, Mo..... ................................................ ..................................... ......
San Francisco-Oakland, Calif.........................................................................
Washington, D.C_________________ _____ ____ ____ _______ ____________

2

X

i The 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.
* 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
components.




X
X
X

X

X
X
X

X
X
X
X

X
X

X
X

X

X
X
X
X
X
X

X

X
X

X
X
X
X

X
X

X
X

X

X

X
X
X

X
X

X
X
X

X
X

* 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

78

BLS HANDBOOK OF METHODS

covering the period 1960-61, except for Anchor­
age, Alaska, which was surveyed for 1959. Ex­
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 SMS A 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
weights.1
*
1
3
The list of items priced, includes all the most
important goods and services and a sample of the
1 Six additional B size Standard Metropolitan Statistical Areas
1
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.
1 Minor weight revisions are introduced by linking.
4
1 For a more complete discussion, see The Consumer Price
5
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 1 in the index calculation between major
4
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 shops.1
5
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. Co­
operation 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

79

CONSUMER PRICES
T a b l e 2.

C o n s u m e r P r ic e I n d e x ( N e w S e r ie s ) 1 R e l a t iv e I m p o r t a n c e o p M a jo r G r o u p s , S p e c ia 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 * D e c e m b e r 1963

Components

Percent of
all items
December 1063
100.00

M ajor G roups

22.43
33.23
10.63
13.88
19.45
5.70
2.75
5.04
5.06
S pecial G roups

All items less shelter..____ ____ ____ _____________________

Nondurables_______ ______ ___ _____________________
Services... _________
..
_________________________
Commodities less food_______ _________________________ _
Nondurables less food.___
________ ____ __________

Housefumishings_______ _________ _________________

79.85
77.57
65.97
47.19
18.78
34.03
43.54
24.76
9.16
7.65
15.60
5.69
4.72
28.53
13.47
4.86
4. 56
5.64

I ndividual I tems

Food__________________ ________ _______________________
Food at home______________________________ ____ ___
Cereals and bakery products_____________________

Meats, poultry, and fish_________________________
Meats___II . . ___________________________ ..

Pork..

_____________________________

Other priced items________ __________

Fish.......................... ............................................
Dairy products___________________ _________
Mflk, fresh (grocery).. _____________________
Milk! fresh (delivered)___ __________ _______
• Butter_____________ 1 ___________
_ ... ...
Other priced items.. ______ _________ ______ _
Fruits and'vegetables____________________________

Potatoes_________________________ _______

Other priced items__________ ____ _______

Other priced itemsl______________________

Between meal snacks____ _______________________
See

footnotes at end of table.




22.43
17.89
2.45
.80
1.65
.60
1.05
5.63
4.45
2.21
.57
.80
.84
1.30
.36
.30
.64
.94
.73
.51
.22
.45
2.80
.85
.68
.25
1.02
3.02
.76
.17
.15
. 20
.24
.94
.16
.24
.14
.40
1.32
3.99
.64
.55
.15
.40
.64
1.01
.40
.61
1.15
4.54
3.75
.79

Components

an d

Percent of
all items
December 1963

Housing____________ __________________________________
!3helter_________ ___________________________________
Rent_____________ ____ _______________ _____ ____
Hotels and motels_______________________________
Homeownership___ _____________________________
Purchase and financing............. ................. ........
Home purchase.. .1________ ________ _____
Mortgage interest____________ _____ ______
Taxes and insurance_________________________
Real estate taxes____ _____ _______________
Property insurance.__________ ___________
Maintenance and repairs_____________________
Commodities...I_________________________
Services____________________________ ____
Fuel and utilities______ _____________________________
Fuel oil and coal________________________________
Fuel oil__________ ____ ____________ _____ ___
Coal____ _______ ______________________ _____
Gas and electricity______________________________
Gas_____ ____________ _____________________
Electricity.................................................... ..........
Other utilities_________________ _______ _____ ___
Telephone________ ______ ____________________
Water and sewerage___________________ ____ _
Household furnishings and operation_________________
Textile housefumishings!_______________ _________
Furniture__________ _____ ______________________
Bedroom suite______________________________
Living room suite...________ ________________
Other”priced items.. _________ __________ . . .
Floor coverings____________________ _____________
Rugs, soft surface____ ____ ____________ ____ _
Other priced items.. . . . __________________
Appliances______ ____ __________ _______________
_____________________
’ 'Refrigerators_____ .
Other priced items____________ _______ ______
Other housefumishings _____ _______________ . .
Housekeeping supplies________ _____________ ..
Housekeeping services_________ __________________
Domestic service.___________________________
Babysitter__ ________________ ____________
Postage__ ___________________________ _____
Other-priced items______________ _______ _____

33.23
20.15
5.50
.38
14.27
9.11
6.28
2.83
2.13
1.72
.41
3.03
.98
2.05
5.26
.73
.67
.06
2. 71
1.30
1.41
1.82
1.38
.44
7.82
.61
1.44
.28
.28
.88
.48
.34
.14
1.36
.28
1.08
.83
1.55
1.55
.26
.29
.23
.77

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.............. .............. ................ ..................
Footwear..".!..... ...............................................................
Street shoes, men’s...
.
__________ ______ ____
Street shoes, women’s.. . . _____________________
Other priced items_________________________ _____
Other apparel_______________________ _______________
Commodities____________ ____ _______ ____________
Services_______________________________ _________
Dry cleaning__________ _____________________
Men’s suit_________ _____ _______________
Women’s dress____________ ____ __________
Other priced items__________ _______________

10.63
2.86
2.21
.36
1.85
.65
4.08
3.23
.28
.50
.39
2.06
.85
1.51
.26
.26
.99
2.18
.71
1.49
.79
.44
.35
.68

Transportation_______ _________________________________
Private transportation______________________________
Autos and related goods_____ _____ ______________

13.88
12.64
9.02
5.02
2.55
2.47
3.28
3.05
.23
.72
3.62
.98
2.64
1.42
.37
.04
.18
.63
1.24
.78
.14
.07
.20
.05

New cars_______ ____ ____ _______________
Used cars.................. ............ .......................
Gasoline and motor oil................................... ......
Gasoline________________________ ________
Motor oil________________ ____ ______ ____
Auto parts___________________ ____ _______ _
Automobile services___________________ ______
Auto repairs and maintenance.......................
Other automobile expenses________ _______
Auto insurance.."_____________________
Registration fees.____________________
Drivers’ license______ ________________
Parking fees_________________________
Auto financing chargesJ
..........................
Public transportation_______I____ I___ ____ _____ ____
Local transit____________________________________
Taxicabs_________________________________ ______
Train fares___________ ______________ ___________
Airplane fares............... ............... ..............................
Intercity bus fares........................................................

BLS HANDBOOK OF METHODS

80
T a b l e 2.

C o n s u m e b P r ic e I n d e x ( N e w S e r ie 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 ia 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 * D e c e m b e r 1963— Continued
Percent ol
all Items
December 1963

Components

Health and recreation................. ............................................
Medical care........................................................................
Drugs and pharmaceuticals_______ ______ _________
Professional services...............
Family doctor, house visit
Family doctor\office visit
Dentists’ tees
Other priced Items

..........

.

.

7 . . . . .
. . . .

Nonhospital services_________________________
Toilet goods
Services________________________________________
Rending and recreation__
_ .......
Recreation
.................................
Recreational goods................................................
TV sets
.............
Other priced Items..........................................

19.45
5.70
1.14
.50
.64
2.59
.12
.77
.29
.86
.55
.36
1.61
.66
.71
.24
2.75
1.52
1.23
.51
.72
5.94
4.36
2.78
.63
2.15

Percent o
all Items
December 1963

Components

Health and recreation—Continued
Reading and recreation—Continued
Recreation—Continued
Bowling fees_____________________________
Other priced items_______________________
College tuition
._
. .
. . . _____ .
Other priced items___________________________

Other priced items___________________________
Alcoholic beverages______________________________
Beer_______~_______________________________
Away from home

M iscella n eou s 3 3

_

an d

. .......

. . . . . . . . . .

1.58
.38
.36
.84
1.58
.50
.23
.85
5.06
1.89
1.74
.15
2.64
1.06
.78
.80
.53
.28
.12
.13
.38

» For a description of the new series, see The Consumer Price Index,
(.Revised January 1981). 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.
* 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 CPI,"
Monthly Labor Review, November 1964, pp. 1299-1300.
* Personal financing charges other than mortgage interest and auto financ­
ing.

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




CONSUMER PRICES

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.

Sampling
A complicated index such as the CPI must be
based on a whole complex of samples.1 A sample
6
o f 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 pair­
ing 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



81

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.
City Sample. 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 (PSU’s) 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 PSU’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
1,400,000;
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
Roe Goodman and Leslie Kish in the September
w For 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.

82

BLS HANDBOOK OF METHODS

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 BLS 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.1 The
7
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.
Sampling of Items. 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
1 The selection of the CES sample Is discussed In chapter 8
7
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 o f 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
specifications.
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

CONSUMER PRICES

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 Sampling. 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




83

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 SMSA (central business district,
neighborhood centers, and suburbs) and, in
some cases, to multiunit and independent
establishments.
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

84

BLS HANDBOOK OF METHODS
T

a b l e

4.

L

is t

o f

I

t e m s

P

r ic e d

f o r

t h e

R

e v is e d

C

o n s u m e r

Sample A

EC-1
EC-2

Bakery products_________

EC-3
. 3A

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

3B

Pork.

3C

Other meats.

EC-4

Poultry___

EC-S

Fish...........

EC-fl

Dairy products

EC-7

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

EC-8

EC-9

EC-10
EC-11
EC-12
EC-13
EC-14

EC-15

EC-16
EC-17
EC-18

EC-19

EC-20

See footnotes at end of table.




I

n d e x

a s

o f

D

e c e m b e r

19 63

Sample B

Corn 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 cut____ __________
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.
Llverwurst 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__________ ___________
Tomatoes__________________________
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..............................
Eggs, fresh, large, Grade A ---------------Fats and oils.................
Margarine, colored......... ......................
Salad dressing, Italian...........................
Sugar and sweets..........
Sugar, white, granulated......................
Chocolate bar, plain milk.....................
Coflee, can or bag..................................
Nonalcoholic beverages.
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:
Lunch..............................................
Breakfast.........................................
Between meal snacks:
Coflee, cup......................................
Carbonated beverages, cup............
Frankfurter on roll..........................
Ice cream, dish...............................
Housing:
Shelter:
Rent.................................................... Rent of house or apartment..........................
Hotel, motel room rates................................
Homeownership:
Homo 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........................... .
Services......................

r ic e

Priced items

Groups, subgroups, expenditure classes

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

P

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.
Tomatoes.
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:
Lunch.
Dinner.
Between meal snacks:
Coflee, cup.
Carbonated beverages, cup.
Pie, slice.
Candy bar.
Rent of house or apartment.
Hotel, motel room rates.
Home purchase.
Mortgage 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.

85

CONSUMER PRICES
T a b l e 4.

L is t

of

I t e m s P r ic e d

for th e

R e v is e d C o n s u m e r P r ic e I n d e x

Sample A

EC-22

Housing—Continued
Fuel and utilities.

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

EC-23

Furniture.

EC-24

Floor coverings.

EC-25

Appliances____

EC-20

Other housefurnishings.

EC-27

Housekeeping supplies.

EC-28

Housekeeping services.

EC-29

Apparel and upkeep:
Men’s and boys’ apparel:
Men’s apparel______

EC-30

Boys’ apparel.................

EC-31

Women’s and girls’ apparel:
Women’s apparel______

EC-32

EC—
33

Girls’ apparel.

Footwear.

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...........

EC-34
EC-35

Services.........

See footnotes at end of table.




ecember

1963— Continued

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.
Cribs.
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.
Nalls, 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.
Reuphoistering 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
ties.
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_________________________ ____
Men’s:
Shoes, street, oxford, 2 qualities........................
Women’s:
Shoes, street, pump, 2 styles.
Shoes, evening, pump...........
Shoes, casual, pump_______
Houseslippers, scuff..............
Children’s:
Sneakers, boys’, oxford type.

Other apparel:
Commodities.

D

Priced items

Groups, subgroups, expenditure classes

EC-21

as o f

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

Sport coats, wool or wool blend.
Undershorts, cotton.
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 blend.
Dresses, street, cotton.
Blouses, cotton.
Bathing suits, one piece.
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, work, high.
Shoes, street, pump, 2 styles.
Shoes, evening, pump.

Shoes, oxford.
Dress shoes, girls’, strap.
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.

BLS HANDBOOK OF METHODS

86

T a b l e 4.

L is t

of

I t e m s P r ic e d

for t h e

R e v is e d C o n s u m e r P r ic e I n d e x

as of

D

ecember

1963— Continued

Priced items
Groups, subgroups, expenditure classes
Sample A

EC-36

Transportation:
Private:
Auto purchase-

EO-37

Gasoline and motor oil.

EC-38
EC-39

Auto parts........................................
Automobile services:
Auto repairs and maintenance.

EC-40

Other automobile expenses.

EO-41

EC-42

Public transportation.

Health and recreation:
Medical care:
Drugs and prescriptions.

EC-43

Professional services.

EC-44

Hospital services and health in­
surance:
Hospital services..................
Health insurance ’ .

New cars:
Chevrolet, Impala, 2-door, hardtop.................. Chevrolet, Impala, 2-door hardtop.
Chevrolet, Chevelle, 2-door hardtop................ Ford, Falcon, Future, 4-door sedan.
Ford, Galaxie 600,2-door hardtop..................... Ford, Galaxie 600,2-door hardtop.
Plymouth, Fury III, 4-door sedan.................... Pontiac, Catalina, 4-door sedan.
Rambler, Classic 660,4-door sedan................... Volkswagen, Deluxe, 2-door hardtop.
Used cars:
2 years old, Chevrolet and Ford_____________ 2 years old, Chevrolet and Ford.
3 years old.........do........ .................................... . 3 years old,
Do.
4 years old.........do........ .................................... . 4 years old,
Do.
Do.
6 years old.........do............. —............................ . 5 years old,
Gasoline, regular and premium.............................. . Gasoline, regular and premium.
Motor oil, premium.................................................. Motor oil, premium.
Tires, tubeless, retread............................................ . Tires, tubeless, new.
Chassis lubrication, complete............................... .
Motor tune-up.............. ......................................... .
Automatic transmission repair............................... .
Auto insurance rates, liability and physical
damage.
Auto financing charges *.......................................... .
Auto registration and inspection fees..................... .
Driver’s license fees..................................................
Parking fees, private and municipal...................... .
Local transit fares.................................................... .
Taxicab fares............................................................ .
Railroad fares, coach.................................................
Airplane fares, chiefly coach................................... .
Bus fares, intercity.................................................. .

EC-46

EC-46

Personal care services.

See footnotes at end of table.




Water pump replacement.
Replacing muffler.
Front end alignment.
Auto insurance rates, liability and physical damage.
Auto financing charges.9
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.

Over-the-counter items:
Multiple vitamin concentrates........................ . Aspirin compounds.
Liquid tonics................................................... . Cough syrups.
Cold tablets or capsules..................................... Adhesive bandages, package.
Prescriptions:
Anti-infectives:
Penicillin G buffered tablets..................... . Tetracycline capsules.
Sulflsoxazole tablets.
Sedatives and hynotics:
Phenobarbital tablets.............................. . Secobarbital sodium capsules.
Ataractics:
Chlordlazepoxide-hydrocbloride capsules.. Meprobamate tablets.
Antispasmodics:
Propantheline Bromide tablets.................. Phenobarbital and belladonna extract.
Cardiovasculare and antihypertensives:
Reserpine tablets........................................ . Crystalline digitoxin tablets.
Chlorothiazide tablets.
Antiarthritics...................................................... Prednisone, tablets.
Cough preparations:
Terpin hydrate with codeine, elixir.
Family doctor, office visits_____________________ Family doctor, office visits.
Family doctor, house visits________ _____ _______ Family doctor, house visits.
Pediatric care, office visits..... ..................... .......... . Obstetrical cases.
Psychiatrists, office visits......................................... Chiropractors and podiatrists, office visits.
Routine laboratory tests........................................... Herniorrhaphy, adult.
Examination, prescriptions and dispensing of Examination, prescriptions and dispensing
eyeglasses.
eyeglasses.
Fillings, adult, amalgam, one surface.................... . Fillings, adult, amalgam, one surface.
Dentures, full upper................................................ . Extractions, adult.
Daily service charges:
Semiprivate room............................................. .
Private room..................................................... .
Hospital services:
Daily service charges, semiprivate room..........
Daily service charges, private room............... .
Operating room................................................. .
Nonhospital services:
Family doctor, office visit.................................
Surgeon’s fees, tonsillectomy/adenoidectomy..
Prescriptions and drugs.....................................
Retained earnings (overhead)..................................

Personal care:
Toilet goods.

Sample B

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, herniorrhaphy, adult.
Obstetrical cases.
Prescriptions and drugs.
Retained earnings (overhead).

Toothpaste, standard dentrifloe............................. . Toilet soap, hand milled.
Hand lotions, liquid................................................ . Shaving cream, aerosol.
Face powder, pressed.............................................. . Deodorants, cream or roll-on.
Cleansing tissues.................................................... . ■Home permanent refills.
Men’s haircuts......................................................... . Men’s haircuts.
Shampoo and wave sets, plain............................... . Shampoo and wave sets, plain.
Women’s haircuts_____________ ________________ Permanent waves, cold.

87

CONSUMER PRICES
T a b l e 4.

L is t

of

I t e m s P r ic e d

fo r t h e

R e v is e d C o n s u m e r P r ic e I n d e x

a s of

D

ecember

1963— Continued

Priced items
Groups, subgroups, expenditure classes
Sample A

EC-47

EC-48

EC-49

EC-50

Health andrecreation—Continued
Reading and recreation:
Recreation:
Recreational goods-----

Recreational services.

Reading and education.

Other goods and services:
Tobacco products___

EC-51

Alcoholic beverages.

EC-52

Financial and miscellaneous personal
expenses.

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:
Adult.
Children’s.
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.

Funeral services, adult_____________ ___ ________ Funeral services, adult.
Bank service charges, checking account.................. Legal services, short form will.

• 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.

s 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.

Formula. In the absence of major weight revi­
sions, and ignoring the problems of sampling, the
index formula is most simply expressed as:
( 1)

2

\P oQ .o)

or by its algebraic equivalent, the dollar weighted
average of price relatives:

Calculation Procedures
(la)
The index is a time series. As previously ex­
plained,TtTs~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.*8
1
The Bureau began publication of seasonally ad­
justed indexes in 1966, for selected components
which show a significant seasonal pattern of price
change.1
9



Iio=-

Xioo

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

1 For a discussion of the problems involved in using varying
8
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.
1 Factors used to compute seasonally adjusted indexes are avail­
8
able on request.

BLS HANDBOOK OF METHODS

88

items, and the current index is computed by a
chain computation procedure, as shown below:
(2)

2 (p0q0
)

(p -sQ )
_a

xioo

where q is a derived composite of the annual quantities
purchased 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).

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

September
price

October
price

$0.75
.80
1.00

$0.7725
.82
1.02

Pork chops___
Ham________
Bacon.........

(PiQ.a) = ( P i- lfi.

In practice, then, the index formula is as follows:
(Dec. 1963 (Change from Dec.
Index)
1963 to month i—
1)

(3)

Iuo

2(p0 ) Az(^;_,ga
?0
)
(Change from month
i— to month i)
1

2

( £
x - ^ S(^-lga)

>

Xioo

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




$15.00
8.00
10.00

$15.45
8.20
10.20

33.00

1.03
1.025
1.02

Total........

33.85

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

The (p 0_ or (p'i-sqa) base “ weights” for a
g0)
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 g’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!

October
Ratio
September
October-icost
cost weight
September
weight
(Sept. X
ratio)

Pork chops______
Ham____________
Bacon___________
Total............

Implied September September
cost
quantity
price
weight
(pounds)

October
price

$15.00
8.00
10.00

$0.7725
.82
1.02

20
10
10

$0.75
.80
1.00

33.00

October
cost
weight
$15.45
8.20
10.20
33.85

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
follows:
October cost weight________ $33.85X100=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

CONSUMER PRICES

89

BLS HANDBOOK OF METHODS

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=100.2 This means that current prices are
0
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~yr ar, the index calcula­
e
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.2
1
Average bills for specified quantities of gas and
electricity and average prices o f 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
calculation.
“ The index base was 1 9 4 7 -4 9 = 1 0 0 from 1953 through 1961,
1935-39 = 100 from 1935 through 1952, and 1913 = 100 from
1913 through 1934.
" F o r a more detailed discussion, see article by Doris P.
Rothwell, “ Calculation of Average Retail Food Prices,” Monthly
Labor Review, January 1965, pp. 61-66.

90

BLS HANDBOOK OF METHODS

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 regionaLoffices pre­
pares and mails a press reteaseforeach 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 Re­
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 Estimated 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.2
2
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 Bu-2
3
2 See article by Francis S. Cunningham, “ The Use of Price
3
Indexes In Escalator Contracts,” Monthly Labor Review, August
1963, pp. 948-952. Reprint 2424.

CONSUMER PRICES
T a b l e 5.
D efinition

of the

Sum m ary

of

C h a r a c t e r s rics

of th e

91
CPI B e g in n in g 1964

I ndex

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

Formula (Simplified expression).

Base period...... .......................................
Duration............................. ............
Definition of index expenditure weights.

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

(£ )

2p.-j0.

1957-69=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 opulation C o verage of
E xpenditure Su r vey

Place of residence..................
Family size.................................. '* .......
Occupation...............................
Length of employment......................... .
Income______ ______________________

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 halfof tota lfamiW
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 ity C overage

Population weights................................
Sample of priced cities_______________
Published indexes___________________

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.

I tem Sample

Basis of sample selection........................
Basis for allocation to priced items____
Commodity coverage________________
Number of items priced_____________
Pricing cycle.......................................... .

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 I d re­
maining cities.

R epo rt er C o v e ra ge

Location.............................................
Number of reporters..... ......... , ___ ”
Number of quotations............... ......
Pricing technique............ ................

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
ditions.

reau is attempting to measure the sampling error
in the index.2
3
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*
** Preliminary estimates of sampling error were computed
and published in Measurement of Sampling E rror in the Con­
sumer Price In d ex: F irst R esults, 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
reporting.
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

92

BLS HANDBOOK OF METHODS
CON SU M ER

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.2 Nevertheless,
4
some residual effects of quality changes on quoted
prices undoubtedly do affect the movement of the
CPI either downward or upward from time to
time.2
5
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

PRICES

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.*
M Hoover, Ethel D.f “ The CPI and Problems of Quality
Change,” M on thly L abor R eview , November 1961, pp. 1175-1185.
Reprint 2378, and Larsgaard, Olga A., and Louise J. Mack,
“ Compact Cars In the Consumer Price Index,” M onthly Labor
R eview , May 1961, pp. 519-523. Reprint 2368.
* See testimony of Evran Clague In Hearings before th e Sub­
com m ittee on Econom ic S tatistics, Joint Economic Committee,
Congress of the United States, 8 7 /1 , Part II, p. 588, Washing­
ton, May 1, 2, 3, 4, and 5, 1961.

Technical References
Number
1. Cunningham, Francis S. “ The Use of Price Indexes in Escalator Contracts,” Monthly Labor 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
(W PI). 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,” Monthly Labor 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,” Monthly Labor 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,” Monthly Labor Review, December 1948, pp. 631-637.
“ The Rent Component of the Consumers’ Price Index. Part II— Methodology of Measure­
ment,” Monthly Labor Review, January 1949, pp. 60-68. Combined in Reprint No. 1947.
Part I discusses trie 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 comperfsating for depreciation of quality caused by aging.
5. Jaffe, Sidney A. “ The Statistical Structure of the Revised CPI,” Monthly Labor 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 Consumer Price Index— Technical Questions and Practical Answers. 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.




CONSUMER PRICES

Technical References—Continued
ffumber

7. Lamale, Helen Humes. “Housing Costs in the Consumer Price Index, Part I ,” Monthly Labor
Review, February 1956, pp. 189-196. “ Housing Costs in the Consumer Price Index. Part II,
Monthly Labor 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,” Monthly
Labor 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
cars.
9. National Bureau of Economic Research. The Price Statistics of the Federal Government: Revieiv,
Appraisal, and Recommendations, (Washington, National Bureau of Economic Research, General
Series, Number 73, 1961), 496 pp. Also appears in Government Price Statistics: Hearings,
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 NBER
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,” Monthly Labor 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 Consumer Price Index: Pricing and Calculation 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,” Journal of the American
Statistical Association, 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. Consumers’
Price Index, 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 Price Statistics: Hearings, 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.




93

BLS HANDBOOK OF METHODS

94

Technical References—Continued
Number

15. U.S. Congress, Joint Committee on the Economic Report. The Consumers' Price Index— Report
of the Joint Committee on the Economic Report on the Consumers' Price Index of the United States
Bureau of Labor Statistics, 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.
16. U.S. Department of Labor, Bureau of Labor Statistics. The Consumer Price Index (Revised January
1 9 6 4 ): A Short Description, (U.S. Department of Labor, Bureau of Labor Statistics, September
1964).
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.
17. ------- . Seasonal Factors, Consumer Price Index: Selected Series, June 1958-M ay 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.
18. ------- . “ Taxes and the Consumers’ Price Index,” Monthly 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.
19. -------. Interim Adjustment of Consumers' Price Index: Correction of Neiv Unit Bias in Rent Component
of Consumers’ Price Index and Relative Importance of Items, 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.
20. ------- . Consumer Prices in the United States, 1953-58: Price Trends and Indexes, 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.
21. U.S. Office of Economic Stabilization. Report of the President’s Committee on the Cost of Living,
(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.
22. Wilkerson, Marvin. Measurement of Sampling Error in the Consumer Price Index: First Results.
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.




95

CONSUMER PRICES

Technical References—Continued
Number

23. Wilkerson, Marvin. Sampling Aspects 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
tions.
24. -------. “ The Revised City Sample for the Consumer Price Index,” Monthly Labor 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

P.

R

othw ell

and

C arlyle

P.

S t a l l in g s

Office of Prices and Living Conditions




Chapter 11. Wholesale Prices
Background

The Wholesale Price Index (W PI) is the
oldest continuous statistical series published by
the Bureau of Labor Statistics (BLS) and one
of the oldest in the Federal Government. It was
first published in 1902, and covered the years
1890-1901. The origins of the index are as­
sociated with a resolution of the U.S. Senate
in 1891, which authorized the Senate Com­
mittee on Finance to investigate the effects of
the tariff laws “ upon the imports and exports,
the growth, development, production, and prices
of agricultural and manufactured articles at
home and abroad.” 1
The index published in 1902 on the base
1890-99 was an unweighted average of price
relatives and included from 250 to 261 com­
modities. 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 com­
pleted at the end of 1914, when a system of
weighting was introduced and the index was
recalculated back through 1890.1 By 1940, the
2
number of commodities had increased to ap­
proximately 900, based on about 2,000 individual
price quotations. Then, in 1952, the most ex­
tensive revision in the history of the index was
completed.3 The number of commodities and
quotations was doubled, weights were based on
1947 Censuses and changes were made in the
calculation method. Some changes in classifica­
1 W h olesa le P rices, W a ges, and T ransportation, Senate Re­
port No. 1394, "T he Aldrich Report,” Senate Committee on
Finance, Congress of the United States, March 3, 1893, Part
I, (52d Cong., 2d session), Government Printing Office (1893);
and C ou rse o f W h olesa le P rices, 1890-1901, Bulletin of the
Department of Labor, No. 39, March 1902, pp. 205-209.
2 See also, Allan D. Searle, “Weight Revisions in the Whole­
sale Price Index, 1890-1960,” Monthly Labor Review, February
1962, pp. 175-182.
3 A large number of the newly introduced commodity
prices were carried back to 1947. The presently published
index contains the new commodities for the period 1947-51
and displaces the older, less comprehensive index on the 1926
base published for the same period.




tion were made also, including expansion to the
present 15 major groups. A major reclassifica­
tion was implemented in January 1967, when
the 8-digit classification structure was initiated.
Also at that time, new weights from the 1963
industrial censuses were introduced.
By January 1971, the number of commodities
had increased to more than 2,500, the number
of price quotations had increased to over 8,000,
and the index had become increasingly repre­
sentative of general primary market price
changes.

Description o f Survey
Concepts

Throughout its history, the WPI has* been a
measure of price changes for goods sold in
primary markets in the United States. “ Whole­
sale” as used in the title of the index refers to
sales in large quantities, not prices received by
wholesalers, jobbers, or distributors.
From its inception, the index has been con­
sidered a general purpose index designed to
measure the general price level in other than
retail markets. From the beginning of the in­
dex, however, attention was directed to some
specific needs of users, and indexes for indi­
vidual commodities and for major commodity
groups were published. As early as 1903, two
special group indexes by stage of processing—
Raw Commodities and Manufactured Com­
modities— were published “ to meet the wishes
of students of price statistics.” In recent years,
emphasis has been placed on the development
of more subdivisions within major groups and
special combinations of indexes such as by
stage of processing and by durability of
product.
Most of the quotations reported to the Bu­
reau are the selling prices of representative
manufacturers or producers, but some prices
97

98

BLS HANDBOOK OF METHODS

are those quoted on organized exchanges (spot
prices) or at central markets. Prices for im­
ported commodities are those received byimporters— the first commercial transaction
involving the commodity 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., commodities included in the index are
defined by precise specifications which in­
corporate their principal price-determining
characteristics.* So far as possible, prices are
4
3
2
f.o.b. production point, and refer to sales for
immediate delivery. Prices applicable to longrun contracts and “ futures” are usually not
included.
Universe

The WPI universe consists of all commodi­
ties 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 manu­
factured and processed goods and the output of
industries classified as manufacturing, agricul­
ture, forestry, fishing, mining, gas and elec­
tricity, public utilities, and goods competitive
with those made in the producing sector, 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-company transfers). Goods
sold at retail by producer-owned retail estab­
lishments also are excluded because they
conceptually belong to a retail (consumers')
universe, rather than to primary market
transactions.
Civilian goods normally purchased by the
Government are in the universe, but military
goods are not. Government sales of some com­
4 An 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.”
3 The prices used in the index through 1951 were the simple
arithmetic averages of prices for all Tuesdays in the month.
From January 1952 through December 1966, Tuesday of the
week containing the 15th was the pricing date.




modifies (e.g., electric power) are included if
they can be considered competitive with free
market sales.
Prices

To the extent possible, the prices used in
constructing the index are those that 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 trans­
formed into semifinished and finished goods,
the resulting products 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 con­
taining the 13th day; 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 transaction prices. Companies are re­
quested to report prices less all discounts, al­
lowances, rebates, free deals, etc., so that the
resulting net price is the actual selling price
of the commodity for the specified basis of quo­
tation. The Bureau periodically emphasizes to
reporters the need to take into account all dis­
counts and allowances. However, list or book
prices are used if transaction prices are un­
obtainable.
Prices are generally f.o.b. production or cen­
tral marketing point to avoid reflection of
changes in transportation costs. Delivered
prices are included 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 excluded since they are not
considered part of the price, but import duties
are included as part of the selling price of im­
ported 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 im­
portant when new commodities are introduced,

WHOLESALE PRICES

but even when specifications of existing com­
modities 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 its
selling terms is substituted for the earlier series
by direct comparison or by linking. The objec­
tive of the linking procedure is to insure that
the index will reflect only those changes due
to actual price differences.6 Each time a change
in the item priced occurs, the Bureau appraises
the significance of the specification change to
ascertain whether an actual price change oc­
curred. If the specification change is minor and
does not involve price-making factors, the sub­
stitution is effected by direct comparison, and
any reported price change between the old and
the new specification is reflected in the index.
If changes in specification are major, and if
either no real price change occurred or no in­
formation can be obtained concerning the value
of the difference in specification (perhaps in­
dicative of a change in quality), the substitu­
tion is made by linking and no change is
reflected in the index. In this case, any reported

6 The following example illustrates the linking procedure:
The September price for a certain machine used in the cal­
culation 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 considered 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 Sep­
tember price of former model plus $186.20 increase in value
of motor and tires). The price comparison between Septem­
ber and October was based on the estimated September price
of $2,533.70 and the reported October price of $2,562.60. Thus
a 1.1-percent increase was reflected in the October index, but
the price change due to quality improvement (more powerful
motor and larger tires) was not reflected.




99

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 that
formerly sold at extra cost are not reflected in
the index. Conversely, price changes attrib­
utable to deletion of equipment which was
formerly standard are not treated as decreases.
In the event production of a specified com­
modity is discontinued by a reporter, or its im­
portance is reduced, the Bureau collects price
data for a similar 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 orig­
inal price is taken as a measure of the quality
difference between the two items.
Linking is also used for the addition to or
deletion of commodities or groups of commodi­
ties from the index; the addition to or deletion
of a company report from the sample of com­
panies priced, or, on occasion, a change in the
source of price. Whenever a new commodity
is added to an existing commodity group, link­
ing of the new item to any one of the existing
items is not pertinent. 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 that drop out
of the index.

BLS HANDBOOK OF METHODS

100
W h o l e s a l e P r ice I n d e x

R e l a t iv e I m p o r t a n c e , N u m b e r o f I t e m s a n d P r ic e Q u o t a t i o n s for M a jo r G r o u p s a n d S u b gr o ups

Grouping

Relative importance
in total 1963 weights
December 1970

All commodities ________________________________
01 Farm products _____________________________________
01-1 Fresh and dried fruits and vegetables_________
01-2 Grains ____________________________________________
01-3 Livestock _________________________________________
01-4 Live poultry _____________________________________
01-5 Plant and animal fibers ________________________
01-6 Fluid milk _______________________________________
01-7 Eggs ---------------------------------------------------------------------01-8 Hay, hayseeds, and o ilseed s_____________________
019 Other farm products _________________________
02 Processed foods and f e e d s ________________________
021 Cereal and bakery products ________________
02-2 Meats, poultry, and fish ________________________
02-3 Dairy products ___________________________________
02-4 Processed fruits and vegetables _______________
02-5 Sugar and confectionery________________________
02-6 Beverages and beverage m aterials_____________
02-7 Fats and oils 1 ____________________________________
02-8 Miscellaneous processed foods _________________
029 Manufactured animal f e e d s __________________

100.000
10.038
1.176
1.198
2.851
.255
.437
2.081
.510
.694
.835
16.298
2.017
4.153
2.300
.866
1.268
2.118
.627
1.185
1.765

03 Textile products and a p p a r e l_____________________
031 Cotton products _______________________________
03-2 Wool products ___ ______________________________
03-3 Manmade fiber textile p rodu cts________________
03-4 Silk' products1 ___________________________________
03-5 Apparel _________
_____________________ .________
03-6 Textile housefumishings ________________________
037 Miscellaneous textile products ______________

6.875
1.086
.346
1.320

04 Hides, skins, leather, and related p rodu cts______
04- 1 Hides and skins _______________________________
04-2 Leather ___________________________________________
04-3 Footwear _________________________________________
044 Other leather and related produ cts_________

1.239
.077
.179
.694
.289

05 Fuels and related products, and p o w e r__________
051 Coal ___________________________________________
05-2 Coke _ ___________________________________________
05-3 Gas fuels _________________________________________
05-4 Electric power ___________________________________
05-6 Crude petroleum and related products1 _______
057 Petroleum products, refined _________________

7.322
.700
.092
.677
1.767
.628
3.459

06 Chemicals and allied products ____________________
061 Industrial chemicals ____ _____________________
06-2 Paint and paint materials 1 ______________________
06-3 Drugs and pharmaceuticals _____________________
06-4 Fats and oils, inedible __________________________
06-5 Agricultural chemicals and chemical products
06-6 Plastic resins and materials ____________________
067 Other chemicals and allied products _______
07 Rubber and plastic products _____________________
071 Rubber and rubber products _______________
072 Plastic products ______________________________
08 Lumber and wood p rodu cts_______________________
08- 1 Lumber ________________________________________
08-2 Mi 1 work _________________________________________
1
08-3 Plywood __________________________________________
084 Other wood products ________________________
09 Pulp, paper, and allied p rodu cts_________________
09- 1 Pulp, paper, and products, excluding building
paper and board ______________________________
092 Building paper and b o a r d ____________________

5.937
1.814
.698
.807
.189
.545
.366
1.517

3.631
.359
.132

December 1966

Number of items
and price quotations,
January 1971
Price quotations
Items

7
9
193
22
40
15
35
11
18
17
20
15

8,017
143
48
8
12
4
29
8
4
13
17
664
79
99
42
172
23
104
24
67
54

183
38
17
45

519
93
39
108

67
11
5

235
32
12

49
12
11
17
9

100
12
12
55
21

67
7
6
2
18
6
28

275
39
6
2
173
22
33

335
97
30
107
7
43
10
41

550
101
71
167
7
43
38
123
335
163
172
367
224
68
41
34
266

100.000
10.637
1.171
1.357
3.086
.332
.553
2.001
.576
.780
.781
16.533
2.038
4.404
2.275
.856
1.192
2.047
.603
1.183
1.935
7.149
1.152
.403
1.488
.021
3.562
.384
.139
1.264
.097
.196
.667
.304

2,503
101
30
8
12
4
19

7.130
.439
.070
.691
1.808
.614
3.508
6.378
1.968
.706
.888
.163
.675
.456
1.522

8
4

!
i
|

2.346
1.475
.872
2.483
1.259
.686
.399
.140
4.796

2.339
2.339
2.418
1.215
.658
.416
.129
4.877

88
45
43
77
50
15
7
5
74

4.655
.141

4.719
.158

66
8

222
44

10 Metals and metal p rodu cts________________________
101 Iron and steel _________________________________
10-2 Nonferrous metals ______________________________
10-3 Metal containers ________________________________
10-4 Hardware ________________________________________
10-5 Plumbing fixtures and brass fittin gs___________
10-6 Heating equipment _____________________________
10-7 Fabricated structural metal p ro d u cts__________
108 Miscellaneous metal products _______________

13.456
4.797
3.376
.488
.575
.181
.256
1.771
2.012

12.799
4.547
3.205
.462
.548
.177
.254
1.716
1.890

375
108
92
8
44
13
21
41
48

1,295
404
275
28
144
65
90
133
156

11 Machinery and equipment ________________________
111 Agricultural machinery and equipment ____
11-2 Construction machinery and equ ip m en t_______
11-3 Metalworking machinery and equipm ent______
11-4 General purpose machinery and equipment ___
11-6 Special industry machinery and equ ip m en t_
_
11-7 Electrical machinery and equ ip m en t__________
11- 9 Miscellaneous machinery ____________________

12.508
.705
.885
1.545
1.770
1.641
4.348
1.614

627
55
49
105
78
76
182
82

2,098
202
169
319
287
226
633
262

12 Furniture and household d u rables_______________
12- 1 Household furniture _________________^_______
12-2 Commercial furniture ___________________________

3.529
.925
.449

12.110
.665
.807
1.469
1.665
1.506
4.462
1.536
3.584
.904
.416

93
21
7

498
151
34

See footnotes at end of table.




WHOLESALE PRICES

101

W h o l e s a l e P r ic e I n d e x — C o n t i n u e d
R e l a t iv e I m p o r t a n c e , N u m b e r o f I t e m s a n d P r ic e Q u o t a t i o n s for M a jo r G r o u p s a n d S ubgro ups
Relative importance
in total 1963 weights

Grouping

December 1970

December 1966

Number of items
and price quotations,
January 1971
Price Quotations
Items

.336
.862
.383
.574

.385
.891
.459
.529

9
28
10
18

54
134
52
73

13 Nonmetallic mineral Droducts
13-1 Glass 1
13-2 Concrete ingredients
13-3 Concrete products
13-4 Structural clay prods., excluding refractories
13-5 Refractories
13-6 Asphalt roofing
13-7 Gypsum products
............ ...
13-8 Glass containers
13-9 Other nonmetallic minerals

3.175
.388
.630
.922
.170
.159
.120
.090
.305
.392

3.040
.364
.612
.882
.168
.139
.127
.106
.273
.369

45
4
4
4
6
5
5
3
7
7

405
14

14 TransDortation eauiDment
14-1 Motor vehicles and equipment
14-2 Aircraft 7
14-4 Railroad eauiDment

7.441
6.642
.490
.310
2.556
.515
.863
.101
.371
.705

7.244
6.932
.312

77
62
10
5

152
132
10
10

2.498
.513
.802
.102
.386
.695

119
30
9
5
40
35

12-3
12-4
12-5
12-6

Floor coverings
Household appliances
Home electronic equipment
Other household durable goods

15 Miscellaneous Droducts
15-1 Toys, sporting goods, small arms, etc.
15-2 Tobacco products
15-3 Notions
15-4 Photographic equipment and supplies
15-9 Other miscellaneous products
"’

__

1 Subgroup index not published.
NOTE: Relative importance represents the basic value
weight of an item or items multiplied by the relative of
price change between the weight date and a later date, and

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

The classification system of the WPI follows
commodity lines. Products are grouped by sim­
ilarity of end-use or material composition,
rather than by industry of origin. The WPI
classification does not match the Standard In­
dustrial 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
classification. No single classification plan can



169
68
49
5
18

9
26
47

350

103
48

10
69
120

the result is expressed as a percentage of the total for all
commodities. The differences between the relative impor­
tances as of December 1966, the date of last major weight
change, and that of December 1970 are the result of pnce
changes only.

meet all of the requirements for wholesale price
statistics, but the plan adopted should be flex­
ible enough to facilitate regrouping of price
series to make special grouping indexes. In
January 1971, the index was made up of 15
major groups, 87 subgroups, 291 product
classes, 554 subproduct classes, and 2,503
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 ma­
terials, in addition to about 22 other special
group indexes.
Except for the stage of processing indexes,
these special groupings consist of rearrange­
ments of the WPI data into different combina­
tions of price series, so that the appropriate
prices and weights are those of the WPI. The
7 See table for the major groups and subgroups included in
the WPI.
8 The broad stages of processing are: Crude materials for
further processing; Intermediate materials, supplies, and
components; and Finished goods. Each of these is subdivided
further.

102

BLS HANDBOOK OF METHODS

stage of processing indexes, however, regroup
each item priced in the WPI according to the
amount of processing, manufacturing, or as­
sembling it undergoes before entering the
market. A commodity may appear in several
different categories in this scheme. Thus, 29
percent of the fresh vegetables (by valueweight) was assigned to crude foodstuffs and
feedstuffs for further processing and 71 per­
cent 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 trans­
action study made for the year 1958 by the
Office of Business Economics.

Data Sources and Collection Methods

Prices
Price data are collected by mail question­
naire, and reporting is voluntary and confi­
dential. Most prices are collected each month.
For a few commodities, for which price
changes are infrequent, the shuttle schedule is
mailed quarterly, but monthly prices are re­
quested. 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
Government agencies.
Price reporting is initiated, wherever possi­
ble, by a personal visit by a Bureau representa­
tive to the prospective respondent. Pricing of
additional products from established reporters
often is started by mail. In any event, a de­
tailed report describing all of the price-making
characteristics of the commodity is prepared
for each new price series. This commodity
price information sheet (BLS 1810) is shown
on pages 107 and 108. The form becomes a part
of the permanent record for the series. After
the initial collection of prices, monthly infor­
mation is collected by mail on a shuttle sched­
ule. (BLS 473, shown on pages 109 and 110.)



Weights

The price data are combined using weights
based on value of shipments. The major sources
of the value data are:
Bureau of Census --Census of Manufactures
Census of Mineral Industries
Bureau of M in e s___ Various publications, e.g.,
Minerals Yearbook
Department of
Agriculture______ Various publications, e.g.,
Agricultural Statistics
Bureau of
F ish eries_________Various publications, e.g.,
Fisheries of the United States

In addition, many other sources of data, such
as trade associations, are used. Import data
are obtained from a report of the U.S. Depart­
ment of Commerce, United States Imports for
Consumption.
Sampling

The monthy index is based on a judgment
sample of commodities, a sample of specifica­
tions (descriptions), and a sample of reporters.
The sample of commodities is chosen after a
review of the data of the industrial censuses
and other statistics of value of transactions.
Generally, the commodities chosen are those of
the largest shipment values. Starting with
January 1967, expansion of Industry Sector
Price Index sample coverage has been a major
influence in selecting new products for the
WPI. New items are not added until they have
become established in the market.8 They are
9
added, normally, in December of any year, and
have their first effect on the index in January.
Samples of specifications and of reporters
are selected after consultation with trade as­
sociations or other industry representatives
and with staff of other government agencies.
Individual commodity specifications are se­
lected also on the basis of net dollar sales. That
is, the “ volume seller” of the industry (not of
8 If new items are added before they become fully estab­
lished, 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.

103

WHOLESALE PRICES

the company) is preferred. The specification
describes not only the popular physical charac­
teristics but also the most common quality,
grade, level of distribution, and market. How­
ever, terms of sales (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 spe­
cification taken as standard; all other prices
are quoted as differentials from the standard.
The latter is true for some farm products such
as wheat and cotton. When no standard com­
modity basis exists, the specification to be
priced is selected with the help of industry
experts.
The number of reporters is determined, to
some extent, by the variation of price move­
ments among them and the degree of price
leadership. Whenever possible, a minimum of
three companies is obtained, so that data for
specified commodities can be published without
disclosure of information supplied by individ­
ual companies. For commodities with more
than one major production area and a definite
regional pattern, a larger sample is selected.
Among these commodities are waste materials
and building materials such as brick, cement,
and stone.
A comparatively small list of properly se­
lected commodities would produce a reliable
index, if only an All Commodities index were
desired. However, historically interest has been
great in indexes for groups of commodities and
for individual commodities. To meet these
needs, the Bureau has increased the sample in
order to provide more detailed indexes as well
as many special-purpose indexes.

An alternative formulation more closely ap­
proximates the actual computation procedure:
2. I, =

[2 (Q.P„)

(P ,/P „) / 2 Qap „] x 10°-

In this form, the index is a weighted average
of price relatives for each item (P i/P G The
).
expression (QaP0) represents the weights in
value form and the “ P” and “ Q” elements (both
of which originally relate to period “ a” but are
adjusted for price change to period “ o” ) are
not derived separately. Each value weight in­
cludes not only the value of items priced but
also the values of unpriced items whose price
movements are assumed to behave similarly.
When new weights are introduced, the index
with new weights is linked to the index con­
structed 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
(multiplying) the relatives for the separate
periods for which the data are precisely com­
parable. (For a somewhat more detailed treat­
ment, see chapter 10, Consumer Prices.)
Base Period

The Wholesale Price Index has been com­
puted on the government-wide standard refer­
ence base 1967=100 since January 1971.1 It
0
had been based 1957-59=100 from January
1962 through December 1970. Earlier bases
were 1947-49, 1926, and 1913. New items (or
new index groupings consisting primarily of
new items) introduced into the index after
1967 cannot be calculated on the 1967 base.
Such indexes are published with separate bases
related to the date of introduction.

Estimating Procedures

Weights
Formula and Calculation

In concept, the Wholesale Price Index is cal­
culated according to a modified Laspeyres
formula:
1.

I, = [2 QaP ,/2 Q ,P 0] x 100> where P0 is the price of
a commodity in the comparison period and P, is its
price currently. Q, represents the quantity shipped
during the weight-base period.




The WPI weights represent the total net sell­
ing value of commodities produced, processed,
1 Conversion of indexes from the 1967=100 base to the
0
1957-59 base may be accomplished by multiplying the
1967 = 100 based index by that item’s 1967 annual (12-month)
average (1957-59=100) and dividing the result by 100. Aver­
ages for 1967, calculated from 12 months’ data, and appear­
ing in the December 1968 issue of “ Wholesale Prices and
Price Indexes” may be used for base conversion.

104

BLS HANDBOOK OF METHODS

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 inter­
plant transfers, military products, and goods
sold at retail directly from producing establish­
ments also are excluded. Thus the definition of
the weights conforms to tho universe definition.
Each commodity price series is considered
representative of a class of prices and is as­
signed its own weight (the shipment value of
the commodity) plus the weights of other re­
lated commodities not directly priced but whose
prices are known or assumed to move sim­
ilarly.1 The assignment of price movements
1
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 ma­
chinery— it is not possible to obtain direct
measures of price movement. The weights for
such items are assigned to other commodities
or groups of commodities for which prices are
available. Usually, this assignment is made to
priced commodities that have a similar manu­
facturing process, on the assumption of similar
price movements. Price movements for attach­
ments and parts for certain machinery often
are imputed to the machine itself.
The Bureau’s policy is to revise the WPI
weights periodically when data from the in­
dustrial censuses become available.1 The
1
2
weights beginning in 1967 are based on the
1963 industrial censuses. The next revision
which normally would follow the 1967 indus­
trial censuses has been postponed because of
lack of resources. 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 re­
ported in the Annual Surveys of Manufactures.
Weights based on the 1954 census shipment
values were introduced in January 1958. From
1961 through 1966, weights were based on 1958
census values. In January 1967, new weights
from the 1963 industrial censuses were in­
corporated with the comprehensive reclassifica­
tion mentioned previously. Subsequent minor
redistributions of the weights have been made



each January to account for additions and
deletions of commodities.
The Bureau publishes the relative impor­
tance 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 impu­
tations, multiplied by the relative of price
change from the weight date to a later date;
the result is expressed as a percent of the
total for all commodities or for some index
grouping.1
3
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 prod­
ucts and processed foods groups, out of the
market seasonally, the price in off-season is
imputed from the combined movement of the
related commodities for which prices are avail­
able for the two periods being compared. For
other commodities, delinquent prices are held
unchanged from the preceding month.
Prices for some custom-made items are re­
ported to BLS as estimates. For example,
prices for fabricated structural steel for build­
ings 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 transportation and in­
stallation cost— for use in the WPI.

Analysis and Presentation

The monthly WPI is published first in a press
release, usually issued in the first week of the
1 Before the 1952 revision (calculated back to 1947), priced
1
items in the index were weighted only by their own market
values.
1 In general, the censuses are collected at 5-year intervals.
2
1 The use of relative importance data to construct indexes
3
for groups of products is discussed in the January 1970 issue
of “ Wholesale Prices and Price Indexes.” Relative importances
as of December for all WPI items are published in each
January issue.

WHOLESALE PRICES

month following the reference month. Indexes
are shown for all groups and subgroups as well
as for All Commodities, Farm Products and
Processed Foods and Feeds combined, and In­
dustrial Commodities. Analytic tables also are
included which show monthly percent changes
for the preceding 12 months for major group­
ings, and selected seasonally adjusted and un­
adjusted changes for some stage of processing
classifications. A brief description and analysis
of the causes of price movements are included.
The monthly detailed report, issued some time
after the press release, carries all data for
which wholesale price indexes are published,
including item indexes and all special group
indexes. Prices for many individual commodi­
ties also are included. Each quarter this report
includes a more comprehensive analysis than
that given in the press release. Annual sum­
maries appear in the monthly report as they
become available. In addition, numerous his­
torical tabulations at various levels of detail
are available on request.
The monthly indexes are published as final.
Beginning with data for January 1967, only
major corrections are made and published im­
mediately at the time the error is discovered.
Each year, after calculation of the December
index, all corrections reported during the year
are made, and the indexes for all 12 months
and the annual average are republished as
revised.
Selected seasonally adjusted indexes or per­
cent changes are published in the press release
and monthly detailed report. About 50 indexes
which historically show significant and con­
sistent seasonal movement are presented each
month seasonally adjusted and unadjusted. The
applicable season adjustment factors are avail­
able on request from the Bureau. These factors
are recalculated annually to include more
recent data, and the most recent set of factors
may differ somewhat from those previously in
use.

Uses and Limitations

The WPI is used by government and private
research agencies for many purposes, including



105

market analysis, escalation of long-term pur­
chase and sales contracts, and formulation of
monetary policies. It is used, as well, as an
indicator of economic trends.
A 1961 survey of users of the WPI revealed
that more than one-half use the All Commodi­
ties index as a general economic indicator.
About 40 percent use that index or its com­
ponents 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 purchase of material or
lease of industrial property are escalated ac­
cording to 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
managers. In most of these cases, it is not the
All Commodities index, but rather the group
indexes and the individual price series that
are employed. Buyers of commodities are able
to check both the amounts which they pay for
goods and the general movement of their pur­
chase 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 direction
and amount of change, and only incidentally to
measure actual selling prices.
The index, as a measure of general and speci­
fic price trends, also is used widely in budget
making and review, both in government and in
industry; in planning the cost of plant expan­
sion programs; in appraising inventories; in
establishing replacement costs; etc. Compo­
nents of the index also are used in LIFO
(Last-In, First-Out) inventory accounting by
some organizations.
Although the WPI often is used to measure
change in purchasing power of the dollar, it
should not be used to measure changes in gen­
eral purchasing power, prices at retail, securi­
ties prices, etc. Comparisons between the level
of the WPI, the Consumer Price Index, and
the indexes of prices of farm products show
relative change from a base period, but com-

106

BLS HANDBOOK OF METHODS

parisons of the index levels should not be used
as a measure of the actual margins between
farm prices and manufacturing or between
manufacturing and retail. Its commodity clas­
sification structure should be borne in mind
when using it to measure price changes for
industries, many of which make divers prod­
ucts not classified as their “ primary” products.1
4
Again, as in other measures, the WPI has
some limitations even in the field for which it
is conceptually designed. Segments of the in­
dex 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 WPI is based on a purposive, judgment
sample. The All Commodities Index can be as­
sumed 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.1 As
5
the economy has produced an increasing pro­
1 See Chapter 12 on Industry-Sector Indexes.
4
1 The sample of priced items doubled in 1952 to about 1,850
5
items and has increased to about 2,500 since then.




portion of fabricated finished goods (whose
price changes are relatively infrequent), over
the years, movement of the WPI has become
somewhat smoother. Currently, new products
are added each year. In earlier decades, there
were also major additions of large numbers of
new items at one time, in commodity areas
previously underrepresented. These sudden ex­
pansions could have made it appear that prices
had stabilized suddenly.
To the extent that quality improves (or de­
teriorates) over the years, the index errs when
no adjustment is made. However, the Bureau
makes suitable adjustments whenever possible.
Assuming quality improvement, the index
would have an upward bias if direct compari­
son were made between unimproved and im­
proved 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.

107

WHOLESALE PRICES
BLS 1810

THIS FORM WILL BE HELD IN CONFIDENCE

B udget Bureau No. 44-R 0602

U.S. DEPARTMENT OF LABOR

E xpiration date:

1 2 /3 1 /7 3

B U R E A U OF L A B O R S T A T I S T I C S
Washington, D.C.

20212

Code No.

Commodity

COMMODITY PRICE INFORMATION SHEET

Mfr. | |

Plant or division

Other
(S p e e lly )

Address
(Street)

(C ity and State)

Information authorized bv

(Z ip co d e)

Title

Information furnished by

Title

Mail schedule to

Title
(R eporter)

Address

(C ity and State)

(Street)

1.

2.

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

PRICE HISTORY FOR COMMODITY DESCRIBED ABOVE

Price

Date

3.

(Z ip code)

Remarks

Date

Price

Remarks

CHECK OR FILL-IN PERTINENT INFORMATION ABOUT PRICES REPO RTED ABOVE
A.

C la ss of seller

B.

Prices are:

________ _______ ________________________to C lass of customer __________________________________
(m lr., Importer, e tc .)

Actual transaction prices’ Q

L ist prices less discounts Q

; Other Q

;

(w h o lesa ler, u se r, e tc .)

L ist prices subject to discounts

Q

;

(s p e c ify )_____ ______________________________________________________

C.

Unit quoted_____________________________________________ D. Size of order____________________________________________

E.

Shipping terms (f. o. b .; frt. allowed; ect.)

F.

Type of package used;

Crate Q

;

Carton Q

G. Is refund allowed for returnable container?

;

Y es H ]

Bag. Q
No Q

;

O th er_______________________ ._______________
If “ Y e s ” , e x p la in __________________________

A ctual s e llin g p r ic e s to c la s s o f custom er, for s iz e o f order, shipping term s, and d isc o u n ts reported.




BLS HANDBOOK OF METHODS

108

4.

ENTER DISCOUNTS AND ALLOWANCES A P P L IC A B L E TO R EPO RTED PRICES WHEN COMMODITY DESCRIBED
IS SOLD TO THE CLASS OF CUSTOMER SPECIFIED IN 3A .
A.

Trade discount ________________________ %

B.

Quantity discount (based on size of order specified in 3D) __________________

%

FOR THE FOLLOW ING, INDICATE DISCOUNT TERMS AND ESTIMATE THE P E R C E N T OF SALES
AFFECTED
Estimated
% of sale s

Terms

C.

Cash d is c o u n t ____________________________________________________________________________________

--------- ------------------------------------------------------------------------------------------------------- --------------------------%
D.

Seasonal discounts

%
E.

Cumulative volume discount

%
F.

Rebates (monthly)

%
G.

Other discounts, allow ances, free d eals, etc. (explain fully)

CIRCLE A L L DISCOUNTS OR ALLOWANCES ABOVE WHICH
HAVE BEEN DEDU CTED IN ARRIVING A T PRICES R EPO RTED
5.

LIST DUTIES OR EXCISE T A XE S A P P L IC A B L E TO R EPO RTED PRICES.

A.
B.

6.

These are included in prices quoted

Not included Q ]

If tax is included, give example of how to calculate price excluding tax.

ENTER APPROXIM ATE P ER CENTAGE OF SALES TO EACH CLASS OF CUSTOMER.

O th e r m fr.
( O . E .M . or

D ist r ib u t o r

Jo b b e r

W h o le s a le r

S a la s (a p p ro x . % )

U ser

O th e r

T o ta l

100%

Remarks:

 BLS Representative


R e t a ile r

( s p e c if y )

a s s e m b le r )

Date

WHOLESALE PRICES
U S. DEPARTMENT OF LABOR
B ureau

of

La b o r

WAS HIN GT ON. O C

S tatistics
202 12

INFORMATION FOR THE WHOLESALE PRICE INDEX
ALL REPORTS WILL BE HELD IN CONFIDENCE

Dear Sir:
The price data which you provide is used in computing the Wholesale
Price Index which is the officially accepted indicator of primary market price
movements. The index is widely used by industry and government.
These voluntary reports, submitted by you and other businessmen, are
the major source of information used in preparing this index. The information
you provide is strictly confidential and open to inspection only to sworn
employees of the Bureau of Labor Statistics.
Please use the enclosed envelope, which requires no postage, for
returning this schedule. Your continued cooperation is greatly appreciated.
COMMISSION EH OF LABOR STATISTICS

IMPORTANT INSTRUCTIONS
In the boxes provided on the other side, pleose be sure to indicote oil chonges in
COMMODITY D E S C R IP T IO N , BASIS OF QUOTATION, DISCOUNTS, ALLO W AN CES. AND T A X E S
that m ay hove occurred since your last report.
Your cooperation in keep ing oil inform ation current
is a g reat ard in com puting a r e lia b le , a c cu ra te W h o le sa le Price In d e x.

(Remarks)




109

110

BLS HANDBOOK OF METHODS

INFORMATION FOR THE WHOLESALE PRICE INDEX
C o da No
COM MODITY DESCRIPTION (Pleoia md.cote a l l changes ]•

CHANGES
C iv e dote, notwre. ond
estim ated vo I im of chonge

2

RASIS OF Q U O TA TIO N (P h o t * *nd*cote a l l chonges >

Dot* tn d noKirt of (K on g *

U ni*

C ln i of Mlhr and c w iH w
S>ie af order
Shipping termy

Omar (Jpa«V )

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

In d ica ta a ll d is c o u n ts , a llo w o n c a s , and tan as a p p lic a b le to
obova b a s is of q u o tatio n . T h is inform ation is naadad 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
i n d i c a t e a l l chonge S.)nmffffwn^^wwm^“ m ,“ , *mw"*^"mn,“'w“ w«““™ * “ ,™ r™ ,“ “ n,*» ™ "^ m *"r™ “ w“ ™^mmmnmmi^m.
“
m

DoN and notwre of (Kongo

HO

OventPy d'KOun*
TreRe diKaunl

f

Ce«*> dnceuni

\ dixauntt boon deducted

Saesanef d>tceuni

l

Hero #ny indicated
tram IK rapartad prue*
#

O W . t 'K a v r l

1

Oma* (Karpai

Hare any at mate
| bean included*

(■I'M (*RR«

4

PRICE IN FO R M A TIO N

For the com m odify described in item I , pleo te enter below the current price for the dote in d icate d , on the b o u t quoted in
•tern 2 .
Price a* of Mar

O M IC IH e
O tT I

Apr

13

1971

Moy

11

1971

June

15

1971

e m c t

O t T I

o r

C H I N K

9 17
91

O ffice

M IC IN I

(H any)

OAT|

Oct

12

9

o a r s

O ffice

o p

C H I N K

ft# a n y )

1971

Nov

* m c «

1971

Dec

14

1971

Ju ly

13

1971

Jon

11

1972

Aug

10

1971

Feb

15

1972

Mar

14

1972

Sept

14

1971

r

n

P E R M A N EN T O F F IC E

R EC O R D

KIN O LY RETURN

L



THIS FORM PROM PTLY

WHOLESALE PRICES

Technical References
Number

1. Cunningham, Francis S., “ The Use of Price Indexes in Escalator Contracts,” Monthly La­
bor Review, August 1963, pp. 948-952 Reprint No. 2424.
A statement of the use of the Wholesale and Consumer Price Indexes in escalating pur­
chase 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 Studies
in Income and Wealth, New York, National Bureau of Economic Research, (1955) Vol. 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 Economics and Statistics, May 1952, pp. 97 -142.
A description of scope, uses, and method of the U.S. Government’s interindustry sta­
tistical study of 1947. Includes discussions of computational problems, areas of use, data
requirements, etc.
4. National Bureau of Economic Research. The Price Statistics of the Federal Government:
Review, Appraisal, and Recommendations, Washington, D.C., NBER General Series, Num­
ber 73 (1961).
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,” Monthly Labor
Review, February 1962, pp. 175-182.
History of weight changes and weighting concepts, from inception of the Wholesale
Price Index.
6. U.S. Congress, Joint Economic Committee. Government Price Statistics: Hearings: Subcom­
mittee on Economic 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
NBER (q.v.) ; Part II contains statementer of Labor Statistics.
the response of Ewan Clague, Commissions of private and government economists including
7. U.S. Department of Labor, Wholesale Prices, 1890 to 1899 (Bulletin 27, 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. U.S. Department of Labor, Course of Wholesale Prices, 1890 to 1901 (Bulletin 39, March 1902).
Describes United States Senate Finance Committee index (pp. 205-211), and Depart­
ment of Labor index (pp. 212-243).
9. U.S. Department of Labor, Bureau of Labor Statistics, Wholesale Prices and Price Indexes,
195U-56 (Bulletin 1214, 1957).
Method of Calculating Special Group Indexes, (pp. 1 2 -1 3 ), Calculating Relative Im­
portance 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. U.S. Department of Labor, Seasonal Adjustment Factors; Wholesale Price Index: Selected
Series 1948-1961 (BLS Bulletin 1379, 1963).
Seasonal adjustment factors for 183 commodities and commodity groups, and descrip­
tion of BLS seasonal adjustment method.
11. U.S. Department of Labor, Wholesale Prices and Price Indexes, 1957, BLS Bulletin 1235
(1958).
Indexes by Durability of Product (Economic Sectors by Durability of Product), pp.
11-14.
12. U.S. Department of Labor, Bureau of Labor Statistics, Wholesale Prices and Price Indexes,
1958, BLS Bulletin 1257 (1959).
Describes Supplementary Inquiry on Wholesale Price Reports (discount study), pp. 1012, and January 1958 Revision of the Weighting Structure, pp. 14-16.
13. U.S. Department of Labor, Wholesale Prices and Price Indexes, 1961, (BLS Bulletin 1382,
1964).




111

112

BLS HANDBOOK OF METHODS

Technical References— Continued
Number

14.

15.

16.

17.

18.

19.

January 1961 Revision of the Weighting Structure, pp. 14-16.
U.S. Department of Labor, Bureau of Labor Statistics, Wholesale Prices and Price Indexes,
January 1967 (final) and February 1967 ( final).
Describes introduction of new 1963 weight values and major reclassification effected in
January 1967.
U.S. Department of Labor, Wholesale Prices and Price Indexes, January 1970.
Describes derivation and use of relative importances (weights) and lists all W P I weights
for December 1969.
U.S. Department of Labor, Wholesale Prices and Price Indexes, January 1971.
Introduces the new standard reference base, 1967=100, and describes conversion from
the former base.
U.S. Senate, Wholesale Prices, Wages, and Transportation, 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.
Searle, Allan D., “ Toward Comprehensive Measurement of Prices,” Monthly Labor Review,
March 1971, pp. 9-22.
Describes how a general price index could be constructed, what it should accomplish,
and virtues and limitations of various approaches.
Clorety, Joseph A ., Jr., “ Measuring Changes in Industrial Prices,” Monthly Labor Review,
November 1970, pp. 30-36.
Relates Stigler-Kindahl study (N B E R , 1970) to BLS program.




— T h o m a s R . T ib b e t t s

Chapter 12. Industry-Sector Indexes
Background

During recent years a growing need for com­
prehensive 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 es­
sentially 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 commodity rather than
industry groupings. The number of sectors or
industries for which these series are compiled
depends largely upon resources available for
additional pricing.
A set of industry-sector price indexes cover­
ing the years 1947 through 1953 was prepared
in the early 1950’s as part of the Bureau’s
project on interindustry economics. These in­
dexes, generally a regrouping of the Wholesale
Price Indexes into the interindustry (inputoutput) classification structure, were designed
for revaluing bills of goods and industry out­
puts. In 1959, another set of such indexes was
compiled for the Bureau of the Census in con­
nection with that agency’s construction of the
1958 production index benchmark.1 This sec­
ond group of price indexes was used in deflat­
1 None of these earlier indexes was published, but they
were made available to other government research agencies.
See Chapter 31, 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 Office of
Management and Budget in its Standard Industrial Classifica­
tion (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 classi­
fied. 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.
3 Government Price Statistics, Hearings before the Subcom­
mittee on Economic Statistics, of the Joint Economic Com­
mittee, Congress of the United States, Part 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.




ing values of shipments in those census product
classes where physical production data were
lacking or unsatisfactory. Again, these were
essentially indexes of commodity prices, classi­
fied as primary to a given industry.1
2
The need for the Bureau to develop IndustrySector Price Indexes became increasingly ap­
parent 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 ob­
jectives of an industrial price program should
be comprehensiveness, maximum detail in re­
porting, and groupings most useful in economic
analysis.3 The committee stated “ . . . It seems
desirable that the subclassification should aim
at fitting into the Standard Industrial Classi­
fication.”
In 1962, the Bureau of Labor Statistics intiated the development of industry and sector
price indexes. Because of its scope, the pro­
gram was viewed as a long-run program to be
accomplished in several stages. The first stage
was devoted to the study of conceptual and data
problems with only a gradual expansion of
commodity pricing.
The first indexes to be developed were out­
put price indexes utilizing gross shipments
weights. Priority was given to indexes for the
manufacturing and mining divisions of the
Standard Industrial Classification. The next
stage will be the development of output price
indexes for the trade and transportation and
other nonmanufacturing sectors. Input price
indexes, i.e., indexes representing the price of
industrial purchases, will come later. Eventu­
ally the work will expand sufficiently to permit
the development of a general price index of
the entire economy.

Description o f the Survey

Concepts
An industry or sector price index is a com­
posite index derived from several series of
prices that closely match the economic activity
of a specified industry or industry sector. These
118

114

BLS HANDBOOK OF METHODS

indexes may be either output or input price
indexes based upon either the products and
services sold or the products and services pur­
chased by an industry. An output price index
for a given industry represents price indexes
for a sample of the products produced by that
industry, averaged together according to the
relative importance of production of each sam­
ple product to the industry. An input price
index for an industry consists of an aggre­
gation of price indexes for a sample of the
commodities and services purchased by the
industry, averaged together according to the
relative magnitude of the purchases.
The Bureau’s work has been directed first
toward two sets of output indexes. One set is
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 shipments classified on
the industrial basis but weighted by shipments
of the product produced anywhere in the eco­
nomy. A principal use of the second set is for
input-output analysis.4
Universe

Ultimately, the scope of the universe will be
defined in terms of the Office of Management
and Budget’s Standard Industrial Classification
(SIC) system as revised in 1967, which covers
all domestic economic activity. This system
groups together related products or services
and assigns them industry and sector codes.
Currently, however, the scope of pricing is
effectively restricted to the commodities’ cover­
age of the WPI because of the use of WPI price
data.
If price indexes are to parallel industry out­
put data, the indexes should cover the total
output of each industry including the value of *
8
* See Chapter 31, 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 Statistics, May 1952, pp. 97-142; and “Input-Output Rela­
tions and Appraisal,” Studies in Income and Wealth, vol. 18,
National Bureau of Economic Research, New York, 1955.
BThe SIC provides no product codes.
8 The Input-Output model referred to is that compiled by
the Office of Business Economics, U.S. Department of Com­
merce. See National Economics Division, “ Input-Output Struc­
ture of the U.S. Economy: 1963,” Survey of Current Business,
November 1969, pp. 16-47.




interplant transfers, the value of sales to all
classes of customers, and the value of indus­
trial services. They should include the value of
sales for export but exclude excise taxes and
costs of transporting finished goods to pur­
chasers. This is consistent with the “ total
activity” coverage of statistical series on em­
ployment and production.
Input price indexes of materials consumed
in production should cover total materials in­
puts of the industry. This would include im­
ports for consumption, and also transportation
and delivery costs.
Prices and Base Period

The prices used in the current IndustrySector Indexes are in general those used in the
Wholesale Price Index. In the Wholesale Price
Index, primary market prices, f.o.b. production
point, are used. For the Industry-Sector In­
dexes, pricing eventually should be extended to
all classes of customers, including retail, for
use in the output indexes. Buyers’ prices, in­
cluding shipping costs, should be used for input
indexes and should represent the particular
mix of products purchased by the buying in­
dustry.
As of January 1971, the reference base
period for Federal Government indexes is
1967=100.
Classification

The Standard Industrial Classification (SIC)
of the Office of Management and Budget pro­
vides the framework for the Industry-Sector
Index classification scheme. Within this frame­
work, individual products 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
Sampling

Currently the Industry-Sector program
largely depends on price data already available

INDUSTRY-SECTOR INDEXES

in the Wholesale Price Index. However, ex­
tension of industrial pricing is being geared
to the new program. A sampling plan has been
prepared which outlines industrial sectors that
should have priority as new pricing is under­
taken. As the result of this analysis, pricing
may be cut back in some sectors to permit ex­
tension of pricing into inadequately covered
sectors.
Price data used in computing an industryoutput 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 secondary industries may be
different, sometimes because these industries
sell to different types of users. As a rough
guide to the adequacy of sampling, the im­
mediate objective is to represent at least 50
percent by value of the commodities included
in each 5-digit Census product class.7 This per­
8
centage 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 ex­
perience makes possible the use of more sophis­
ticated sampling approaches.
Weights

Since January 1967, weights for the output
indexes are 1963 value of shipments obtained
from the Census of Manufactures, the Census
of Mineral Industries, and data of the U.S.
Department of Agriculture. Indexes for 1957
through 1966 are weighted by 1958 values.
Values include interplant transfer values,
values for goods produced and consumed in the
same establishment, and the value of goods
sold for export. Values of imported commodi­
ties 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



115

weight plus the weights of other products not
directly priced in the index but whose prices
are known or assumed to move similarly.
Values for unpriced products which cannot be
assigned to a specific priced commodity are
imputed to the average movement of the prod­
uct classes in which they fall.7
For use in deflating industry shipments, the
4-digit (SIC) Industry Indexes are derived
from 5-digit product class indexes weighted
together by their shipments value for the par­
ticular industry, i.e., the “ made-in-the-industry” value.
Formula and Calculation

A modification of the Laspeyres fixed-weight
formula is used. The underlying formula is:
h = 2 ( P t/ P „ Q a)/zPoQa where P«/P„ is an indi­
vidual product price index. P „Q a, the value
weight (base year price times weight year
quantity), currently is the 1963 value of ship­
ments adjusted for price change from the year
1963 to 1967. In succeeding years, new weights
will be introduced whenever weights are re­
vised for the comprehensive Wholesale Price
Index.8
In actual practice the calculation may be
somewhat more involved than indicated by the
simple formula above. For example, indexes
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 available only at the 5-digit (productclass) level, it is necesary first to construct
product-class indexes based upon total output
(wherever made) weights. Then, using 5-digit
made-in-the-industry weights, the productclass indexes are combined to the 4-digit indus­
try level.
7 This procedure is the same as that employed in the WPI.
However, as the product classes are defined differently, an
unpriced commodity may have a different price movement
imputed to it in the Industry Index program than it has in
the WPI.
8 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.

116

BLS HANDBOOK OF METHODS

Analysis and Presentation

The published indexes for selected 5-digit
product classes and 4-digit industries are an­
nual averages for the period 1957 through
1964, and cover 44 manufacturing and eight
mineral industries. Monthly indexes are avail­
able beginning January 1965 for the same
limited number of industries and products.
Additional indexes are published as they be­
come available. By January 1971, indexes were
published for 102 four-digit industries and
about 350 five-digit product classes.
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 measures of output in constant dollars. Most
measures of output and productivity rely pri­
marily 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 suitable 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 unitvalue 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. If the dollar values themseves are divided by the price index, the result­
ing dollar values express the sales value in
terms of purchasing 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 index.9



The output indexes also may be used for
comparing movements of prices with other in­
dustry-based statistical measures such as em­
ployment, earnings, productivity, etc. Price
indexes consistent with total shipments weights
will be useful for deflating industry inputs. For
example, the appropriate index for deflating
the value of aluminum purchased by an indus­
try would be the index whose components rep­
resent 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 stud­
ies. They should be consistent in coverage with
BLS series on average hourly earnings, another
important element of cost. For contract escala­
tion, 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
output 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 contructing 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 market level of sales repre­
sented by the Census data. To include inter­
plant transfer values and values of goods
produced and consumed in the same industry, it
is necessary to assume that price movements
of goods in commercial markets represent the
price changes of goods not sold in commercial
markets.
Until additional pricing can be done, these
new indexes will be limited by the coverage—
commodity and class of customer— of the com­
prehensive Wholesale Price Index.
0 It can be shown that division of the value index by the
Laspeyres (base-year-weighted) price index yields a produc­
tion index of the Paasche (current-year-weight) form. Di­
vision by the Paasche price index, conversely, yields a
quantity index of the Laspeyres type. See Chapter 2S, Output
per Manhour Measures: Industries.

117

INDUSTRY-SECTOR INDEXES

Technical References
Number

1. Evans, W . Duane and Hoffenberg, Marvin, “ The Inter-Industry Relations Study for 1947,”
The Review of Economics and Statistics, 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 Vaccara, Beatrice N., “ The Inter-Industry
Structure of the United States,” Survey o f Current Business, November 1964, pp. 10-29.
A report showing preliminary results of the 1958 interindustry relations study and con­
taining tables of the percent distribution of 1958 gross output in 86-industry detail.
3. Moss, Bennett R., “ Industry and Sector Price Indexes,” Monthly Labor Review, August 1965,
pp. 974-982.
Contains price indexes for about 50 4-digit (Standard Industrial Classification) indus­
tries, together with a technical note on concepts, methodology, and uses.
4. National Bureau of Economic Research, “ Input-Output Relations and Appraisal,” Studies in
Income and Wealth, vol. 18, New York, 1955.
A description of scope, uses, and significance in economic analysis of the U.S. Govern­
ment’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 Price Statistics, Hearings before the
Subcommittee on Economic Statistics, Part I and Part II, Washington, D.C., 1961.
Part I represents the report, The Price Statistics of the Federal Government, prepared
by the NBER (q .v .); Part II contains statements of private and government economists in­
cluding the response by Ewan Clague, Commissioner of Labor Statistics.




Mary E. Lawrence




Chapter 13. Spot Market Prices
Background

As early as January 1934, at the request of
the U.S. Department of the Treasury, the Bu­
reau of Labor Statistics began the computation
of a daily commodity price index, using quota­
tions for sensitive comodities. It was released
first to the general public in January 1940. In
1952, in connection 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=100 base to cor­
respond to the base period adopted for other
Federal Government general purpose indexes.
In January 1971, the index was rebased again
in accordance with government-wide practice,
this time to a 1967=100 base. In 1969, com­
putation of the index on a daily basis was dis­
continued. Since then the index has been pre­
pared for Tuesday of each week.

Description o f Survey

The Spot Market Price Index is a measure
of price movements of 22 sensitive basic com­
modities whose markets are presumed to be
among the first to be influenced by changes in
economic conditions. As such, it serves as one
early indicator of impending changes in busi­
ness activity.
The commodities used are in most cases
either raw materials or products close to the
initial production stage which, as a result of
daily trading in fairly large volume of stan­
dardized qualities, are particularly sensitive to
factors affecting current and future economic
forces and conditions. Highly fabricated com­
modities 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 nominal
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 publica­
tions 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
subdivisions: Raw 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 commodities do not fall into one of these
four groupings. For example, sugar is not in­
cluded 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 Govern­
1 Exchanges which issue spot prices have committees to
make a determination of the spot price for the standard
commodity.

119

120

BLS HANDBOOK OF METHODS

ment agencies. 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
comprehensive monthly Wholesale Price In­
dex.2 Prices for butter, corn, hides, hogs, lard,
tallow, and tin are either differently specified
spot prices or from different markets.
Selection o f Products

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 homo­
geneous 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 pos­
sible. Also, the influence of international mark­
ets upon the economy was taken into account by
the inclusion of some key commodities (such as
crude rubber and tin) which are important in
international trade. Both in the sample and in
the index structure, an attempt was made to
prevent price movements of agricultural prod­
ucts from dominating the movement of the
index.

each of the commodities is unweighted in the
index means that a price change for rosin, a
comparatively 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 obtain­
ing 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. This product is then multiplied
by 100 to obtain the index number for each
period. The calculation is made by means of
logarithms. The formula reduces to
Log

Log Pk—2 log P0+44

22
where
7fc = Index for a given day
P k = Price for a given day
P0 = Average (geometric) price in base
period
44 = Logarithmic constant which when di­
vided by 22 equals log of 100.
Monthly average indexes are obtained ac­
cording to the previous procedure, except that
P7=the geometric average of the Tuesday
c
prices (daily prices prior to 1969) over the
month. In maintaining the index over time, it
may be necessary to change commodity specifi­
cations or substitute entirely new products.
These changes are handled by a statistical link­
ing procedure so that only actual price move­
ments are reflected in the index.

Estimating Procedures

The Spot Market Index is an unweighted
geometric mean 3 of the individual commodity
4
*
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 move­
ments of individual commodities. Since ex­
tremely 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
2 See Chapter 11, Wholesale Prices.
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 = 2 7 . 3V27—3). The arithmetic mean,
is 4.2.




Analysis and Presentation

Tuesday spot market indexes and prices are
published each week, on the Friday following
the day of reference. A summary of weekly
indexes and the average for each month are
published with the first weekly release of the
following month. Beginning with 1950, his­
torical indexes are shown for Tuesday of each
week together with monthly averages; from
July 1946 through 1949 indexes are listed for
Tuesday of each week only. In addition, in­
dexes are published for selected earlier dates:
August 15, 1939, December 6, 1941, August 17,
1945, and June 28, 1946.

121

SPOT MARKET PRICES

Uses and Limitations

A survey of users in 1964 showed that the
Index is frequently used as a general economic
indicator, for gaging the direction of basic
prices, for forecasting general price move­
ments, and for current prices of specific com­
modities. Other uses, frequently mentioned, are
for market research and for comparing price
trends with the user’s selling or buying prices.
The Tuesday Index of Spot Market Prices dif­
fers from the Wholesale Price Index in method
of construction and weighting, as well as in the
sample of items for which prices are included.
While it is independent of the monthly compre­
hensive index, changes in the Tuesday Index
or its components may foreshadow turns in
Wholesale Price Indexes. However, the Tues­
day 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 Tuesday Spot
Market 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 committee of experts from the com­
modity exchange for a standardized com­
modity. Also, when there are not enough trans­
actions 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 vol­
ume of private transactions occurring outside
the organized market. However, it is believed
that the reported exchange prices generally are
used as the basis for private negotiations.

Composition o f Grouping Indexes

Metals: Copper scrap, lead scrap, steel scrap,
tin, and zinc.



Textiles and Fibers: Burlap, cotton, print
cloth, and wool tops.
Livestock and Products: Hides, hogs, lard,
steers, and tallow.
Fats and Oils: Butter, cottonseed oil, lard,
and tallow.
Specifications for Commodities Included in the Index as of
March 1971
Market
Specifications
10 oz„ 40", ex-dock or ex­ New York.
warehouse, duty paid, per
yd.
Chicago.
Grade A, 92 score, per lb. —
Butter
New York.
_
Cocoa beans__ Accra, per lb.
No. 1 heavy copper and wire, New York.
Copper scrap
refiners’ buying price, car­
load lots, delivered buyers’
works, per lb.
Chicago.
Corn
_______ No. 3 yellow, per bu. - ..
12 markets.
Middling, ll£ f " , per lb.
Cotton
Memphis.
Crude, Valley, per lb. ___ Cottonseed
oil.
Cow, light native, packer 30/53 Chicago.
Hides
lbs., fleshed, packer to tan­
ner, dealer, or exporter per
lb., f.o.b. shipping point.
U.S. No. 2’s and 3’s, 200-220 Omaha.
Hogs
lbs., per 100 lb.
Prime Steam, in tanks, per Chicago.
Lard
lb.
Battery plates, smelters’ buy­ New York.
Lead scrap
ing price, East, carload lots;
delivered buyers’ works, per
lb.
New York.
Print cloth __ 48", 78x78 count, 4 yds./lb.
spot and nearby, per yd.
Print cloth __ 48", 78x78 count, 4 yds./lb., New York.
most distant contract, per
yd.
Gum, W G grade, carlots, per New York.
Rosin
100 lb.
Crude, natural, No. 1 Ribbed New York.
Rubber
Smoked Sheets, per lb.
Steel scrap __ No. 1 heavy melting, (dealer), Chicago.
consumers’ buying price, in­
cluding
brokerage,
de­
livered, per gross ton.
Steers ________ Choice, 900-1100 lbs., per 100 Omaha.
lb.
Raw, 96°, duty paid, per 100 New York.
Sugar ______
lb.
Tallow _______ Fancy, bleachable, inedible, Chicago.
per lb. _
Tin __________ Grade A, spot delivery, per New York.
lb.
No. 1 Dark Northern Spring, Minneapolis.
Wheat ___
per bu.
Wheat _______ No. 1 Hard Winter Ord., per Kansas City.
bu.
Certificated spot price, nom­ New York.
Wool tops
inal, per lb.
Prime
Western,
for New York.
Zinc _________ Slab,
prompt delivery, delivered,
(f.o.b. New York equivaCommodity

Burlap -

L

loyd

E.

W

ig r e n




Wag es and Industrial Relations

Chapter 14. Occupational Pay and Supplementary Benefits
Background

The Bureau of Labor Statistics, for many
decades, has conducted studies of wages by oc­
cupation and industry, based upon employer
records. The Bureau’s first such study, growing
out of a study by the U.S. Senate in 1891, re­
sulted in a wage rate record extending back
continuously to 1860. Systematic collection of
wage data by occupation and industry has con­
tinued since the turn of the century; changes
in coverage has been dictated mainly by gov­
ernment requirements. A large survey program
undertaken for the War Industries Board in
1919 produced occupational pay rates by indus­
try 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 Re­
covery Act, Public Contracts Act, and the Fair
Labor Standards Act, with emphasis on nation­
wide data for relatively low-wage industries.
Survey activity shifted in the 1940-41 de­
fense period to heavy industries essential to
war production. Implementation of wage sta­
bilization policy during the war required a
large-scale program of occupational wage stud­
ies 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.
Area wage surveys, initiated in the late
1940’s, were designed to meet the growing de­
mand for pay data related to office clerical and
manual jobs that are common to a wide variety
of manufacturing and nonmanufacturing in­
dustries within metropolitan areas. This survey
program was firmly established and tem­
porarily expanded for use in the wage stabili­
zation 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 re­
sulted in a survey design that would produce
national averages, based on an area sample.
Data for individual areas studied also serve the
wage administration needs for other govern­
ment agencies.
Prior to 1960, studies in a very few profes­
sions provided salary data. Beginning in that
year, salary surveys have been made on a
nationwide basis covering professional, admin­
istrative, and technical jobs in a broad spec­
trum of industries. Averages for these jobs,
together with national averages for clerical and
drafting jobs included in the area wage sur­
veys, are utilised by the administrative agen­
cies directly concerned with Federal pay
matters.

Description o f Surveys

Although differing in industrial, geographic,
and occupational coverage, the three types of
surveys described form an integrated program
of occupational wage surveys based upon a
common set of administrative forms, manual of
procedures, and common concepts and defini­
tions. Employer cooperation in surveys is on a
voluntary basis. Confidential individual estab­
lishment data compiled by the Bureau’s field
economists are grouped in published reports in
a manner that will avoid possible disclosure of
an establishment’s rates. Establishments in­
cluded in all surveys are classified by industry
as defined in the 1967 edition of the Standard
Industrial Classification Manual prepared by
the U.S. Office of Management and Budget.1
Survey reports identify the minimum size of
establishment (measured by total employment)
studied. Definitions for Standard Metropolitan
Statistical Areas are employed in all pro­
grams.2
1 See appendix B.
a See appendix C.

123

124

BLS HANDBOOK OF METHODS

Industry wage surveys provide data for occu­
pations selected to provide representativeness
of the range of rates, methods of wage
payment, and of men’s and women’s work ac­
tivities. Consideration also is given, in their
selection, to the prevalence in the industry,
definiteness and clarity of duties, and impor­
tance as reference points in collective bargain­
ing.
In addition to collecting straight-time firstshift rates (or hours and earnings for incentive
workers) for individual workers in the selected
occupations, surveys in most industries also
establish the wage frequency distribution for
broad employment groups, i.e., production and
related workers or nonsupervisory workers.
Weekly work schedules; shift operations and
differentials; paid holiday and vacation prac­
tices; and health, insurance, and pension bene­
fits are included in the information 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 un­
der incentive pay plans, and the extent to which
establishments provide a single rate or range
of rates for individual job categories.
Fifty manufacturing and 20 nonmanufactur­
ing industries, accounting for about 22.5 mil­
lion employees, are surveyed on a regularly
recurring basis. A majority are studied on a
5-year cycle, but a number of comparatively
low-wage industries are on a 3-year cycle. In
addition, special wage surveys also are under­
taken at the request of other government
agencies.
Nearly all of the manufacturing, utilities,
and mining industries are studied on a nation­
wide basis and estimates are provided also for
regions and major areas of concentration. Sur­
veys in trade, finance, and service industries
usually are limited to a score or more of metro­
politan areas. Nationwide surveys generally
develop separate estimates by size of establish­
ment, size of community, labor-management
agreement coverage, and type of product or
plant group.
Area wage surveys provide data for occupa­
tions common to a wide variety of industries in
the communities surveyed. The 76 occupational



categories studied include 31 office clerical; 15
electronic data processing, drafting, and indus­
trial nurses; and 30 maintenance, toolroom,
powerplant, and custodial and material move­
ment jobs. Thus, they provide representation
of the range of duties and responsibilities as­
sociated with white-collar, skilled maintenance
trades, and other “ indirect” manual jobs.
Weekly salaries reported for individuals in
white-collar 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) manu­
facturing; (2) transportation, communication,
and other public utilities; (3) wholesale trade;
(4) retail trade; (5) finance, insurance, and
real estate; and (6) selected service industries.
Establishments employing fewer than 50 work­
ers are excluded— 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,
separate data are provided for manufacturing
and nonmanufacturing in each area and,
wherever possible, for individual industry di­
visions in the nonmanufacturing sector. Among
the 89 Standard Metropolitan Statistical Areas
in this annual survey program as of 1971, sep­
arate data are provided for transportation,
communication, and other public utilities in 88
areas; for retail trade in 25 areas; for whole­
sale trade and finance, insurance, and real
estate in 17 areas; and for the selected service
industries in 8 large areas. In 22 of the larger
areas, wage data are presented separately for
establishments that have 500 workers or more.
Data on weekly work schedules; paid holiday
and vacation practices; and health, insurance,
and pension benefits are recorded separately
for nonsupervisory officeworkers and plant
workers (nonoffice). Shift operations and dif­
ferentials are collected for plantworkers in
manufacturing. Data on minimum entrance
rates for inexperienced officeworkers are col­
lected in all industries. These items are studied
biennially in all areas. This survey program

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS

also has developed information on profit-shar­
ing plans, characteristics of sick leave plans,
wage payment systems, and other items related
to employee compensation.
Special area wage surveys have been con­
ducted annually since 1967 at the request of the
Employment Standards Administration for use
in administering the Service Contract Act of
1965. The surveys also meet the needs of the
general public and provide information on
hourly earnings for 14 office occupations; 22
maintenance, toolroom, powerplant, and cus­
todial and material movement jobs; 7 laundry
jobs; and 6 food service jobs.
The industrial scope includes manufactur­
ing; transportation, communication, and other
public utilities; and wholesale trade industry
divisions; and general merchandise stores;
eating and drinking places; real estate, hotels
and other lodging places; engineering and
architectural services; personal services; and
miscellaneous business services. Data on inci­
dence of paid holidays and vacation practices,
and health insurance and pension benefits are
provided biennially.
The National Survey of Professional, Administative, Technical, and Clerical Pay provides
a fund of broadly based information on salary
levels and distributions in private employment.
The 80 occupation-work levels studied were
selected from the following fields: Accounting,
legal services, personnel management, engi­
neering and chemistry, buying, clerical super­
visory, drafting, and clerical. Definitions for
these occupations provide for classification of
employees according to appropriate work levels
(or classes). Although reflecting duties and
responsibilities in industry, the definitions were
designed to be translatable to specific pay
grades in the General Schedule applying to
Federal Classification Act employees. This sur­
vey, thus, provides information in a form suit­
able for use in comparing the compensation of
salaried employees in the Federal civil service
with pay in private industry.
Average salaries, monthly and annual for
all occupations and also on a weekly basis for
clerical and drafting, relate to the standard
salaries that were paid for standard work
schedules, i.e., to the straight-time salary cor­




125

responding to the employee’s normal work
schedule, excluding overtime hours. Nationwide
salary distributions and averages are presented
for men and women combined. Averages also
are presented for establishments in metropoli­
tan areas combined and for establishments
employing 2,500 workers or more.
Industry divisions included are: (1) manu­
facturing, (2) transportation, communication,
electric, gas and sanitary services, (3) whole­
sale trade, (4) retail trade, (5) finance, insur­
ance, and real estate, and (6) engineering
and architectural services, and commercially
operated research, development, 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 non­
metropolitan counties. The minimum establish­
ment size included in the survey was raised
from 100 to 250 in 1961. In 1966, the minimum
establishment size was lowered to 100 in trans­
portation, communication, and the other public
utilities; wholesale trade; and the service in­
dustries studied. The minimum was also low­
ered to include establishments with 50 workers
in the finance, insurance, and real estate in­
dustries. 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
surveys summarize a highly specific wage
measure— the rate of pay, excluding premium
pay for overtime and for work on weekends,
holidays, and late shifts, for individual work­
ers. In the case of workers paid under piece­
work or other types of production incentive
pay plans, an earned rate is computed by di­
viding straight-time earnings for a time period
by corresponding hours worked. Production
bonuses, commissions, and cost-of-living bo­
nuses are counted as earnings. In general,
bonuses that depend on factors other than the
output of the individual worker or group of
workers are excluded; examples of such non­
production payments are safety, attendance,
3 The terms “in scope" or “within scope” are used through­
out this chapter to refer to the coverage of the particular
survey being described.

126

BLS HANDBOOK OF METHODS

year-end or Christmas bonuses, and cash dis­
tributions under profitsharing plans.
Unless stated otherwise, rates do not include
tips or allowances for the value of meals, room,
uniform, etc. The earnings figures, thus, rep­
resent 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
work.
Hours shown for salaried occupations relate
to standard weekly hours for which the em­
ployee receives his regular straight-time salary.
Occupational classifications are defined in
advance of the survey. Because of the emphasis
on interestablishment and interarea compar­
ability of occupational content, the Bureau’s
job descriptions may differ significantly from
those in use in individual 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 purposes. The pri­
mary objective of the descriptions is to identify
the essential elements of skill, difficulty, and
4 An example of a job description:
Machinist, Maintenance
Produces replacement parts and new parts in making re­
pairs of metal parts of mechanical equipment operated in an
establishment. Work involves most of the following: In­
terpreting written instructions and specifications; planning
and laying out of work; using a variety of machinist’s handtools and precision measuring instruments; setting up and
operating standard machine tools; shaping of metal parts to
close tolerances; making standard shop computations relating
to dimensions of work, tooling, feeds, and speeds of machin­
ing; knowledge of the working properties of the common
metals; selecting standard materials, parts, and equipment
required for his work; and fitting and assembling parts into
mechanical equipment. In general, the machinist’s work
normally requires a rounded training in machine-shop prac­
tice usually acquired through a formal apprenticeship 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,
particularly office and skilled production-worker categories,
workers may regularly perform a combination of duties in­
volving more than one occupation. Unless indicated other­
wise 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.




responsibility that establish the basic concept
of the job.4
Although work arrangements in any one es­
tablishment may not correspond precisely to
those described, those workers meeting the
basic requirements established for the job are
included.5
In applying these job descriptions, the Bu­
reau’s field representatives exclude working
supervisors, apprentices, learners, beginners,
trainees, handicapped workers, part-time or
temporary workers, and probationary workers
unless provision for their inclusion is spe­
cifically stated in the job description.
Paid holidays, paid vacations, and health,
insurance, and pension plans are treated sta­
tistically 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 eventually 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 em­
ployer. Informal provisions are excluded.

Survey Methods

Planning. Consultations are held with appro­
priate management, labor, and Government
representatives to obtain views and recom­
mendations related to scope, timing, selection,
and definitions of survey items, and types of
tabulations. Particularly in planning surveys in
specific industries, these discussions impor­
tantly supplement comments and suggestions
received from the regional offices at the con­
clusion of the previous study. Reflecting its use
in evaluation of Federal white-collar pay, the
design of the National Survey of Professional,
Administrative, Technical, and Clerical Pay
was developed in conjunction with the Office
of Management and 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 identi­
fied in terms of the classification system pro­
vided in the Standard Industrial Classification
Manual. The scope may range from part of a

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS

4-digit code for an industry study to a uniform
combination of broad industry divisions and
specific industries for the area wage surveys or
the salary survey of professional, administra­
tive, technical, and clerical jobs. The needs of
major users are a major consideration in de­
signing the multi-purpose occupational studies.
The minimum size of establishment included
in a survey is set at a point where the possible
contribution of the excluded establishments is
regarded as negligible for most of the occupa­
tions surveyed. Another practical reason for
the adoption of size limitations is the difficulty
encountered in classifying workers in small es­
tablishments where they do not perform the
specialized duties indicated in the job defini­
tions.
Considerations in timing of industry surveys
include date of expiration of major labormanagement agreements, deferred wage ad­
justments, seasonality of production (e.g.,
garments), and interests of users. Wherever
possible, area wage surveys are timed to follow
major wage settlements as well as to meet the
needs of government 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 reasonably complete range of rates in the
wage structure for the employment categories
involved, i.e., production and related workers
in a specific manufacturing industry or nonsupervisory office, maintenance, material han­
dling, 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 (BLS
2751A) includes items relating to products or
services, employment, shift operations and dif­
ferentials, work schedule, overtime premiums,



127

paid holidays and vacations, insurance and
pension plans, union contract coverage, and
other items applicable to the establishment. The
second (BLS 2753G) is used in recording oc­
cupation, sex, method of wage payment, hours
(where needed), and pay rate or earnings for
each worker studied. Supplementary forms are
used to meet particular needs.
Collection. Bureau field economists collect data
by personal visit to each of the sample estab­
lishments. Job functions and factors in the
establishment are carefully compared with
those included in the Bureau job definitions.
The job matching may involve review of rec­
ords such as pay structure plans and organiza­
tional charts, company position descriptions,
interviews with appropriate officials, and, on
occasion, observation of jobs within plants. A
satisfactory completion of job matching per­
mits acceptance of company-prepared reports
where this procedure is preferred by the re­
spondent. Generally, however, the field eco­
nomist secures wage or salary rates (or hours
and earnings, when needed) from payroll or
other records and data on the selected employer
practices and supplementary benefits - from
company officials, company booklets, and labormanagement agreements.
Area wage surveys in all areas involve per­
sonal visits every second year with partial
collection by mail in the intervening years.
Establishments participating in the mail collec­
tion receive a transcript of the job matching
and wage data obtained a year earlier by the
field economist, together with the job defini­
tions. 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 yearearlier data.
The work of all field economists is checked
for quality of reporting, with particular atten­
tion directed to accuracy in job matching. The
revisits are made by supervisory and senior
economists. Systematic technical audits of the
validity of survey definitions, made by staff
with specialized training, also are maintained
for the technically complex nationwide whitecollar salary survey.

BLS HANDBOOK OF METHODS

128
B L S 2751A

Budget Bureau No. 44-R0338
Approval expires November 30, 1972

U.S. D E P A R T M E N T OF LA BO R
Bureau of Labor Statistics

(R e v . January 1970)

WAGE S U R V E Y
G E N E R A L E S T A B L IS H M E N T IN FO RM A TIO N
1.

Your report will be
held in confidence

E S T A B LISH M E N T ID E N T IF IC A T IO N

A . Survey

P a y ro ll P e rio d

P a y ro ll P e rio d
19___

e s t a b l i s h m e n t

STREET

19.

n a m e

ADDRESS

COUNTY, STATE

AND ZIP CODE

NAME AND T I T L E

AREA

OF AUTHORIZING OFFICIAL

NAME AND T I T L E

CODE - TELEP H O N E

OF OFFICIAL SUPPLYING DATA

1 9 ___

ADDRESS

OF

O FFICE

FROM

WHICH

DA TA

WAS O B T A I N E D .

IF D I F F E R E N T

FROM

ABOVE

T E L E P HONE

B . C entral O ffic e
NAME

OF

S TR E E T

(

C o m p le te

|

i f clea r a n c e

|

a n d / o r d a ta

|

|

o b t a i n e d fr o m

COMPANY

th is

sou rce)
NAME

ADDRESS

CITY,

S TA TE

AND

OF

AUTHORIZIN G

O FFICIA L

ZIP C O D E

T IT L E

2.

C U R R E N T PR O D U C T S OR S E R V IC E S AND P R O C E S S E S

A. P R O D U C T O R S E R VI C E

[APPROXIM ATE
% ANNUAL
VALUE

APPROXIMATE
% ANNUAL
VALUE

B. S C O P E O F O P E R A T I O N S

19____

19____

3.

P R O D U C T OR S E R V I C E

O F F IC E USE O N LY
S C H E D U L E NO.
1 -5

1 9 _____
1 9 _____




1D E N T .

6 -8

AREA

9-11

REGION

STATE

CITY
SIZE

SIC C O D E

EST.
SI Z E

UNION

12

13 -1 4

15

1 6 -1 9

20

21

•
WEIGHT

22 -2 4

SPECIAL CHAR­
ACTERISTICS
(1 )
2 5 -27

(2)
2 8 -3 0

129

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS
U .S . D E P A R T M E N T O F L A B O R

B L S 2751 A —Continued
(Rev. January 1970)

Bureau of Labor Statistics

SURVEY

P A YR O LL

PERIOD

SCHEDULE

NO.

ESTABLI SHMENT

5. UNION C O N T R A C T C O V E R A G E
No

A . Are a m ajority o f your p rod u ction w orkers c o v e r e d by union agreem ents?
B . Are a m ajority o f your o f f i c e w orkers c o v e r e d by union a g r e e m e n t s ? _____

a
a

31
32

nam e and

a ffilia tio n

□
□

D. What o c c u p a t io n a l group s are co v e r e d by the co n tra ct?

C . With what u n ion s d o e s th is esta b lis h m e n t h a ve co n tr a c ts ?
( G iv e

Y es

b e lo w .)

(L is t grou p s

b elo w

o p p o site

th e

a p p r o p ria te

u n io n .)

P ro d u ctio n W orkers:

O ffic e W orkers:

6 . E S T A B LISH M E N T E M P L O Y M E N T (A P P R O X IM A T E )
A . What is the approxim ate tota l em ploym ent* in th is e sta b lish m e n t?
B . H ow many are n on su p e rv iso ry p rod u ction (p la n t) w orkers?
M e n ___________________________________________________________________ ____
W omen____________________________________________________________________
C . N o n s u p e rv is o ry o ffi c e w o r k e r s ? _____________________________
M en _______________________________________________________________________
Women________ ____________________________________________________________
D . O ther e m p lo y e e s (e x e c u t iv e , p r o fe s s io n a l, s u p e rv is o ry , e t c . ) ? ____
E.

__________________________________________________________________________

F .

__ __________________________________________________________________

In clu d es s a la rie d o ffi c e r s o f c o r p o r a tio n s but d o e s not in clu d e p r o p rie to rs , members o f u n in co rp o ra te d firm s, p e n s io n e r s ,
members of the arm ed fo r c e s ca rried on the p a y r o ll, or unpaid fa m ily w o r k e rs .

G . Rem arks




1

130

BLS HANDBOOK OF METHODS

OCCUPATIONAL RATES

* s 275 3 0
Rev. 65

Payroii period
'

Schedule No.

r

Est. N am e
O C C U P A T IO N A N D G RAD E




P age
Occupational
code
(1)

Sex
(2)

M ethod of Num ber of
pay
workers
(3)
(4)

Hours
(5)

Salary, rate, or
earnings
(6)

Line
No.
(7)
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

(8)

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS

Sampling

The sampling design employed is almost al­
ways highly stratified. Before the sample is
selected, information on all known establish­
ments that might possibly fall within the scope
of the survey is compiled from lists provided by
regulatory governmental agencies (primarily
State unemployment insurance agencies), sup­
plemented by data from trade directories, trade
associations, labor unions, and other sources.
Establishments then are stratified as pre­
cisely as available information permits. Each
geographic-industry unit for which a separate
analysis is to be presented is sampled inde­
pendently. Within these broad groupings, a
finer stratification by product (or other perti­
nent attributes) and size of establishment is
made. Stratification may be carried still further
in certain industries: textile mills, for instance,
are classified on the basis of integration, 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 estab­
lishments through progressively declining pro­
portions of the establishments in each smaller
size group, in accordance with the principles
of optimum allocation. Thus, each sampled
stratum will be represented in the sample by a
number of establishments roughly propor­
tionate to its share of the total employment.
Though this procedure may appear at first to
yield a sample biased by the over-representa­
tion of large firms, the method of estimation
employed avoids the possibility of bias by the
assignment of proper weights to the sample
establishments.
In the event a sample establishment within
scope is unable to supply usable data, a sub­
stitute is assigned in the same industry-loca­
tion-size class. (Since no close relation exists
between failure to participate in these surveys



131

and the items being studied, little bias is intro­
duced by this procedure.) The overall non­
response 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 di­
versity of occupations, and their distribution,
the relative dispersion of earnings among es­
tablishments, the distribution of the establish­
ments by size, and the degree of accuracy
required. Estimates of variance based on data
from previous surveys are used in determining
the size of the sample needed.
As indicated earlier, area wage surveys
are limited to selected metropolitan areas.
These areas, however, form a sample of all
such areas, and, when properly combined
(weighted), 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
industrial activity as measured by percentage
of manufacturing employment, major indus­
tries, and level of earnings in manufacturing.
Each area was selected with probability pro­
portionate to its nonagricultural employment.
The largest metropolitan areas are selfrepresenting, 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 1970.
In some strata it was impossible to give some
areas their proper chance of selection, because
of difficulties in making surveys in these areas
primarily due to predominance of single em­
ployers. 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 oc­
cupation are computed as the arithmetic mean

132

BLS HANDBOOK OF METHODS

of the individual employees’ earnings. They are
not estimated by dividing total payrolls by the
total time worked, since such information al­
most never is available on an occupational
basis.
All estimates are derived from the sample
data. The averages for occupations, as well as
for industries, are weighted averages of indi­
vidual earnings and not computed on an es­
tablishment basis. The proportion of employees
affected by any fringe provision likewise is
estimated from the sample; all workers in each
establishment are considered 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 vari­
able sampling ratio in different strata of the
population would result in biased estimates if
straight addition of the data for the various
establishments were made. Therefore, each es­
tablishment 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
universe were 7 establishments, from which a
sample of 3 was selected. Assume that estab­
lishment 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 probability 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 calculations are
made in estimating average earnings for a
given occupation.
Workers in occupation
in sample establish­
ments at specified rate
Average
Estimates of total
hourly
in stratum
Total
earn- ______________________
Establishment Weight number
ings
Workers earnings
A
B ______ ___

2
1

C_____ ___
4
Estimated universe




40
30
20
10

$1.60
1.70
1.95
1.65

2X40
1X30
1X20
4X10
..... 170

2 X 40 X $1.60
1X 30X 1.70
1X 20X 1.95
4X 10X 1.65
$284.00

The estimated average hourly earning is thus
$284.00
or $1.67.
170
A similar method applies to any character­
istic estimated from the sample. To estimate
the proportion of employees in establishments
granting paid vacations of 2 weeks after 2
years of service, for instance, the establish­
ments are classified according to the length of
vacation granted after 2 years’ service, estab­
lishment 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 employment
then is computed. Using the same three estab­
lishments as in the previous example, this can
be illustrated as follows:
Actual total Weighted
establishment em ployment
Establishment Weight employment
200
100
A _____________
2
500
500
1
B _________
300
75
C_____________
4
1,000
Estimated universe

Vacation
provisions
after 2 years
1 week.
2 weeks.
1 week.

Thus, the estimated percentage of workers in
establishments granting 2 weeks’ vacation
500
or 50 percent.
after 2 years of service is
1,000
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 sim­
ilar units. Should any segment be affected by
a substantial amount of such noncooperation,
the publication of materials will be diminished
by omitting separate presentation 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 esti­
mated stratum totals by multiplying by the
inverse of the chance of selection. This pro­
cedure provides the ratio of nonagricultural
employment in the stratum to that in the

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS

sample area (one in the case of the large selfrepresenting areas). Summing all such esti­
mated stratum totals yields the earnings and
employment 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 available 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 com­
munity, size of establishment, product, or
labor-management agreement coverage that
generally are included in bulletin reports on
nationwide surveys. Regardless of geographic
scope, industry survey reports record the inci­
dence of incentive pay plans and, to the extent
possible, average pay levels separately for time
and incentive workers.
Individual bulletin reports on individual area
wage surveys are supplemented by two sum­
mary bulletins. The first compiles the results
of individual area surveys made during a fiscal
year. The second contains information on oc­
cupational earnings, employer practices, and
suplementary wage benefits for all metropoli­
tan areas combined and by industry division
within the four broad census regions.
Wage-rate indexes are constructed for broad
occupational categories, e.g., office clerical
workers, skilled maintenance workers, and un­
skilled plant workers. These indexes are pub­
lished annually, separately for all industries
and manufacturing, for each metropolitan area
studied, and for all metropolitan areas com­
bined and by region. Area pay relatives for the
three occupational categories are published
annually, permitting ready comparisons of
average pay levels among areas. Estimates of
labor-management agreement coverage are also
presented annually. Occupational pay relation­
ships within individual establishments are
summarized periodically.
Bulletin reports on the National Survey of
Professional, Administrative, Technical, and



133

Clerical Pay present occupational averages and
distributions on an all-industry basis, nation­
wide and separately for all metropolitan areas
combined, and for establishments employing
2,500 workers or more. Average pay levels for
industry divisions are shown as percentages of
the all-industry averages. Year-to-year per­
centage changes for occupation-work levels and
trend estimates for occupations are reported.
Industry and area 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 March.
Summaries of the data in the bulletins and
special analyses appear also in the Monthly
Labor Review.

Uses and Limitations
Occupational wage data developed in these
surveys have a variety of uses. They are used
by Federal, State, and local agencies in wage
and salary administration and in the formula­
tion of public policy on wages, as in minimum
wage legislation. They are of value to Federal
and State mediation and conciliation services
and to State unemployment compensation agen­
cies in judging the suitability of job offers.
Knowledge of levels and trends of pay rates by
occupation, industry, locality, and region is re­
quired in the analysis of current economic de­
velopments and in studies relating to wage
dispersion and differentials.
Bureau data are used in connection with
private wage or salary determinations by em­
ployers 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
questions of pay policy. As suggested earlier,
limitations 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 determina­
tion of periodicity and timing of particular

134

BLS HANDBOOK OF METHODS

surveys. Depending upon his needs, the user
may find it necessary to interpolate for occupa­
tions or areas missing from the survey on the
basis of knowledge of pay relationships.
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 relative relationships found in the
majority of establishments. To illustrate, em­
ployment in the specialized maintenance crafts
tends to be concentrated in the larger establish­
ments, whereas employment 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 establish­
ments vary from the average level, the skill
differential measure based on the survey aver­
ages will differ to some degree from that
obtainable within each of the larger estab­
lishments.
The incidence of incentive methods of pay­
ment may vary greatly among the occupations
and establishments studied. Since hourly aver­
ages for incentive workers generally exceed
those for hourly-rated workers in the same job,
averages for some incentive-paid jobs may
equal or exceed averages for jobs positioned
higher on a job evaluation basis but normally
paid on a time basis. Wherever possible, data
are shown separately for time workers and
incentive workers in the industry surveys. In­
centive plans (generally plant-wide in applica­
tion) apply to only a very small proportion
of the workers in the indirect plant jobs studied
in the area 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 ex­
pansions and reductions for other reasons, as
well as changes in the proportion 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
occupations, the dispersion of earnings among
establishments and frequency of occurrence of
the occupation 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 enumeration of all estab­
lishments in the universe. That error applies to
the smallest breakdown published. 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 practices almost always will ap­
pear in their true position. Small percentages
may be subject to considerable 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 employees may be
given as 2 percent, when the true percentage
for all establishments might be only 1 percent.
Such a sampling error, while considerable, does
not affect the essential inference that the prac­
tice 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 establishments in the proportion of
workers found in individual occupations. (It is
not unusual to find these estimates subject to
sampling error of as much as 20 percent.)
Hence, the estimated number of workers can
be interpreted only as a rough measure of the
relative importance of various occupations.
The greatest degree of accuracy in these em­
ployment counts is for those occupations found
principally in large establishments. This sam­
pling error, however, does not materially affect
the accuracy of the average earnings shown for
the occupations. The estimate of average earn­
ings is technically known as a “ ratio estimate,”
i.e., it is the ratio of total earnings (not pay­
rolls) to total employment in the occupation.
Since these two variables are highly correlated

OCCUPATIONAL PAY AND SUPPLEMENTARY BENEFITS

(i.e., the errors tend to be in the same direc­
tion), the sampling error of the estimate (aver­
age hourly earnings) is considerably smaller
than the sampling error of either total earnings
or total employment.
Since completely current and accurate infor­
mation regarding establishment products and
the creation of new establishments is not avail­
able, the universe from which the sample is
drawn may be incomplete. Sample firms in­
correctly 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 ac­
counted for.

Since some measure of subjective judgment
enters into the classification of occupations and
other characteristics, there is some reporting
variability in the results. A repetition of the
survey in any establishment with different in­
terviewers and respondents would undoubtedly
produce slightly different results. However,
when spread over a large number of establish­
ments the differences, being random, would
tend to balance out. Hence, analyses based on a
small number of respondents must be used with
care, even when all eligible establishments are
included. No evidence of any consistent error
has been uncovered.

Technical References
Number

1.

Cohen, Samuel E., “ Studies of Occupational Wages and Supplementary Benefits.” Monthly
Labor Review, March 1954 (pp. 292-297).
An eariler description of the methods of wage surveys, similar to the present article.
2. Douty, H. M., “ Survey Methods and Wage Comparisons.” Labor Law Journal, April 1964 (pp.
222-230).
A discussion of the uses of wage survey results, and the pitfalls to be avoided. A short
discussion of the factors affecting survey methods is also included.
3. Kanninen, Toivo P., “ New Dimensions in BLS Wage Survey Work.” Monthly Labor Review,October 1959 (pp. 1081-1084).
An outline of the occupational wage survey programs, as expanded in fiscal 1960. Lists
the type of survey and cycle for each of 70 industries studied separately, and identifies the
area sample as originally determined for the labor market survey program.




135

— G eo r g e L . S t e l l u t o




Chapter 15. Employee Earnings and Hours Frequency Distributions
Background

An extensive program of studies of the fre­
quency distribution of employee earnings has
been maintained by the Bureau of Labor Sta­
tistics since 1954. 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 section 4(d ) of
the Fair Labor Standards Act, directed the
Secretary of Labor to include in his annual
report an evaluation and appraisal of minimum
wages established by the act and his legislative
recommendations.1 To meet these requirements,
the program of employee earnings distributions
studies was established. Since 1964, frequency
distributions of weekly hours of work have
been a part of this program.
This program was initiated to provide knowl­
edge 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 lowest
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.

Description

Studies of employee earnings and hours dis­
tributions generally include within their scope
all nonsupervisory employees. No attempt is
made to classify employees by occupation, al­
though at times information is collected sep­
arately for some types of employees. The data
collected relate to straight-time hourly eam1 The Bureau’s studies conducted in this area before 1954
frequently related to selected occupations, or were limited to
relatively narrowly defined industries.
2 See technical references for a listing of publications con­
taining full descriptions of both types or studies.
1 See appendix B.
* See appendix C.




ings, excluding premium pay for overtime
work and for work on weekends, holidays, 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 informa­
tion is collected for hours worked during the
week and for hours spent on vacations, holi­
days, 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
conducted— industry and area.1 Industry stud­
2
ies may have broad coverage, such as manu­
facturing or wholesale trade, or they may be
limited to specific industries, such as motor
carriers. While these studies usually are con­
ducted on a nationwide basis, they sometimes
(especially in the case of specific industry stud­
ies) are limited to one or several areas in which
the industry is concentrated. Industries are
defined on the basis of descriptions in the cur­
rent Standard Industrial Classification Manual,
published by the Office of Management and
Budget.3
Area studies are limited in geographic cover­
age, perhaps to a Standard Metropolitan Sta­
tistical Area,4 to a county or a group of
contiguous counties, or to a region or part of
a region. This type of study usually is con­
ducted on a cross-industry basis— that is, earn­
ings and hours data are collected for most
industry divisions, including mining; manu­
facturing; transportation (except railroads),
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
payroll records, using one of two collection
methods. Establishments may be solicited for
187

138

BLS HANDBOOK OF METHODS

information by mail or by personal visit. Those
establishments contacted by mail receive a
questionnaire form and a note requesting co­
operation in the survey and explaining the
nature and purpose of the study. Typically, the
questionnaire requests information on the prod­
uct or service of the establishment and number
of employees, as well as other establishment
information which may be pertinent to the
survey. 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 incentive 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
requested, are visited personally by the Bu­
reau’s field economists who may prepare the
data for the survey from company records, or
arrange with the company for the completion
of the questionnaire form. In addition, a sam­
ple 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 con­
sistency. Where data are questionable, a letter
is written to the respondent asking him to re­
view the item to assure its accuracy.

Sampling

Earnings and hours distribution studies are
conducted on the basis of a sample of all es­
tablishments 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 UI lists may be augmented by sources



such as other government agencies, or trade
directories.
The size of the sample depends on several
factors, among which are the size of the uni­
verse, the distribution of establishments by
number of employees, 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 de­
termining the appropriate size of the sample.
The sample usually is selected using a highly
stratified probability sampling design. Estab­
lishments are first grouped, or stratified, ac­
cording to industry, geographic location, and
employment size. Establishments in specific
industries or areas for which earnings and
hours data are to be presented separately are
grouped independently of establishments in
other strata, and sampled separately.
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 establishments 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 in­
clusion in the sample is greater for the large
than for the small establishment. Frequently
the entire stratum containing the largest estab­
lishments is included in the sample.
The following example shows a hypothetical
universe of 47 establishments in three employ­
ment-size groups. By applying the appropriate
sampling ratio to each size group (stratum),
the number of sample establishments is de­
termined.

Employment
size group
Under 25 _____
25-49 __________
50 and over

Number of
establishments
in universe
32
12
3

Sampling
ratio
1/4
1/3
1/1

Number
of sample
establishments
8
4
3

Weight of
each
sample
establishment
4
3
1

No assumption is made that the earnings and
hours structures of establishments not respond­
ing to the mail questionnaire are similar to
those of establishments which do respond.

139

EMPLOYEE EARNINGS AND HOURS

Therefore, a sample is taken of the non­
respondents following a procedure similar to
the one just described. Establishments in this
subsample are visited by Bureau field repre­
sentatives in order to obtain the required data.
Data are not always obtained for every es­
tablishment in the original sample. Generally,
approximately 60 to 70 percent of the sample
establishments supply usable data to the
survey.

Estimating Procedure

Although a greater proportion of large than
of small establishments is included in the sam­
ple, any possible bias which might result from
this difference is avoided by means of the esti­
mating procedure. Each establishment in the
sample is assigned a weight which is the re­
ciprocal of the sampling ratio in the stratum
from which it was selected. That is, an estab­
lishment 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 represents itself and
three other establishments in the stratum (see
previous example). Data for each establish­
ment are multiplied by the weight assigned to
the establishment. Thus, all establishments,
regardless of their size, are represented ap­
propriately in the final estimates.
An establishment in the subsample of non­
respondents 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 frac­
tion. 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 assigned. In the
case of an establishment included in the sam­
ple with certainty, another establishment 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 establish­
ment employment series.5 This adjustment is
5 See chapter 2, “Employment, Hours, and Earnings” .




necessary to reduce the hazards of sampling,
and because the State UI listings, which con­
stitute the universe or a large part thereof, are
prepared prior to the time of the survey and
thus do not account for establishments opened
or closed between the compilation of the lists
and the date of the survey.
Estimated average hourly earnings or
weekly hours of work are the arithmetic mean
of weighted individual employees’ earnings or
hours. Generally, they are derived by totaling
weighted individual hourly earnings or weekly
hours and dividing the sum by the weighted
number of employees 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.
The following example illustrates the more
common method of estimating group average
hourly earnings. Referring to the sampling
scheme described in the preceding example, as­
sume 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.

Weight X
4

3

1

employm ent=
30
20
15
50
40
30
150
100
50

Weighted
employm entx
120
80
60
150
120
90
150
100
50
920

Hourly
earnings=

Weighted
group
hourly
earnings

$1.50
2.00
3.00
1.50
2.50
3.00
2.00
2.50
3.00

$180.00
160.00
180.00
225.00
300.00
270.00
300.00
250.00
150.00
2,015.00

Estimated
group —Total weighted group hourly earnings = $2,0 15 .0 0
average
Total weighted group employm ent
920
hourly
earnings

^

140

BLS HANDBOOK OF METHODS

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

a description of the area or industry studied
and provides a summary and analysis of the
survey results. Important relationships and dif­
ferences among areas and industries are high­
lighted 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 report.

Analysis and Presentation

The Bureau generally issues a bulletin on the
results of each survey of employee earnings
and hours distributions. The report contains a
description and analysis of the survey results
as well as tabulations of data collected.
In industry studies, tabulations are pre­
sented for the entire industry, and frequently
they are also presented for important segments
of the industry. For example, in a survey of
manufacturing, in addition to data for the en­
tire industry division, data also might be pre­
sented 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 presentation, 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.
In area studies, which include most industry
divisions, data obtained are tabulated for all
the industries in the area combined, and then
separately for manufacturing and nonmanu­
facturing industries. Where sufficient data are
available, tabulations are presented for selected
industry groups as well.
Earnings are tabulated to show the distribu­
tion of employees by intervals of 5- and 10-cent
average hourly earnings. Hours are tabulated
to show the distribution of employees by
weekly hours of work. The total number of
employees, their average hourly earnings and
average weekly hours, of course, also are
shown. Data also may be presented to show
average weekly earnings and cross-tabulations
of average hourly earnings by weekly hours
of work.
The text accompanying the tabulations gives



Uses and Limitations

The design of these employee earnings and
hours distributions studies makes them par­
ticularly useful in the analysis of Federal
minimum wage and maximum hours legislation
— in analyzing the effects of legislation, in con­
sidering new legislation, and in formulating
wage and hours policy. The information is used
by the executive and legislative branches of
government, organized labor, business,, aca­
demicians, etc. Special tabulations are prepared
for the U.S. Department of Labor’s Employ­
ment Standards Administration, for use in the
Secretary of Labor’s annual report to the Con­
gress 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, earings, and hours.
Employment estimates are subject to some
error, and in smaller groups this error may be
relatively large. Therefore, these estimates fre­
quently 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 changes.
The average earnings reported are straighttime, and any premium pay for overtime or
late shift work is not reflected. Similarly, dif­
ferences in prevailing supplementary compen­
sation practices ( “ fringe benefits” ) among
establishments, industries, and areas are not
considered.

EMPLOYEE EARNINGS AND HOURS

BLS 3010b

O ffic e o f Managem ent
and Budget No. 4 4 -S -7 1 0 0 6
Approval expires 1 2 -31 -7 1

U.S. DEPARTMENT OF LABOR
B U R E A U O F L A B O R S T A T IS T IC S
W A S H IN G T O N , D.C. 20212

Report Number

*

r

Y our report will be
held in confidence

n
Keep this copy for
your company tile.

T h e d a ta , e x c e p t fo r Ite m

1 w h ic h

r e la te s t o th e

e n t ir e c o m p a n y , s h o u l d c o v e r t h e e s t a b l i s h m e n t s )
in t h e l o c a t i o n d e s i g n a t e d t o t h e le f t .

l_
SURVEY OF INDIVIDUAL HOURS AND EARNINGS
OF NONSUPERVISORY EMPLOYEES
(Check appropriate box.)
(2)

(O

ANNUAL GROSS SALES OR RECEIPTS FOR THE
COMPANY, ENTERPRISE, OR INSTITUTION

1

(3)

1

1
1
1
----------------1
Under
$250,000 to
$1,000,000
or more
$250,000
$1,000,000
Check the box which indicates the annual gross volume of sales or receipts (exclusive of excise taxes) from all related activities of
the company, enterprise or institution, including receipts from the establishment(s) covered by this report. Use the last calendar or
fiscal year.
2. ESTABLISHMENT INFORMATION:
Please enter the information requested in the columns below for each separate establishment covered by this report. Each place of
business, warehouse, or central office in a separate location is considered a separate establishment for the purpose of this survey.
(a)

Location: Identify each establishment by its street address and city.

(b)

Type o f Activity: Describe the activity of each establishment listed, for example, gas station; drug store; department store;
women’s apparel store; shoe store; dime store; hotel; motel; dance hall; bowling alley; race track; amusement park; movie
theatre; etc.

(c)

Employment: Report all full-time, part-time, seasonal, and casual employees who received pay for any part of the payroll

period including May 12, 1971. Include those paid for sick leave, holidays, vacations, etc. DO NOT INCLUDE proprietors,
m em bers o f u n in co rp o rated firm s, o r unpaid fam ily workers.
Total - Enter total number of employees. Include all classes of employees—
executive, administrative and professional, as well

as nonsupervisory employees.
Nonsupervisorv - Enter number o f workers below the supervisory level. Include employees such as inside salespersons, shipping

and receiving clerks, routemen, laborers, warehousemen, repairmen, installers, room clerks, waiters and waitresses, nurses,
office clerks, janitors, watchmen, etc. DO NOT INCLUDE driver salesmen, outside salesmen, executive, professional, and
supervisory personnel.
(d)

Collective Bargaining: Are a majority of nonsupervisory employees covered by union-management agreements? (Check appro-

(e)

Annual Gross Sales or Receipts fo r the Establishment: Check the column which indicates the annual gross volume of sales or

priate column).
receipts (exclusive of excise taxes).
(b )

(a )

(c)
E m p lo y m e n t
f o r p a y r o ll p e r io d
in c lu d in g

L o ca tio n

M ay 12,

T y p e o f A c t iv it y

(s tre e t a d d ress

N on­
s u p e rv iso ry

s a le s

A g reem en t
( A r e m a jo r it y

( w e r e la s t y e a r ’s s a le s

covered ?)

$ 2 5 0 ,0 0 0 o r m o r e ? )

1971

a n d c ity )
T o ta l

(d )
(e )
U n io n -m a n a g e m e n t G r o s s e s ta b lis h m e n t

Yes

No

Y es

No

0 )

(2 )

to

(2 )

(f

)

Fo r
BL S
On iy
i

2
3
4
5
6

7
8

9

(0

How many hours per day and days per week are the majority o f your nonsupervisory employees scheduled to work?

(g)

After how many hours per day would your nonsupervisory employees receive overtime premium pay?
A fter______hours per day
or.
I
I we do not pay daily overtime.

Hours per day |

[

Days per week [

|

Check here if you want a copy o f the Bureau’s report on this survey. I

E n te r h o u r s t o o n e
d e c im al, e.g., 7 .5

I

Name and title of person furnishing data




(Please Type or P rint)

Area code, Phone No.

142

BLS HANDBOOK OF METHODS
Page 2
3. HOURS AND EARNINGS OF NONSUPERVISORY EMPLOYEES

Definitions and Instructions
(See also the examples below)
Hours worked and paid for and straight-time earnings of all nonsupervisory employees are requested for the payroll period (weekly,
biweekly, semimonthly, or monthly) which includes May 12, 1971. Please report and identify the data separately for each of the
locations listed in item 2(a).
• Nonsupervisory Employees are defined in item 2(c) on page 1 of this questionnaire. Remember to EXCLUDE outside sales­
men, driver salesmen, executive, administrative, professional, and supervisory personnel.
• Hours should include all hours worked and hours paid for sick leave, holidays, vacations, etc. Also include the number of over­
time hours actually worked.
• Straight-time Earnings should relate to the hours worked and paid for, including straight-time pay for overtime, holidays, week­
ends, and late shifts. Also include commission and bonus earnings. Exclude, however, all premium payments. (For example, if
overtime is paid at time and one-half report only two-thirds of this pay).
• These data are requested separately fo r each nonsupervisory employee. However, one entry may be made if two or more emplovees worked identical hours and received identical straight-time pay (see example A below). To avoid correspondence, please
do not report aggregate hours and earnings for several employees.
• Detailed instructions and examples for reporting the necessary data in each column are listed below.
Column (2 ) - Use a separate line for each employee and enter “ 1,” unless two or
more employees worked the same number of hours and received identical hourly
or salary rates (see example A). Please check to see that the sum of the employees

ALL NONSUPERVISORY EMPLOYEES,
COMPLETE COLUMNS (2) and (3)

in this column equals the number of nonsupervisory employees reported in item
2(c) on page 1.
Column (3 ) - Enter the number of hours paid for during the week of .May 9-15,

1971 regardless of the length of the payroll period. Include hours paid for sick
leave, holidays, vacations, etc. Count the number of overtime hours actually
worked. (Not hours “ boosted” to reflect premium pay.)
EMPLOYEES FOR WHOM STRAIGHT-TIME
HOURLY EARNINGS ARE AVAILABLE,
USE COLUMN (4)

Column (4) - Enter the average straight-time hourly earnings. Exclude premium
payments for overtime and for weekend, holiday, and late shift work. (You
should use this column to report earnings of employees even when paid on a
salary basis if average straight-time hourly earnings are available). If the hourly
earnings exclude commissions or bonuses, also complete columns 7 and 8 (see
example D).
Column (5 ) - Enter the straight-time salary for the salary period (weekly, bi-

EMPLOYEES PAID ON A SALARY BASIS
(IF HOURLY EARNINGS ARE NOT
AVAILABLE),
USE COLUMNS (5) and (6)

EMPLOYEES RECEIVING COMMISSION
AND/OR BONUS PAY,
USE COLUMNS (7) and (8)

weekly, semimonthly or monthly) which includes May 12, 1971. Include
straight-time pay for overtime but exclude overtime premium. Do not include
“draws” against commission as salary. For employeespaid a commission or bonus
in addition to a salary, also complete columns 7 and 8 (see example E).
Column (6 ) - Enter the number of hours paid for during the payroll period for
which earnings were reported in column 5 to permit accurate calculation of aver­
age straight-time hourly earnings.
Column (7 ) - Enter for each employee the total commission and/or bonus pay,
including “PM’s,” “Stims,” or any special bonuses based on sales paid quarterly
or oftener. This pay is to be reported for the commission or bonus period includ­
ing May 12, 1971. If the commissions earned during that pay period are not
representative of normal commission earnings, a longer period may be used. If
employees receive both commission and bonus payments for an identical period
of time, report the combined figure (see example D). If bonus payments cover a
period longer than the commission period, add only the prorated am ount of the
bonus to the commission earnings that correspond to the commission period
(see example E).
Column (8) - Enter the number of hours worked during the commission or bonus
period. (The hours should refer to the total hours worked during the period and
not necessarily 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).

EXAMPLES
(See illustrations on next page)
A.

Two employees each worked 36 3/4 hours during the survey week and each was paid a straight-time hourly rate of $2.20.

B.

One employee, who worked 40.0 hours during the survey week, received a salary of $168.50 (exclusive of any premium pay) for
88.0 hours worked during the salary period (!4 month).

C.

One employee, who worked 32.5 hours during the survey week, was paid on a straight commission basis and received $361.20
for 168 hours worked during the commission period (1 month).

D.

One employee, who worked 40.0 hours during the survey week, was paid a straight-time hourly rate of $1.95 and also received
$35.00 in commissions and $7.50 in “PMY" for 173.6 hours worked during the commission period (1 month).

E.

One employee, who worked 37.5 hours during the survey week, was paid a weekly salary of 575.00, received commissions of
$102.00 for 162.0 hours worked during a 1-month period, and $150.00 in bonuses earned during a 3-month period. Only 1/3 of
the bonus ($50.00) is reported so that the bonus period corresponds to the commission period.




143

EMPLOYEE EARNINGS AND HOURS

3. HOURS AND E A R N IN G S O F N O N S U P E R V IS O R Y
C o m p le te th e s e co lu m n s
for a ll n o n su p e rv isory
e m p lo y e e s

EM PLOYEES

U s e th is
colum n if

U s e th e s e co lu m n s for n o n su p e rv isory e m p lo y e e s paid

hourly e a rn in gs

other than on an hourly b a s is .

a v a ila b le

Num ber
L in e
N o.

of
e m p lo y e e s

H ou rs p a id for
during the
w eek o f
M ay 9 15,

S alary

C o m m is s io n s a n d /o r B o n u s

S tra ig h t-tim e
s a la r y for

1971

during

in c lu d in g
M a y 1 2 , 1971

hourly ea rn in gs

H ou rs p a id for

sa la ry p eriod

S traigh t-tim e

s a la r y period

T o ta l

H ou rs p a id for

co m m is s io n s
a n d /o r

co m m issio n

during

b on us pay

p eriod

( N e a r e s t tenth)

(i)

(N e a r e s t c en t)

( N e a r e s t cen t)

(N e a r e s t tenth)

( N e a r e s t c en t)

(N e a r e s t tenth)

(3)

(4)

(5)

(6 )

(7 )

(8)

(2 )

I llu s tr a tio n s o f e x a m p le s on p a g e 2.

A.

2

3 6 .8

B.

1

4 0 .0

C.

1

3 2 .5

D.

1

4 0 .0

E.

1

3 7 .5

$

2 .2 0

$

$

1 6 8 .5 0

8 8 .0
3 6 1 .2 0

7 5 .0 0

37. 5

1 6 8 .0

4 2 .5 0

1 .9 5

17 3 .6

1 5 2 .0 0

1 6 2 .0

DATA FOR EACH ESTABLISHMENT SHOULD BE REPORTED SEPARATELY AND THE ESTABLISHMENT IDENTIFIED.
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36




144

BLS HANDBOOK OF METHODS
3. HOURS AND EARNINGS OF NONSUPERVISORY EMPLOYEES

C o m p le te th e s e co lu m n s
for a ll n o n su p e rv isory
e m p lo y e e s

U s e th is
colum n if

U s e th e s e co lu m n s for n o n su p e rv iso r y e m p lo y e e s p a id

h ourly e a rn in gs

oth er than on an h ourly b a s i s .

a v a ila b le

Salary

H ou rs p a id for
Number

during the

of
L in e
N o.

w eek o f

e m p lo y e e s

M ay
15,

12-

(i)
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42




(2 )

s a la r y for

during

in c lu d in g

s a la r y p eriod

M ay

$

$

H ou rs p a id fo r

c o m m is s io n s

during

a n d /o r

c o m m issio n

b o n u s pay

(6 )
$

( N e a r e s t tenth)

(7 )

( N e a r e s t tenth)

pe rio d

( N e a r e s t cen t)

1971

(5)

(4 )

(3 )

12,

( N e a r e s t c en t)

( N e a r e s t c en t)

T o ta l

H ou rs p a id fo r

s a la r y period

S tra ig h t-tim e
hourly e a rn in gs

1971

( N e a r e s t ten th )

C o m m is s io n s a n d /o r B o n u s

Stra ig h t-tim e

(8)

145

EMPLOYEE EARNINGS AND HOURS

Technical References
Number

1. U.S. Department of Labor, Bureau of Labor Statistics: Employee Earnings and Hours in
Eight Metropolitan Areas of the South, June 1965 (Bulletin 1533, 1966).
2 . ______Employee Earnings and Hours in Retail Trade, June 1966 (Bulletin 1584, 1968).
3 . ______Earnings in Wholesale Trade, June 1958, (Bulletin 1253, 1959).
4 . ______ Factory Workers’ Earnings, May 1958 (Bulletin 1252, 1959).
5 . ______ Employee Earnings in Nonmetropolitan Areas of the South and North Central Re­
gions, June 1962 (Bulletin 1416, 1964).
6 . ______Employee Earnings in Retail Trade, June 1962 (Bulletin 1380, 1963).
Each of these bulletins contains a detailed description of the method used in the survey.




— A

l v in

B

a u m a n




Chapter 16. Union Wage Rates
Background and Description o f Survey

Annual studies of union wage rates and
hours are conducted in four industries: build­
ing construction, local transit, local trucking,
and printing.1 Union wage rates and hours are
those agreed on through collective bargaining
between employers and trade unions; they 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 maximum
number of hours per week at straight-time
rates. Rates in excess of the negotiated min­
imum, which may be paid for special qualifica­
tions or other reasons, are excluded.
The use of union agreements or other union
records in studies of occupational wages is
practicable in industries that are characterized
by a high degree of organization and in which
(1) defined craft groupings persist, as in build­
ing construction or printing, or (2) key oc­
cupations can be clearly delineated, as in local
transit.
The Bureau’s annual union wage studies
began in 1907. Originally, information was ob­
tained for 39 cities, but the number was ex­
panded gradually until in 1948, 82 cities were
covered.1 That number was reduced to 77 in
2
1949 and to 52 in 1953. The studies were ex­
panded again, after the 1960 Census of Popula­
tion, to the present coverage of 68 cities having
100,000 inhabitants or more. The scope of the
information for individual industries has also
been expanded. For example, 24 journeymen
crafts and nine helper and laborer classifica­
tions in the building trades are covered current­
ly, in place of the 13 journeymen and 7 helper
and laborer classifications in the initial studies.
The study of union wage rates and hours
in the building trades includes virtually all
journeymen and helper and laborer classifica­
tions. Indexes and other data are shown for
each important trade as well as for all trades
combined.3
The trucking study embraces drivers and
helpers engaged in local trucking. Over-theroad drivers and local city drivers paid on a
mileage or commission basis are excluded. All



data, including indexes, are presented for the
two classifications indicated.
Union wages and hours in the local-transit
industry are limited to operating employees.
Data are shown separately for operators of
surface cars and buses, and elevated and sub­
way lines, except that indexes are shown only
for the industry as a whole.
In the printing industry, 12 book and job
trades, 8 newspaper trades and 6 lithography
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, (except
lithography) for each trade and for all trades
combined.

Data Sources and Collection Methods

The union wage studies are designed to in­
clude 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 that 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 that 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 representatives and entered on
forms designed specifically for this purpose.
1 The coverage at various times also included barbers, line­
men, longshoremen, and workers engaged in breweries,
laundries, metal trades, millwork, restaurants, soft-drink
production, theaters, and bakeries. The Bureau plans to
conduct the first in a series of biennial studies of union wage
rates for grocery store employees in July 1971.
2 In these studies, data relate to individual cities and con­
tiguous suburban areas, rather than to the much broader
Standard Metropolitan Statistical Areas which are used in
most other Bureau surveys.
3 In addition to the annual studies in the building trades, a
quarterly survey of 7 major construction trades is conducted
in 104 cities. Estimated average hourly wage rates for all
trades combined and for each surveyed trade are presented,
together with the estimated change during the quarter and
the year.

147

148

BLS HANDBOOK OF METHODS

Information requested relates to July 1 for
all industries. This date was adopted, after
numerous changes, because most new agree­
ments in these industries have been negotiated
by that time each year. In order to maintain
year-to-year comparability, wage rate, hours,
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 also are requested from union officials
for the purpose of (1) checking the data en­
tered 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 149-151.

Sampling and Estimating Procedures

The current series is designed to reflect union
wage rates and hours in all cities of 100,000
inhabitants or more, excluding Honolulu. All
cities of 500,000 inhabitants or more are in­
cluded, 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 distributed
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
combination of data, each region is considered
separately when city weights are assigned.5
Rates

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

Chain indexes are calculated for each of the
four industries to portray the trend of union
wage rates and weekly hours. In calculating
these indexes, the percent change in aggregates
is computed from quotations for all identical
classifications 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 par­
ticular classification for the current year. The
index for the current year is computed by
multiplying the index for the preceding year
by the ratio of the aggregate change. In the
1970 study of building trades, the rate aggre­
gate for all quotations increased 11.6 percent
over the previous year. The July 1, 1970, index
of union hourly wage rates for all building
trades (128.8) is the result of multiplying the
July 1, 1969, index (115.4) by the ratio of the
aggregates (111.6). This method of index cal­
culation minimizes the influence of year-to-year
changes in membership.
Indexes of union hourly wage rates and
weekly hours are computed for each classifica­
tion as well as for all classifications combined
in the building construction and printing in­
dustries. 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 computation of an
index for union weekly hours in the local
transit industry.
The base period for the indexes of union
wage rates and weekly hours is the 1967 aver­
age. The series for the building trades and
printing industry date back to 1907, for local
4 See chapter 20. “ Collective Bargaining Agreements.”
5 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 expan­
sion in the program.
6 Reported membership, as used in this study, is defined as
members working or immediately available for work.

149

UNION WAGE RATES
B L S 1150.1
(R e v . 1971)

B ud get B ureau N o . 4 4 -R 7 3 8 .1 2 .
A p p ro v a l e x p ir e s 3 -3 1 -7 2 .

U.S. DEPARTMENT OF LABOR
BU R EA U OF L A B O R ST A T IST IC S

r

n

L

J

Dear Sir:
The Bureau of Labor Statistics is now conducting its annual survey of union wage scales and
hours in the building trades. The continued success of these surveys, begun in 1907, depends
largely upon your cooperation and we are asking that you furnish information requested on this
schedule form The completed form should be signed and returned to the Bureau in the enclosed
.
self-addressed envelope which requires no postage. W would appreciate your returning the
e
completed form promptly.
Thank you very m
uch for your cooperation.
Very truly yours,

Regional Director.

INSTRUCTIONS FOR COMPLETING THIS FORM

P A Answ general inform
art :
er
ation questions in this p of schedule.
art
P
art B Check an revise, if necessary, inform
:
d
ation reported for July 1, 1970. If any of you
r
trade classifications h
ave been om
itted, insert th an th negotiated scales in th appropriate
em d e
e
colum
ns. Sh
ould you have m re th one negotiated scale for a trade o classification, rep
o
an
r
ort
separate inform
ation for each scale. For each trade or occupation group show in colum (2),
n
n
enter in colum (5) the w
n
age scale in effect on July 1, 1971, and in colum (10), the n m er of
n
u b
w
eekly h rs before overtim rate is effective. The n m
ou
e
u ber of union m bers at each scale
em
should b entered in colum (15), an th n m er of apprentices, as of July 1, 1971, should be
e
n
d e u b
entered in colum (17) for each trade.
n
Copy of A
greem
ent: Please provide a copy of each of you agreem
r
ents to assist us in o r analy­
u
sis of collective bargaining provisions. If you have only one copy avaiblale, w shall be glad
e
to m e a duplicate an retu the original prom
ak
d
rn
ptly. If requested, th agreem will b kept in
e
ent
e
confidence and m
aterial used only for general analysis w
hich will not reveal the n e of either
am
party to th agreem
e
ent.



150

BLS HANDBOOK OF METHODS

UNION SCALES OF WAGES AND HOURS IN THE BUILDING TRADES
(A nnual Survey)
PART A.-GENERAL INFORMATION:

Please attach a copy of you agreem in effect on July lf 1971, an answ th follow questions:
r
ent
d
er e
ing
1. W en did you agreem go into effect?_________________________________________________________________
h
r
ent
h
W en m it be reopened?__________________________ W en does it expire?_________________________________
h
ay
2. W at is the m u n m er of h rs th can b w
h
axim m u b
ou
at
e orked each day before overtim rate is effective?_______________
e
3 Does you agreem provide for a health an insurance plan (life insurance, hospitalization, m
*
r
ent
d
edical, surgical, an
d
other sim types of health an w
ilar
d elfare program financed—
s)
Entirely by em
ployer? Q Yes □ N
o
In p by em
art
ployer? □ Yes Q N
o
A ou of em
m nt
ployer contribution________________ per H r □ Shift Q W
ou
eek Q M th Q
on
(c e n ts or p e rce n t)

4. Does you agreem provide for a pension plan financed—
r
ent
Entirely by em
ployer? Q Yes □ N
o
In p by em
art
ployer? □ Yes
N
o
ou
eek Q M th Q
on
A ou of em
m nt
ployer contribution________________ .per H r Q] Shift Q W
(c e n ts o f p e rce n t)

5 Does you agreem specifically provide for a paid vacition, financed by em
*
r
ent
ployer paym
ents—
(A) To a vacation fu d □ Yes Q N
n
o
FU PAYM
ND
ENTS DEDUCTED FR M PAYROLL UPON W
O
RJTTEN IM
, PLIED, O CONTRACTURAL AUTHOR­
R
IZATION O W R ER ARE NOT CO SIDERED EM
F OK S
N
PLOYER PAYM
ENTS
(B) To w
orker each pay period, as p of negotiated scale □ Yes □ N
art
o
(C) To w er each pay period, in addition to negotiated scale Q Yes Q N
ork
o
(D O
) ther—
(Explain)____________ ________________________________________________________________
If answ is Yes to any of above, indicate am n of em
er
ou t
ployer payment__________________________________ per h r.
ou
(c e n ts or p e rce n t)

6. Does you agreem provide for em
r
ent
ployer paym
ents to oth funds, such as holiday, educational, prom
er
otional, u em
n ­
ploym benefits? Q Yes Q N
ent
o If YES, list below each type of contribution separately an am n of em
d ou t
ployer
paym
ents (cents o percent).
r

ow an
en?_____________ H m are
ow any
7. W at is th total m bership of you union?_____________ H m y are journeym
h
e
em
r
apprentices?______________ H m y are helpers an laborers?_______________
ow an
d
8. Betw
een July 1, 1970, a d July 1, 1971, h m y apprentices com
n
ow an
pleted their apprenticeship?
(If n e enter zero.)_______________
on
9. Betw
een July 1, 1970, an July 1, 1971, h m y jou eym becam unavailable for w because of death, perm
d
ow an
rn en
e
ork
an­
ent disability, o retirem
r
ent? (If n e enter zero.)_______________________________________________ ________
on

P le ase sign your name here________________________________________ T itle _________________________

D a te ______________________

, 1971.

A ddress_____________________________________________________________ C ity, State, and Z i p C o d e ___________________________________________




U. S. DEPARTMENT OF LABOR
Bureau o f Labor Statistics

Union Scales of Wages and Hours in the Building Trades (Annual Survey)
Part B.

Scale and membership information by trade:

Please enter the data requested below. Do not use columns marked ‘ Code.’
Employer payments to insurance, pension, vacation, or other funds should be excluded from the hourly scales below, and shown as cents or percent of scale in the spaces pro­
vided in part A.
Payments specifically designated as being in the negotiated wage scale for any benefit should be excluded from hourly scales reported below even though they are made or
credited to the worker each pay period.
Membership information will be kept in confidence and used only to compute average wage rates.
Schedule No.______

Trade or occu p a tion

Page_____ of
Number o f union mem bers working or
im m ed ia tely a vailable for work
at each rate on July 1—

Weekly hours before overtim e rate
is e ffe ctiv e on July 1----

W age scale for each trade or occupation
in e ffe ct on July 1—

Number o f apprentices in
ea ch trade on July 1—

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(ID

(12)

(13)

(14)

(15)

(16)

(17)

(18)

C od e

T itle

1969

1970

1971

Code

Code

1969

1970

1971

C ode

C ode

1969

1970

1971

1970

1971

Code

UNION WAGE RATES
151




152

BLS HANDBOOK OF METHODS

transit to 1929, and for local trucking to 1936.
Although data for the latter two industries
were collected for years before the dates of
the index series, indexes were not constructed
because of inadequacies in the available data.

Analysis and Presentation

The averages and indexes mentioned to­
gether with other summary data are contained
in the bulletins published annually for the in­
dividual studies. Included among the informa­
tion shown for individual trade classifications
is the proportion of union members having
hourly rates at different levels, as well as the
proportion of union members having, since the
previous study, wage rate increases of specified
amounts in terms of cents per hour and per­
cent. The increase registered by the trade is
shown also.
In addition, the union rates of wages and
hours in effect on the date of the survey, as
reported by union officials, for both the previ­
ous and current years are published for each
classification by city. These furnish a direct
comparison of union rates between the 2 years
for each of the industries studied. The rates of
wages are indicated as hourly rates and the
hours as the weekly hours of work before over­
time rates are applicable. The current studies
also present data on employer payments for
insurance (health and welfare) and pension
payments; in addition employer payments for
vacation and other funds (except those for
apprenticeship) are shown for the building
trades. These payments are expressed in terms
of cents per hour or as percent of rate.
7 Membership (used for weighting purposes) 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, and it does not necessarily
reflect metropolitan area rates.




Uses and Limitations

The Bureau’s union-wage series provide a
means of determining intercity wage differ­
ences for comparable work, and the relation­
ships 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 wage rates of
building-trades workers are especially impor­
tant in estimating construction costs, because
labor expenditures constitute an important
element in the total cost of building construc­
tion. The index series derived from these
studies provide barometers of year-to-year
changes in rates 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 because 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 rates. For example, if organiza­
tional drives in cities having relatively lower
rates of wages result in sharp increases in
membership, the movement of the rate levels
for the affected trades as a whole is naturally
retarded. Conversely, increases in membership
in cities having high wage rates accelerate the
upward movement of averages.7
The union rates, are not necessarily the
actual rates paid to all workers, and the union
hours are not necessarily the hours actually
worked. Workers with above average exper­
ience and skill may be employed at rates above
the union wage rates, especially during pros­
perous times when a tight job market creates
competitive bidding for the better workmen.
During periods of depressed business activity,
actual hours worked often are less than hours
specified in the union agreement.
— T homas

C.

M obley

Chapter 17. Current Wage Developments
Background

Since January 1948, the Bureau of Labor
Statistics has issued a monthly report listing
general wage changes and changes in supple­
mentary benefits agreed to in selected collective
bargaining situations, identifying the situa­
tions 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 approximately 1,000 or
more production and related workers in manu­
facturing and selected nonmanufacturing in­
dustries.
Current Wage Developments was initiated
because of the rapid increase in wage rates and
prices in the early post World War II period,
the interest in determining the extent to which
settlement patterns spread from industry to
industry, and the discontinuance of an index
of wage rates that had been initiated during
World War II. Interest in the listing was stim­
ulated by the Korean emergency when the
Wage Stabilization Board needed data on the
extent to which wages and benefits were being
changed.
In 1949, and again in 1951 and 1952, sta­
tistical summaries of wage changes were pre­
pared to supplement the listing, but regular
preparation of a statistical summary began in
1954. These quarterly 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 supplementary benefits
were introduced or changed.
Beginning in 1959, another statistical sum­
mary was instituted. It is limited to manu­
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 non­
union workers, including professional, white-collar, and pro­
duction employees, also is presented.
1 Only changes in benefits that represent changes in costs
are included.
3 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 at any one time or all workers
In an occupation. Changes resulting from promotions, merit
increases, etc., are excluded.




facturing, but includes information on general
wage changes for nonunion and small union
situations, as well as for large collective bar­
gaining situations.1 From 1959 through 1970,
this summary also included information on
changes in supplementary benefits.

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 in manufacturing firms
of all kinds is described as the “ manufactur­
ing” series.
The major series describes general wage
changes and changes in benefits1 in all collec­
2
tive bargaining settlements involving 1,000 or
more production and related workers in manu­
facturing and 1,000 or more nonsupervisory
workers in the nonmanufacturing sector, ex­
cluding Government.3
Supervisory or professional employees are
excluded. Large units of technicians are in­
cluded even though they are part of a bargain­
ing unit that is predominantly professional.
Contracts covering multiplant firms are
included if the agreement as a whole covers
1,000 or more workers even though each indi­
vidual plant employs fewer. Also included are
contracts with trade associations or with
groups of firms that bargain jointly with a
union or unions, even though the firms are not
associated formally and each has fewer than
the minimum number of workers covered 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 gen­
eral wage changes,4 regardless of whether the
workers are represented by a union.
Wage change data are presented in cents per
153

154

BLS HANDBOOK OF METHODS

hour and, since 1959, as a percentage of aver­
age 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 economic 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 deferred— that
is, decided upon earlier— or (c) resulted from
operation of cost-of-living clauses. In distribu­
tions 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 situation. The number of workers affected
by changes in supplementary benefits includes
all production and related workers in the situa­
tions where the benefit is changed, whether or
not all are affected immediately. For example,
if a fourth week of vacation is added for work­
ers with 20 years’ service in an establishment
employing 1,000 workers, a vacation change
would be recorded for 1,000 workers, even
though only a relatively small proportion 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
bargaining settlements presented in the
monthly Current Wage Developments listing
which, in turn, is derived primarily from sec­
ondary sources, including general circulation
newspapers and periodicals, as well as union,
management, and trade publications. Other
important sources of information are the file
of union contracts maintained by the BLS and
the U.S. Department of Labor’s files of pension
and health and welfare agreements, maintained
by the Office of Labor-Management and Wel­
fare-Pension Reports.5 By the end of the year,
6
the BLS 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 union­
ized firms is gathered quarterly (semiannually
in 1965 and 1966) by a questionnaire mailed to
participating establishments. The information
on general wage changes is supplemented by
the contract file (unionized establishments)
and from newspaper clippings, purchased from
a commercial clipping service. At the end of the
year, BLS field representatives contact, pri­
marily by telephone, a sample of firms that have
failed to respond to the mail questionnaire or
that have provided incomplete or unclear infor­
mation.

Sampling and Estimating Procedures

As indicated earlier, all bargaining situa­
tions with 1,000 or more workers in manu­
facturing and nonmanufacturing industries are
included in the major series. It is believed that
the current list of nearly 2,500 such situations,
built up since Current Wage Developments was
started in 1948, is very nearly complete. After
a bargaining situation 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 outside 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 em­
ployees, and industry classification.6 The sam­
ple 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 employees.7
5 Information from contracts supplied on a confidential basis
is used only in the statistical summaries, not for the monthly
listing.
6 See appendix B, "Industrial Classification.” For a more
detailed description of unemployment insurance data, see p.
15, chapter 2.
7 In the case of a few companies with large numbers of
establishments each with 1,000 workers or more, a sample of
plants is chosen.

155

CURRENT WAGE DEVELOPMENTS

U.S. D EPA RTM EN T OF LABOR

B L S 2675d

O f f ic e o f M anagem ent and
B u d g e t N o . 4 4 -R 1 1 3 5

BUREAU OF LABOR S T A T I S T I C S
W A S H IN G TO N , D .C .

A p p r o v a l e x p ir e s 3 /3 1 /7 2

20212

Wage Developments in Manufacturing, 1971
T he

B u reau

w ill

h o ld

is h e d
stric t

by

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

a ll

in fo r m a tio n

th e

fu r n ­

resp on d en t

in

c o n fid e n c e .

I d e n t ific a t io n or lo c a t io n o f e s ta b lis h m e n t
fo r

w h ich

in form ation

is

re q u e s t e d ,

if

d iffe re n t from m ailin g a d d r e s s .

( C h a n g e i f i n c o r r e c t , i n c lu d e Z I P

c o d e .)

I.

What w a s the m ajor p rod u ct (in term s o f s a l e s v a lu e ) o f th is plan t during 1970?

II.

P le a s e p ro v id e em p loy m en t and p a y ro ll in form a tion for the p a y r o ll p e rio d in clu d in g January 12, 1971.

S e e p a g e 4 for
e x p la n a tio n .

A.

A ll e m p lo y e e s
1. N um ber

III.

B.

A ll p r o d u c tio n and r e la te d w ork ers
1.

~~|

|

N um ber

2.

~| |$

P a y r o ll

3.

~| |

D o c o l l e c t i v e b arg a in in g a g reem en ts c o v e r a m a jority o f you r p ro d u c tio n and re la te d w o rk e rs ?

M an-hours

Y es [^ j

~

No Q

If " N o , ” p le a s e s k ip to s e c t io n V II. _____________________________________________________________________________________________________
If " Y e s , ” p le a s e a n sw er a ll q u e s tio n s e x c e p t X .
F O R UNION FIRM S O N L Y
IV .

U nion and A greem en t I d e n t ific a t io n :
A.

With what u n ion or u n ion s d o y ou h ave a c o l l e c t i v e b a rg a in in g a greem en t?

B.

A re you a party to an a greem en t s ig n e d by an em p lo y e r a s s o c i a t i o n w ith th is u n io n (s )?

Y es |

No Q

\

If s o , what is the a s s o c i a t i o n ’ s n a m e ? ______________________________________________________________________________________ ,__________
V.

A greem en t E x p ira tion D a te:
When d o e $ you r c o l l e c t i v e b a rg a in in g a g re e m e n t(s ) e x p ire or b e c o m e s u b je c t to r e o p e n in g on w a g e s ?
(S p a ce ha s b een p ro v id e d for tw o e n trie s s in c e the d a te may ch a n g e during the y e a r .)

V I.

N ew or R e v is e d A g reem en t:

P le a s e mark ap p rop ria te b o x e s b e lo w to brin g you r rep ort up to d a te fo r —
J a n .—Mar.

A.

A p r .—June

1971

During the q uarter—

1971

1971

O c t .—D e c .
1971

D id you n e g o tia te a new or r e v is e d
c o l l e c t i v e b a rga in in g c o n t r a c t (s ) for your

Y es

p ro d u ctio n and re la te d w o r k e r s ? ............................................... ........................□
B.

J u ly —S ept.

D id you a gree on an im m ediate or
d e fe rre d ch a n g e in w a g e s ? .........................................................




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

□

No

□
□

Y es

No

Y es

No

Y es

No

□ □ □ □ □ □
□ □ □ □ □ □
PLEASE TURN TO NEXT PAGE.

156

BLS HANDBOOK OF METHODS

FOR UNION AND NONUNION FIRMS
V II.

W a g e-R ate C h a n g e s for P r o d u ctio n and R e la t e d W orkers, 1971:
P le a s e rep ort any g e n e ra l w a g e -ra te c h a n g e s y ou h a ve put in to e f f e c t fo r you r p ro d u c tio n and re la te d w o rk e rs in the p re v io u s
quarter. (F o o d - p r o c e s s i n g c o m p a n ie s s h o u ld a l s o rep ort w a g e -r a te c h a n g e s a ffe c t in g d r iv e r -s a le s m e n .)

In c lu d e :

E x c lu d e:

(1 )

A ll c h a n g e s a ffe c t in g eith er (a ) 10 p e rce n t or m ore o f you r
p ro d u c tio n and r e la te d w ork ers at any on e tim e, or (b ) a ll
w ork ers c o v e r e d b y a s in g le a g reem en t, e v e n i f the a g r e e ­
m ent a p p lie s to fe w e r than 10 p e rce n t o f the w o r k e rs .

(1 )

I n c r e a s e s to in d iv id u a ls re s u ltin g from p ro m o tio n s ,
a u to m a tic i n c r e a s e s w ith len gth o f s e r v i c e , or
p r o g r e s s io n w ith in an e s t a b lis h e d rate ra n g e.

(2 )

I n c r e a s e s d e c id e d on in 1971 but s c h e d u le d to g o

(2 )

A ny ch a n g e in y ou r p a y s c a l e s e v e n though no w ork e rs r e ­
c e i v e d im m ed iate p ay i n c r e a s e s a s a re s u lt o f th is ch a n g e .

(3 )

A ny c o s t -o f - l i v i n g e s c a la t o r a d ju stm en ts w h eth er or not
they are part o f you r perm anent rate stru ctu re.

(4 )

I n c r e a s e s d e c id e d on in e a rlie r y e a r s but g o in g in to e f f e c t
in 197 1 .

(5 )

C h a n g e s in h ou rly ra te s r e s u ltin g from c h a n g e s in hou rs
w ith ou t c o r r e s p o n d in g c h a n g e s in w e e k ly or d a ily p a y .

A.

(3 )

(T h e s e s h o u ld b e lis t e d

T h e c o s t o f any c h a n g e s in su p p lem en ta ry
b e n e fit s .

H ave you put in to e f f e c t a n y s u ch g e n e ra l w a g e c h a n g e s during the quarter?
Jan .-M ar.
1971
Y e s ............................................................................................. . . . .

Q

N o ............................................................................................. . . . .

Q

If you r a n s w e r is " N o ” , p r o c e e d to s e c t io n V III.
and lis t in s u b s e c t io n C :

A p r.-J u n e
1971

J u ly -S e p t.
1971

□
□

□
□

O c t .- D e c .
1971

□
□

If you r a n sw e r is " Y e s ” , in d ic a te b e lo w the form o f the w a g e c h a n g e (s )

(1 )

U niform c e n ts p er h o u r ........................................... . . .

Q

(2 )

U niform p e rce n ta g e c h a n g e ................. - ................ . . .

Q

(3 )

H igh er c e n t s p er hour for s k ille d w ork ers

- -, . . .

Q

(4 )
C.

in to e f f e c t in la ter y e a r s .
in s e c t io n V III.)

O ther ( s p e c i f y in s e c t io n X II. " R e m a r k s ” ) - -. . . .

Q

□
□
□
□

□
□

□
□
□
□

□

□

If a ll w ork ers d id not r e c e iv e the sam e am ount (e ith e r the sam e num ber o f c e n ts or the sam e p e r c e n ta g e ) lis t c h a n g e s for
e a c h grou p on a s e p a ra te lin e with the a p p roxim ate number a f f e c t e d . F o r e x a m p le , if there w a s a uniform a c r o s s -t h e board ch a n g e p lu s a d d ed c h a n g e s for s om e w o r k e rs , lis t the uniform ch a n g e firs t and sh ow a d d itio n a l c h a n g e s b e lo w . If a
c o s t -o f - l i v i n g e s c a la t o r ad ju stm en t w ent in to e f f e c t at the sam e tim e a s a n oth er in c r e a s e , l i s t it s e p a r a t e ly .
In rep ortin g in form a tion for in c e n tiv e w ork ers i n c l u d e , i f p o s s i b l e , e s tim a te d e f f e c t s o f w a g e -r a te c h a n g e s on in c e n tiv e
w o r k e r s ’ e a r n in g s . (F o r e x a m p le , if b a s e ra tes for in c e n tiv e w o rk e rs w ere r a is e d 5 c e n ts and th is in c r e a s e d th e ir h ou rly
ea rn in g s a bou t 7 c e n t s , report 7 c e n t s .)
If any c h a n g e s in s c a l e s w ere m ade that d id n ot a ffe c t any w o rk e rs im m e d ia te ly , in d ic a te the a p p ro xim a te num ber to be
______________________________________________
a ffe c t e d by the en d o f the y ea r.
S e e p a g e 4 for e x a m p le o f
In d ica te w h eth er ch a n g e w a s g iv e n in p e rce n ta g e or c e n ts term s.
re p o rtin g g e n e ra l w a g e c h a n g e s .

Approxi­
mate
number
Effec­ receiving
wage
tive
adjust­
date
ments




Check if this was:
Classes of production and related
workers or jobs affected

Deferred
(effective
in 1971 but
decided
earlier)

Hourly
change
for
workers
(+ or -)

%

t

%
%
%
%

Auto­
matic
cost-ofliving
escalator

157

CURRENT WAGE DEVELOPMENTS
V III.

Future W a g e-R a te C h a n g e s :
L i s t any g e n e ra l w a g e -r a te c h a n g e s a lre a d y d e c id e d upon fo r you r p ro d u c tio n and r e la te d w o rk e rs but s c h e d u le d to g o
in to e f f e c t in the future.

Amount of change for

W orkers a f f e c t e d

E ffe c t iv e d a te
(m on th , d a te , and y e a r)

workers whose rates
E stim a ted
N um ber

C l a s s e s or jo b s

will be changed

_________ L

%

_________ L

%

i

%

%

IX .

C o s t -o f-L i v i n g E s c a la t io n :
D o you h a v e a c o s t -o f - l i v i n g e s c a la t o r p o l ic y w h ereb y w a g e s are c h a n g e d a u to m a tic a lly w ith s p e c i f i e d c h a n g e s in a
p r ic e in d e x ?

Y es |

j No [^ ]

If s o , p le a s e l is t the m onths in w h ich w a g e c h a n g e s w ill g o in to e f f e c t i f the p rice

in d ex w a r r a n t s .______________________________________________________________________________________________________ ________________

F O R N ON U N IO N FIRM S O N L Y
X.

P o l i c y R eg a rd in g W age C h a n g e s :
W hether or not you c h a n g e d w a g e s durin g the firs t quarter o f 197 1 , w h at i s you r p o l ic y re g a rd in g g e n e ra l w a g e c h a n g e s ?
1. W ag es are n orm a lly ch a n g e d o n ly on an in d iv id u a l b a s i s .
2.

X I.
X II.

G en era l w a g e c h a n g e s are s o m e tim e s m ad e.

[[

[

D o y ou w ant a c o p y o f the B u rea u ’ s sum mary on th is s u r v e y ? ..................................... Y e s |

|

No

R em a rk s:

Name and t itle o f p e rs o n fu rn ish in g d ata




( P l e a s e T y p e or Print)

Area c o d e , Phon e No.

158

BLS HANDBOOK OF METHODS

4

Explanations for Report on Wage Developments
in Manufacturing, 1971
Section 11-A A ll employees—o ta l
t

number on the p a y r o ll o f the p la n t c o v e r e d b y th is rep ort w h o w o rk e d fu ll-tim e or p art-tim e or

r e c e iv e d p ay for any part o f the p e rio d re p o rte d .

I n clu d e p e r s o n s on p a id v a c a t io n s and s ic k l e a v e .

E x c lu d e p e r s o n s on le a v e

w ith ou t com p a n y p ay the e n tire p e rio d as w e ll a s p e n s io n e r s and m em bers o f the Arm ed F o r c e s n ot w ork in g during the p e rio d
re p o rte d .

Section ll-B

Production and related workers.

In c lu d e

w ork in g forem en and a ll n o n s u p e r v is o r y w o rk e rs e n g a g e d in fa b r ic 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 iv in g , s to r in g , h a n d lin g , p a c k a g in g , w a re h o u s in g , s h ip p in g , tru ck in g , h a u lin g , m a in te ­
n a n c e , rep a ir, ja n it o r ia l, w a tch m en s e r v i c e s , p rod u ct d e v e lo p m e n t, a u x ilia ry p ro d u c tio n for p la n t’ s ow n u s e ( e . g . , p o w e r -p la n t),
re c o r d k e e p in g , and oth er s e r v i c e s c l o s e l y a s s o c i a t e d w ith t h e s e p r o d u c tio n o p e r a t io n s .

E x c lu d e

e m p lo y e e s e n g a g e d in e x e c u t iv e , p u rc h a s in g , fin a n c e , a c c o u n t in g , p e r s o n n e l, c a f e t e r ia , p r o f e s s i o n a l , and t e c h n ic a l

a c t iv i t i e s ; s a l e s , a d v e r t is in g ; c o l l e c t i o n ; in s t a lla t io n , and s e r v ic in g o f p r o d u c ts ; rou te o f f i c e fu n c t io n s ; fa c t o r y s u p e r v is io n
a b o v e the w ork in g forem en l e v e l ; and e m p lo y e e s on you r p a y r o ll e n g a g e d in c o n s t r u c tio n o f m ajor a d d itio n s or a lte r a tio n s to
the p la n t.

Number.

In clu d e both fu ll-tim e and p art-tim e p r o d u c tio n and r e la te d w ork ers on you r p a y r o ll—w h eth er w a g e or s a la r ie d —w h o

w ork ed during or r e c e iv e d p a y for any part o f the p a y r o ll p e r io d r e p o rte d .

In clu d e p e r s o n s on p a id s ic k le a v e , p a id h o lid a y s ,

and p a id v a c a t io n s .

Payroll.

In clu d e p a y e a rn ed durin g the p a y r o ll p e rio d by p ro d u c tio n and r e la te d w ork ers re p o rte d in the p r e c e d in g b o x .

P a y r o ll s h o u ld b e re p o rte d b e fo r e d e d u ctio n fdr o ld -a g e and u n em p loym en t in s u r a n c e , g ro u p in s u r a n c e , w ith h o ld in g ta x , b o n d s ,
and un ion d u e s .

Man-hours.

In clu d e c o s t -o f - l i v i n g a l lo w a n c e s , p a y for o v e r tim e , h o lid a y s , v a c a t io n s , and s ic k le a v e .

In clu d e a ll h ou rs w ork ed , not s c h e d u le d h o u rs , durin g the p a y r o ll p e rio d b y the p r o d u c tio n and r e la te d w ork ers

rep orted in the firs t b o x p lu s h ou rs p a id for sta n d -b y or rep o rtin g tim e and h o lid a y s , and m an -hours e q u iv a le n t to p a y r e c e iv e d
by e m p lo y e e d ir e c t ly from you r firm for s ic k le a v e and for h o lid a y s and v a c a t io n s for th is p a y r o ll p e rio d .

Section Vll-C

E a ch p e rio d rep ort the gen era l w a g e -r a te c h a n g e s y o u h a v e put in to e f f e c t fo r y ou r p ro d u c tio n and r e la te d w o rk e rs

s in c e you r p r e v io u s rep ort t o u s .
for d r iv e r -s a le s m e n .)

E ffe c ­
tiv e
d a te

E s ta b lis h m e n ts in the f o o d - p r o c e s s i n g in d u s trie s s h o u ld a l s o rep ort w a g e -r a te c h a n g e s

C h e c k i f th is w a s :

A p p r o x i­
m ate
number
r e c e iv in g
w age
a d ju s t ­
m ents

1 /5

1,0 0 0

1 /5

9 50

2 /1 0

(N o te :

E n trie s o f v a riou s ty p e s o f w a g e c h a n g e s ca n be illu s tr a te d a s f o l l o w s :

1 ,0 0 0

C la s s e s o f p r o d u c tio n and re la te d
w ork ers or jo b s a ffe c t e d

D e fe rre d
( e f f e c t i v e A u to ­
in 1971
m a tic
c o s t-o fbut
d e c id e d
liv in g
e a r lie r ) e s c a la t o r

H ou rly ch a n g e for
w ork ers
(+ or - )

A ll

+ 8

A ll e x c e p t com m on la b or

A vg.

A ll
A ll

3

+ 2
4^ —
A vg.

0
<t

%
+ ^ to 5

%

0

%

8 0
6 0

X

%

6 /1 5

1,000

6 /1 5

100

M ain ten a n ce

+ 9 add.

0

%

8 /5

4 00

T im e

+6

0

%

8 /5

600

In ce n tiv e

+ 7

0

%

A ll

- 1

t

%

P ow erp la n t and m a in ten a n ce

+ 10

0

%

1 0 /1

1 ,0 0 0

1 2 /5

125




X

CURRENT WAGE DEVELOPMENTS

The ratios are uniform for all industries. 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
included in the summary for manufacturing.
The sample selected from the UI listings is
compared with this list of establishments for
which information already is available; since
data for these sample members are obtained
from secondary sources, these establishments
are not sent questionnaires. Approximately
6,000 establishments are left for the question­
naire survey.
Although the sampling design yields a
sample in which large firms are relatively over­
represented, this bias is overcome by the esti­
mating procedure. Each establishment in the
sample is assigned a weight which is the re­
ciprocal 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 assigned a weight
of 4, so that it represents itself and three other
establishments. Information for each establish­
ment is multiplied by the weight assigned to the
establishment. Thus, all establishments, re­
gardless of size, are represented appropriately
in the final estimates.
An establishment in the subsample of non­
respondents 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 non­
respondent establishments subsampled from a
group originally sampled at the rate of 1 out of
2 would be assigned a weight of 6. If an estab­
lishment included in the sample with certainty
fails to respond, another similar establishment
would be weighted to represent it.
To the estimates derived from the weighting
of the sample questionnaire are added the data
from secondary sources— the numbers of major
bargaining situations.
The totals thus obtained are further adjusted
to reduce the hazards of sampling and to take
account of opening or closing of establishments
■See chapter 2.




159

between compilation of the State unemploy­
ment insurance listing from which the sample
is chosen and the date of the survey. Adjust­
ments are made to employment levels for prod­
uction workers in the 2-digit Standard Indus­
trial Classification manufacturing industry
groups, as reported in the monthly employment
series of the Bureau of Labor Statistics for the
period covered by the summary of wage
changes.8 For example, if the estimate of prod­
uction worker employment in an industry
group derived from the sample is 100,000 but
the Bureau’s estimate of employment in that
industry group was 110,000 workers, each em110,000

ployment count would be multiplied by ^qq qqq
or 1.1 The resulting industry group estimates
would be added to provide the estimates for all
manufacturing.
The major series for manufacturing and
nonmanufacturing combined is not adjusted in
this fashion, since it is presumed to be all
inclusive.
A new sample of nonunion and small union­
ized plants in manufacturing usually is selected
every 3 years. Establishments with fewer than
four workers are omitted because in many
States they are not covered by these unemploy­
ment compensation programs. After the initial
contact, establishments of any size that indi­
cate 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

Preliminary information on the “ package
cost” and general wage changes resulting from
collective bargaining settlements involving the
major situations is issued in press releases
about 3 weeks after every quarter and the in­
formation is also summarized in Current Wage
Developments (See Chapter 23 for a descrip­
tion of the package cost program ). Final infor­
mation on wage and benefit changes is not
available until the end of the year. It is pre­
sented in Current Wage Developments and the
Monthly Labor Review. Yearend summaries
also present information on total effective

160

BLS HANDBOOK OF METHODS

wage changes— those negotiated in the period
plus deferred and cost-of-living changes re­
sulting from settlements in prior years but
going into effect in the current year. Informa­
tion is presented for manufacturing industries,
for nonmanufacturing industries, and for both
combined.
Quarterly summaries of the manufacturing
data are published in Current Wage Develop­
ments and the article covering the full year also
appears in the Monthly Labor Review. The
3-, 6-, and 9-month summaries, which are based
on preliminary data, stress wage changes re­
sulting from settlements or management de­
cisions made during the period, while the
yearend article, which is based on final data,
also analyzes trends in the size, frequency and
type of wage changes, and the prevalence and
results of wage escalation policies. Because it
is based on data for both large and small union­
ized and nonunionized establishments, the




manufacturing analysis can make many other
useful comparisons of its components.

Uses and Limitations

The data are used extensively by labor, man­
agement, and the Federal Mediation and Con­
ciliation Service in collective bargaining; by
private institutions and universities in studies
of industries or groups of industries; and by
local and Federal Government agencies in­
terested in the current economic picture to
determine trends in wage and benefit changes
as well as for wage, income, and gross national
product forecasts.
Since the sample is relatively small, data are
not presented for individual industries.

— G eorge R u b e n

Chapter 18. Employer Expenditures for Employee Compensation
Background

The measurement of employer expenditures
for employee compensation and the composition
of payroll hours was undertaken 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 pay­
ments were required under various social in­
surance programs, and, later, during the years
of the Second World War, employers were en­
couraged by the policies of the War Labor
Board to grant wage supplements instead of
wage increases, e.g., vacations, hospitalization.
Shortly after the war, the NLRB ruled that
pension plans were within the purview of col­
lectively bargained agreements.1 Expenditures
for these and other compensation elements, in
addition to. pay for working time, began to
comprise a substantial portion of the total com­
pensation of labor.
As early as 1875 the American Express Com­
pany instituted a private pension plan.1 In
2
1929, a private study3 indicated that there
were almost 400 such plans, and by 1968 there
were about 34,100 pension plans and 157,700
private pension or welfare plans in America.4
Paid vacations and holidays also have a rela­
tively short history for most workers. Paid
vacations were fairly well established for
salaried workers by the middle of the nine­
teenth 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, al­
though it had been customary for salaried
workers to receive pay for time not worked on
designated holidays. By 1968, approximately 7
percent of the production worker hours paid
for in manufacturing industries were leave
hours, almost all of which were vacation and
holiday hours.



The importance of recent changes in the
structure of compensation may be illustrated
by examining those that occurred for manu­
facturing production workers between 1959
and 1968. During that period pay for working
time increased from $2.23 to $3.02 an hour or
about 35 percent. At the same time, employee
expenditures for all other elements of compen­
sation increased from 38 cents to 69 cents an
hour or about 76 percent. Thus, the relative
importance of pay for working time decreased
from 85.4 percent of total compensation in
1959 to 81.8 percent in 1968.
The Bureau has for many years recognized
the necessity of studying outlays for employee
compensation. Early attempts, however, were
limited to exploratory work on methodology
and the availability of data.5 By 1959, many
of the technical and conceptual problems had
been sufficiently resolved to permit the initia­
tion of a regular program.
The first survey in the program, 1959 ex­
penditures in manufacturing, was followed by
a 1960 mining study; a 1961 finance, insurance,
and real estate survey, and another manufac­
turing industry study in 1962. The 1963 study
of expenditures for salaried (white-collar)
workers, which covered most nonagricultural
industries in the private sector, represented the
first shift in program emphasis from an indus­
try to an economy-wide orientation. Since then,
the program has been redesigned to cover all
employees in the private nonfarm sector and
to cover all significant items of employee com­
pensation.
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.
“ Latimer, Murray Webb. Industrial Pension Systems in the
United States and Canada, Industrial Relations Counselors,
Inc., New York, 1932.
4 Total active plans covering 26 employees or more for
which descriptions had been filed with the U.S. Department
of Labor by April 1, 1968 under the provisions of the Welfare
and Pension Plans Disclosure Act, as amended (P.L. 85-836 as
amended by P.L. 87-420).
5 Problems in Measurement of Expenditures on Selected
Items of Supplementary Employee Remuneration, Manu­
facturing Establishments, 1953 (BLS Bulletin 1186, 1956);
Wage Structure; Basic Iron and Steel, January 1951 (BLS
Series 2, No. 81, 1951); Wage Structure: Electric and Gas
Utilities, September 1957 (BLS Report 135, 1958).

161

162

BLS HANDBOOK OF METHODS

Description o f Survey

The survey relates to employee compensation
practices, employer expenditures arising from
these practices, and to all hours for which pay­
ment is made— hours worked, paid hours of
vacation, holiday, sick leave, and civic and
personal leave.
The program is designed to provide data
biennially for the entire private nonfarm sec­
tor; separate information is given for manu­
facturing and nonmanufacturing industries. In
the intervening years a number of individual
manufacturing and nonmanufacturing indus­
tries are studied. Both types of studies provide
data for all employees and for office and non­
office employees separately. 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
employers and the hours-paid-for during a
calendar year. Data for a lesser period of time
do not completely reflect the outlays made by
employers and the hours-paid-for. Paid leave
time, for example, usually is spread unequally
throughout the calendar year; similarly, ex­
penditures for many of the legally required
insurance programs stop after a specified
maximum amount is earned by each worker
during the year. These practices result in wide
variations between expenditures in the early
part of the year and in the latter part of the
year.
Employee compensation is increasing contin­
ually in complexity and magnitude. Practices
differ by industry group and new supplemen­
tary practices are being added. Each survey is
designed to show the individual characteristics
of compensation practices in particular indus­
tries without impairing comparability with
others. This objective is accomplished by re­
taining the same basic reference framework.
Expenditures

The expenditures studied are considered to
constitute the major elements of employee
compensation in American industry. The ex­
penditures, and therefore their measurement,



fall into two broad groups determined by the
way they are paid: payments made directly to
the workers and payments made to a third
party on behalf of the workers.
Direct payments include pay for working
hours; pay for hours not worked such as va­
cations, holidays, sick leave, and civic and
personal leave; premium pay for overtime,
weekend, holiday, and shift work; nonproduc­
tion bonuses; and severance pay. The total of
these payments constitutes the gross payroll.
Indirect payments are those made by the
employer on behalf of the worker to funds,
trustees, insurance companies or Government
agencies which may make a payment to the
worker at a later date or provide full or partial
economic security against a future contingency
(i.e. unemployment, retirement, medical ex­
penses etc.). The programs from which these
expenditures arise are either legally required
or voluntary. Legally required programs in­
clude social security, unemployment insurance,
workmen’s compensation, and State temporary
disability insurance. Voluntary plans studied
are life, accident, and health insurance; retire­
ment plans; vacation and holiday funds; severence and supplemental unemployment benefit
funds; and savings and thrift plans.
Payroll Hours

The payroll hours studied are all hours for
which the workers receive pay. These hours
consist of plant or working hours, and vaca­
tion, holiday, 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 ex­
ample, hours worked on a day that would other­
wise 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.
Conversely, some hours of leave are paid for at
less than the regular rate and only the equiva­
lent hours are counted. Each overtime hour
worked at premium rate is counted as 1 planthour.

EMPLOYER EXPENDITURES FOR EMPLOYEE COMPENSATION

Establishment Policies

Although the data on company policies are
used primarily in the review and analysis of
the expenditure and hours data collected in the
survey, these policies also have significance in
their own right, and often signal changes in
supplementary practices before actual expendi­
tures are incurred. In addition, certain charac­
teristics of American industry 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 unionization.
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
requested and approximations in these cases
may be accepted. In general, two types of ap­
proximations are used. First, if the establish­
ment records are kept for a broader grouping
of employees than are being studied, the pro­
rated share for the workers included in the
survey is computed on the basis of employment,
man-hours, or payroll, whichever is most ap­
propriate. Second, by using collateral data,
estimates are made where records are not kept
but the practice is observed. For example, the
expenditures for holiday pay may be approx­
imated by multiplying the number of hours
paid for holiday leave by average straight-time
hourly earnings. Errors occurring from the use
of these approximations would have to be in
the same direction in substantially all the cases
(overstatement or understatement of the actual
values) to have a material effect on the ac­
curacy of the results.
Data are collected primarily by mail, al­
though personal visits are made to many of the
large employers and to a sample of the estab­
lishments that have not responded to a second
mailing of the questionnaire. A questionnaire
form used in the expenditure study is repro­
duced on pages 164-167.



163

Sampling Procedure

The surveys are conducted on the basis of a
highly stratified probability sample of estab­
lishments selected by industry, location, and
employment size. The samples generally are
designed to yield reliable data for an industry
division at the national levels, 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 employ­
ment insurance laws. These lists show the em­
ployment, industry classification, and location
of all establishments 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 treat­
ment of the employment in such establish­
ments.) Some establishments in particular
industries are exempted from the UI 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 associations, 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 allocation. 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 employment size.
A subsample of establishments failing to re­
ply to the mail inquires is selected to represent
all nonrespondents, following the same general
plan as is used in the original sample. Estab­
lishments in this subsample are visited per­
sonally, instead of being solicited again by
mail.
Estimating Procedures

Data for each sample establishment are
weighted in accordance with the probability of
selection of that establishment. In the individ-

BLS HANDBOOK OF METHODS

164

U.S. DEPARTMENT OF LABOR

BLS 2868
(R e v . '69)

B udget B ureau N c. 4 4 -R 1 300
A p p ro v al ex p ires 10-31-71

BUREAU OF LABOR STATISTICS
W a s h in g t o n , D.C.

r

i

L

20212

J

L o c a tio n of u n it for w h ic h d a ta a re req u e sted .

Expenditures for Employee Compensation, 1970

G en tlem en :
The B ureau of L ab or S ta tistic s is conducting an im p ortant su rvey of how m uch
co m p an ies spend fo r em p loyee co m p en sation— fo r w ages and s a la r ie s and for le g a lly
requ ired and private p ro g ra m s w hich provide for e m p lo y e e s ' health and w e lfa r e .
The G overn m en t has u sed such in form ation fro m e a r lie r su rv e y s to aid in fo r m u ­
lating e con o m ic policy and ask s that you a s s i s t it by com p letin g this fo r m .
In form ation fr o m the su rvey w ill be of value to your com pany a ls o , since it w ill
enable you to com p are your expenditures for e m p lo yee com p en sation with those m ade
by in du stry in ge n e ra l.
K eep a copy of your rep ort— an ex tra fo rm is e n c lo se d .
W hen the B u r e a u 's re p ort on the su rvey is issu e d , we w ill send you a copy and
show you how to co m p are data fo r your com pany with the national a v e r a g e s .
Y ou r rep ort w ill be held in confidence and w ill be seen dnly by sw orn e m p lo y e e s
of the B u reau .
Nothing w ill be re le a s e d relatin g to individual co m p a n ie s.
P le a s e co m p lete the fo rm within 3 w eeks and retu rn it in the e n clo se d en velop e.
If you have questions about the in form ation req u ested , or if you need a ssista n c e
in co m p letin g the fo r m , phone the B ureau c o lle c t at a r e a code 2 02, 9 6 1 -4 0 1 9 or
9 6 1 -4 0 3 0 .
Thank you fo r your co op eration.

C o m m is s io n e r
I. Company official to contact if there are questions about this report:

N am e and t it le

D.

(P lease p rin t o r ty p e )

A rea co d e, phone no.

Units covered by this report:
Is th is q u e stio n n a ire b e in g c o m p le te d fo r th e unit(s) d e s ig n a te d ab o v e?
.
I
W hat is th e p rin c ip a l p ro d u ct, se rv ic e , or a c tiv ity of th is u n i t ? _________________________________________ __ *

□

Yes

□

No, o u r reco rd s m ak e it im p o ssib le to rep o rt se p a ra te ly for th e u n it d e sig n a te d ab o v e.
for w h ich d a ta are req u e sted a re in clu d e d in th is rep o rt.

U nits in a d d itio n to th e one

P lea se complete Item VII at end o f questionnaire to describe the units covered.

III. Average 1970 employment in units covered by this report:
P lease e n te r th e a v e ra g e n u m b er o f e m p lo y e e s in e a c h c a te g o ry d u rin g 1970.
T ypes o f e m p lo y e e s in e a c h c a te g o ry are d e scrib ed b elo w .

In clu d e f u ll - an d p a r t- tim e e m p lo y e e s.

A.

O ffic e e m p lo y e e s

B.

N o n o ffice e m p lo y e e s

C.

T o ta l 1970 a v e ra g e e m p lo y m e n t_____________________________ _

________________________________ _

OFFICE EMPLOYEES— In clu d e a ll e m p lo y e e s in e x e c u tiv e , a d m in istra tiv e , and m a n a g e m e n t positions, ab o v e th e w orking
supervisor le v e l. A lso in clu d e supervisory and nonsupervisory pro fessio n al e m p lo y e e s an d th e ir t e c h n ic a l assistants; e m p lo y e e s
en g a g ed in o ffic e c le r ic a l o p e ra tio n s; and a ll salespersons whose sales a c tiv itie s are p rim a rily p e rfo rm e d o u tsid e o f th e
e sta b lish m en t ( e . g . , r e a l e sta te sa le sm e n , and d o o r-to -d o o r sa le sm e n ).
NONOFFICE EMPLOYEES— In clu d e a ll e m p lo y e e s, e x c e p t o ffic e e m p lo y e e s as d e fin e d a b o v e , in nonsupervisory, n o nprofessional
positions. In clu d e e m p lo y e e s en g ag ed in f a b ric a tin g , processing, or assem bling; b u ild in g or e x c a v a tin g ; m in in g , d rillin g , or
p um ping; m a in ta in in g o r re p a irin g ; sh ip p in g , re c e iv in g , h an d lin g , w arehousing, p a c k in g , o r tru c k in g ; r e ta il sales; o p e ra tin g
or w orking on m o v in g v e h ic le s (buses, b o ats, e t c . ); ja n ito ria l work; guard or w a tc h m a n work; and s im ila r a c tiv itie s .
TO TA L EMPLOYMENT— Is th e sum o f o ffic e plus non o ffice e m p lo y e e s. P ro p rieto rs, m em b ers o f u n in co rp o rate d firm s, and
w orkers are n o t c o n sid e red to be e m p lo y ees and a re e x c lu d e d fro m th e survey.

 u n p a id f a m ily


EMPLOYER EXPENDITURES FOR EMPLOYEE COMPENSATION
2
Instructions for Specific Items

Item IV.
A.
G r o s s p ay roll— total of w a g e s, s a la ­
r i e s , and other paym ents m ade during 1970
b e fo re any dedu ction s. The am ount should
equal w ages rep orted on Internal Revenue
S e r v ic e F o r m s W -2 as subject to F e d e r a l
withholding ta x e s , or total rem u n eration
rep orted on IRS F o r m 9 4 0 , E m p lo y e r 's
Annual F e d e r a l U n em ploym en t Tax R eturn
1970, Schedule B , Item 1.
A -2 .
Pay for ov ertim e and w eekend and
h o l i d a y w ork— for o v e r tim e , re p ort the
s tr a ig h t-tim e pay for w ork b e y o n d the
n o rm a l w orkw eek in A - 2 a and p rem iu m
pay in A - 2 b .
F or e x a m p le :
If o v ertim e
is paid at tim e and o n e -h a lf, rep ort tw o th ird s of this pay in A - 2 a and o n e -th ird
in A - 2 b .
F o r w ork on w e e k e n d s ,
h o lid a y s, or
during vacation p eriod s— rep ort the r e g u ­
lar pay fo r w ork in A - 2 a , and the p rem iu m
pay for w ork on w eek en d s, h o lid a y s, or
during vacation p eriod s in A - 2 b . P aym en ts
m ade in lieu of tim e off for h olidays or
va ca tion s should be excluded fro m A - 2 and
included in A - 4 .
F or e x a m p le :
An e m ­
ployee w orked on a h oliday.
He re ce iv e d
h is reg u la r pay fo r w orking; plus o n e -h a lf
his re g u la r pay as a p rem iu m for having
to w ork on a h olid ay; plus his re g u la r h oli­
day pay. R eport h is reg u la r pay fo r w o rk ­
ing in A - 2 a ; the p rem iu m in A -2 b ; and
the re g u la r holiday pay in A -4 b .
A - 3 . Shift d iffe re n tia ls— total expen ditu res
fo r pay above reg u la r d a y -s h ift r a te s for
w ork on late sh ifts . Include pay for hou rs
not w ork ed .
F o r e x a m p le , . if la t e -s h ift
e m p lo y e e s w ork l 1
h hours a day but r e ­
ceive pay for 8 hours re p ort the total of
the o n e -h a lf hour paym en ts.
A-4.
Pay for lea v e — only paym ents that
w ere part of the e s ta b lis h m e n t's g r o s s
p a y ro ll.
E xclude p a y m e n t s
to fu n ds,
tr u s t e e s , e tc .
(R eport th ese in C - 3 . )
If
an em p lo ye e re c e iv e d both "p a y in lieu of
tim e o f f " and "p a y for w o r k " include only
"p a y in lieu of tim e o f f " h e re .

B.
L e g a l l y r e q u i r e d in su ran ce— net
lia b ility in cu rred during 1970 under the
p ro v isio n s of State and F e d e ra l law s for
each p ro g ra m . Exclude p aym ents m ade in
1970 for 1909, but include p aym ents m ade
in 1971 fo r 1970. Include paym ents to g o v ­
ernm ent a g e n c ie s , insurance co m p a n ie s,
and to e m p lo y e e s under s e lf-in s u r e d plans.
E xclude paym ents m ade by or withheld
fr o m e m p lo y e e s.
R e q u ire d P ay m en ts—
S o c ia l S ecu rity : In 1970 th e e m p lo y e r's p a y m e n t was 4 .8
p e rc e n t o f th e first $ 7 ,8 0 0 p a id e a c h e m p lo y e e , or a m a x i­
m u m o f $ 3 74. 40 p er e m p lo y e e .
See Form 9 4 1 —E m p lo y e r's,
Q u a rte rly F ed eral T a x R e tu rn w h ich you file d in A p ril, July,
an d O cto b er 1970 an d January 1971. R eport th e sum o f o n e h a lf th e FICA T a x re p o rte d on lin e 5 for th e 4 q u a rte rs.
U n e m p lo y m e n t in su ran ce— F ed e ra l: In 1970 th e e m p lo y e r's
p a y m e n t was 0.5 p e rc e n t o f th e first $ 3 ,0 0 0 p a id e a c h e m ­
p lo y ee or a m a x im u m o f $15.00 per em ployee. See IRS Form
940 for 1970, S ch ed u le A , Item 16. R ailro ad s in c lu d e to ta l
p a y m e n ts under R a ilro a d U n e m p lo y m e n t Insurance A ct.
S ta te : In m o st S ta te s th e p a y m e n t was a t v a ry in g rates on
th e first $ 3 ,0 0 0 p a id e a c h e m p lo y e e . See IRS Form 940 for
1970, S c h e d u le A , C o lu m n 9.

C.
P r i v a t e w e l f a r e plans— net pay­
m e n ts (after deduction of refu n d s, r e b a te s,
and dividends) m ade during 1970 by the
esta b lish m en t to funds (including u n ionm an agem en t funds), tr u s te e s , insurance
c o m p a n ie s, and paym ents m ade under the
p ro v isio n s of s e lf-in s u r e d plans to e m p lo y ­
e e s or th eir b e n e fic ia r ie s .
Include p ay­
m e n ts fo r cu rren t e m p lo y e e s , e m p lo y e es
on la y o ff, r e tire d e m p lo y e e s and their d e ­
pendents. E xclude em p loyee contributions
and a ll ad m in istra tiv e c o s ts incu rred by
the e sta b lish m en t. A ls o exclude paym ents
m ade by funds, tr u s te e s , and insurance
c a r r ie r s to your e m p lo y e e s or th eir b en ­
e fic ia r ie s .
Life, a c c id e n t, an d h e a lth in su ra n c e: L ife, a c c id e n ta l d e a th
an d d ism e m b e rm e n t, sickness an d a c c id e n t, w age a n d salary
c o n tin u a n c e in su ra n c e, an d d e a th b e n e fits; a n d h o s p ita liz a tio n ,
s u rg ic a l, m e d ic a l, d e n ta l, o p tic a l, an d drug p lan s. Exclude
ex p e n d itu re s for in -p la n t m e d ic a l ca re and v isitin g nurses or
p h y sic ia n s.
Pension an d r e tire m e n t p lans: D irec t p ay m en ts to pensioners
un d er a p a y -a s - y o u -g o p en sio n p la n , p a y m e n ts u n d er p ro fitsh arin g plans d e fe rre d u n til r e tire m e n t, an d p a y m e n ts for past
an d c u rre n t lia b ilitie s un d er fu nded p lan s.

Item V.
A - 5 . Nonproduction bon u ses— total am ount
paid for nonproduction bon u ses including
lu m p -s u m paym ents under p r o fit-s h a r in g
p la n s, and o t h e r ir re g u la r or s e a so n a l
bon u ses (such as attendance, C h r is tm a s ,
or yearen d b o n u ses). P ro c e e d s of p r o fit sharing plans which are paid into savin gs
and thrift funds or re tire m e n t plans should
be rep orted as expen ditu res for a private
w elfa re plan ( I V -C ) .
A-6.
S ev eran ce pay— total of a ll paym ents
m ade by the e stab lish m en t to e m p lo y e es
b eca u se of te m p o r a r y or perm anent s e v ­
era n ce of e m p lo ym en t. E xclude paym ents
to fu n ds, and to p en sion ers under the p r o ­
v is io n s of p a y -a s -y o u -g o pension p la n s.
R eport these in I V - C .




T otal num ber of hours paid fo r— a ll hours
w orked at straigh t tim e , a ll hours worked
at o v e r tim e , a ll paid leave h ou rs, and
hours equivalent to paym ents m ade by the
e sta b lish m en t d ire ctly to the w o rk e rs for
other hours not w orked but paid fo r .
B . N um ber of o v e r tim e h ou rs— all hours
for w ork beyond the n o rm a l workday or
w orkw eek for which s tr a ig h t-tim e or p r e ­
m iu m paym ents w ere m ade during 1970.
C . N um ber of leave hou rs— rep ort hours
equivalent to the paym ents m ade and not
the actual tim e taken off.
F o r e xam p le:
3 hours paid for at tw o -th ir d s the regu lar
rate should be rep orted as 2 h ou rs.
In­
clude leave hours fo r which paym ent was
m ade instead of tim e off.

165

166

BLS HANDBOOK OF METHODS
3
G e n e ra l In structions
If your reco rd s fo r an ite m c o m b in e d a ta for o ffic e and n o n o ffice e m p lo y e e s, p lea se p ro ra te th e c o m b in e d d a ta b e tw e e n
th e tw o e m p lo y e e groups in th e m ost a p p ro p riate m a n n e r.
If it is n o t possible to p ro rate th e c o m b in e d d a ta , e n te r
th e to ta l fig u re un d er o ffic e and in d ic a te " c o m b in e d " u n d er n o n o ffic e.
If your reco rd s c o m b in e d a ta for s e v era l ite m s,
p ro ra te th e c o m b in e d fig u re a m o n g th e ite m s to w h ic h it r e la te d in th e m ost a p p ro p ria te m an n e r O R rep o rt th e c o m ­
b in ed fig u re and c le a r ly in d ic a te to w h ic h ite m s it re la te s .
If it is n o t possibe to m a k e a n e s tim a te of ex p en d itu res
or hours for an ite m , p lea se e n te r "n o t a v a ila b le " on th e a p p ro p riate lin e .
P lease c o m p le te a ll ite m s.
If th e re w ere
no e x p e n d itu re s or hours for an ite m , e n te r "0 . "

For e a c h ite m , e n te r to ta ls for th e y e a r 1970

O ffic e

N o n o ffice

IV. Total compensation in 1970:
A.

Gross p a y r o l l -------------- -----------------------------------------------------------1.

S tra ig h t- tim e p a y --------------------------------------------------------------

2.

Pay for o v e rtim e and w e e k e n d and h o lid a y w o r k --------------a . S tra ig h t- tim e pay for o v e rtim e a n d w ee k e n d an d
h o lid a y w ork -------------------------------------------.
b . P rem iu m p a y for o v e rtim e a n d w e e k e n d and
h o lid a y w o r k ----------------------------------------------------------------

3.

S hift d i f f e r e n t i a l s ---------------------------------------------------------------

4.

Pay for l e a v e ---------------------------------------------------------------------a.

H o l i d a y s ------------------------------------------------------------------------

c.

Sick l e a v e ----------------------------------------------------------------------

d.

C iv ic and p e rso n al le a v e ----------------------------------------------

5.

N o n p ro d u c tio n b o n u s e s -------------------------------------------------------

6.
B.

S e v e ra n c e p a y --------------------------------------------------------------------

S o c ia l s e c u rity o r r a ilro a d r e t i r e m e n t --------------------------------

2.

U n e m p lo y m e n t in s u r a n c e ---------------------------------------------------a.

F e d e ra l or ra ilro a d u n e m p lo y m e n t---------------------------------

b.

S ta te -----------------------------------------------------------------------------

3.

W o rk m e n 's c o m p e n sa tio n and p ay m en ts u n d er F e d e ra l
E m p lo y e r's L ia b ility A c t ------------------------------------------------

4.

O th e r, e .g ., S ta te tem p o rary d is a b ility in su ran ce (sp e c ify ):

E m plo y er E x penditures for P riv ate W e lfa re Plans
1.

L ife, a c c id e n t, an d h e a lth in s u r a n c e ----------------------------------

2.

Pension and re tire m e n t p la n s ----------------------------------------------

3.

V a c a tio n an d h o lid a y f u n d s ------------------------------------------------

4.

S e v e ra n c e pay and s u p p le m e n ta l u n e m p lo y m en t
b e n e fit f u n d s ------------------------------------------------------------------

5.

Savings and th rift plans -----------------------------------------------------

6.

O th e r p riv a te w e lfa re plans (sp ecify ):

V. Total number of hours paid for in 1970
A.

N um b er of s tr a ig h t- tim e hours w orked —

B.

N um ber o f o v e rtim e hours w orked --------

C.

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

E m plo y er E xp en d itu res for L eg ally R e q u ire d Insurance
1.

C.

■■■■■■■

V a c a tio n s ----------------------------------------------------------------------

b.

wmmmmmmmm

N u m b e r of le a v e hours p a id f o r ------------1.

V a c a ti o n --------------------------------------------

2.

H o l i d a y ----------------------------------------------

3.

Sick le a v e -----------------------------------------

4. C iv ic



and p e rso n al le a v e

Hours

Hours

167

EMPLOYER EXPENDITURES FOR EMPLOYEE COMPENSATION

VI. Establishment practices and policies:
A.

Paid v a c a tio n s. R ep o rt th e n u m b er of e m p lo y e e s who r e c e iv e d v a c a tio n pay d u rin g 1970 d ire c tly fro m th e e sta b lish m en t
ac co rd in g to th e am o u n t o f pay.
N u m b er of e m p lo y e e s r e c e iv in g —
Em ployees

U nder
1 w e e k 's
pay

No
v a c a tio n
pay

1 and under
2 w eek s'
pay

2 and un d er
3 w eek s'
pay

3 and u nder
4 w eek s'
pay

4 and un d er
5 w eek s'
pay

5 w eek s'
pay or
m ore

O f f i c e ---------------------------N o n o f f ic e ---------------------B.

Paid h olidays. Enter the n u m b er o f days per e m p lo y e e . If m ore th a n o n e p ra c tic e e x iste d for an e m p lo y e e group, rep o rt
th a t w h ic h a p p lie d to th e g re a te st n u m b er in th e group. If th e g re a te st n u m b er o f e m p lo y e e s re c e iv e d no p a id h o lid ay s,
e n te r "0 . "
Fu ll days
H alf days
O ffic e e m p lo y e e s ---- -----N o n o ffice e m p lo y e e s------

C.

Sick le a v e . Did th e e s ta b lish m en t h a v e a p ra c tic e or a p o lic y (ev e n th o u g h th e re m ay h av e b e e n
no ex p e n d itu re ) for providing p a id sick le a v e for a n y —
Yes

D.

C iv ic and perso n al le a v e .
D id th e e sta b lish m en t hav e a p ra c tic e o r a p o lic y (e v e n th o u g h th e re
m ay h av e b e e n no ex p e n d itu re ) for p ro v id in g p a id c iv ic le a v e (m ilita r y , ju ry , w itn ess, v o tin g ,
e tc . ) o r personal le a v e (su ch as for d e a th in fa m ily ) for any —
O ffic e e m p lo y e e s --------------------------------------------------------------------------------------------------------------------------N o n o ffice e m p lo y e e s - -------------------------------------------------------------------------------------------------------------------

E.

No

B B

O ffic e e m p lo y e e s -------------------------------------------------------------------------------------------------------------------N o n o ffice e m p lo y e e s --------------------------------------------------------------------------------------------------------------

B B

L ife, a c c id e n t, and h e a lth In su ran ce.
O ffic e e m p lo y e e s -------N o n o ffice e m p lo y e e s—

Did the e stab lish m en t finance any of the follow ing insurance p lan s for—
L ife
S ic k n ess and accid e n t
H o sp ita liz a tio n or m edical
Y es
No | |
Yes
No | |
Y es
No Q“]
Yes
No
Y es
No
Y es f~^] No

D id em p lo y e e s pay for p art of an y of th e se in su ran ce plans (answ er NO if payment was only for additional
b e n e fits or c o v e ra g e for d e p e n d en ts)
L ife

O ffic e e m p lo y e e s--------- Y es Q j
N o n o ffice e m p lo y e e s--- Y es
F.

Pension and r e tire m e n t plans.

S i c k n e s s a n d a c c id e n t

No □
No QJ]

Y es L i ]
Y es [ ^ ]

H o s p it a l iz a t i o n or m e d ic a l

Y es C7]
Y es Q- ]

N o C71

No

No |— |
No Q- ]

D id th e e sta b lish m en t fin a n c e such a p lan fo r—

O ffic e e m p lo y e e s ------------------------------------------------------------------------------------------------------------------N o n o ffice e m p lo y e e s --------------------------------------------------------------------------------------------------------------

B B

D id e m p lo y e e s p ay fo r p art of any of th ese plans (answ er NO if p a y m e n t w as o n ly for a d d itio n a l
be n e fits)
O f f i c e e m p l o y e e s ------------------------------------------------------------------------------------------------------------------------------

N o n o ffice e m p lo y e e s -------------------------------------------------------------------------------------------------------------G.

C o lle c tiv e b a rg a in in g .

O ffic e e m p lo y e e s-------------------------------------------------------------------------------------------------------------------N o n o ffice e m p lo y e e s -------------------------------------------------------------------------------------------------------------H.

B B

Did u n io n -m a n a g e m e n t a g re e m e n ts co v er a m a jo rity of th e —

B B

R e g u la r w ork w eek , 1970.
How m an y hours ( e .g ., 44, 40, 3 7 .5 , e t c . ) w ere n o rm a lly w orked e a c h
w eek by th e m a jo rity of th e —
O ffic e e m p lo y e e s------------------------------------------------------------------------------------------------------------------N o n o ffice e m p lo y e e s --------------------------------------------------------------------------------------------------------------

Hours p er w eek
Hours p er w eek

VII. Units included in report ( if d iffe re n t from th a t req u e sted in address b o x ):
If this rep o rt re la te s to units in a d d itio n to th e o n e d e sig n a te d a t th e to p of p ag e 1, p le a se p rovide th e fo llo w in g in fo rm a tio n
for e a c h u n it in clu d e d in the rep o rt.




A v erag e 1970 e m p lo y m e n t
L o catio n
O ffic e

N o n o ffice

R rincipal product,
s e rv ic e , o r a c tiv ity

168

BLS HANDBOOK OF METHODS

ual industry studies the selection is based on
establishment size strata. Thus, a reporting
unit which is in a stratum at which the selection
probability was set, for example, 1 out of 5
establishments, will be given a weight of 5,
representing itself and four other establish­
ments in this same stratum. In the biennial
studies of the entire private nonfarm sector,
the probability of selection is proportionate to
establishment employment size. Thus, a report­
ing unit employing 1000 workers, in a sample
where the employment size probability base
was set at 10,000, will be given a weight of 10,
representing itself and other establishments
having an aggregate employment of 10,000
workers. Under both procedures all establish­
ments over a certain size are included.
The sample of nonrespondents for which
data are collected by Bureau field representa­
tives is weighted appropriately to represent all
nonrespondents.
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 certainty, its weight is assigned to units
in the sample with the most similar industrysize-location characteristics.
All estimated totals derived from such
weighting procedures are adjusted further by
the level of total employment or other appro­
priate measure for the survey year, based on
data from the Bureau’s monthly establishment
employment statistics program, in each of four
broad economic regions. For instance, if the
level of the aggregates, as derived from the
weighting procedures, is 40,000,000 in an in­
dustry-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, including those
having fewer than four employees, in the in­
dustries studied.6
Some improvisation is necessary in the con­
struction of such annual benchmark totals. The
monthly employment series provides data for
only one pay period each month, and the esti­
mate of annual totals is made by multiplying



by the average 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 employment (or hours) estimates into
regional aggregates. Such sources include the
Census of Manufactures and County Business
Patterns (based on Social Security establish­
ment data.)7
Presentation

The expenditure data on the individual ele­
ments of compensation are combined to give a
measure of total employee compensation. The
expenditure data for each individual element
and for groups of elements are presented as a
percent of total compensation, in cents per paid
hour, and in cents per working hour. These
measures are shown for all establishments, as
well as for only those establishments that had
an actual expenditure for a particular practice
during the reference year. Hours data for
working hours and for paid leave hours are
presented as a percent of all paid hours.
A. The expenditure ratios are calculated as follows:
1. Expenditures as a percent of total compensa­
tion for all establishments^:
Aggregate expenditures for the practice

-------------- ~ ~ ~Aggregate compensation in all
a ~ --------r----------------n--------------------- X
100
establishments
2. Expenditures as a percent of total compensa­
tion for establishments reporting expendi­
tures =
Aggregate expenditures for the practice ^
Aggregate compensation in
establishments reporting
expenditures for the practice.
B. The expenditure rates are calculated as follows:
1. Expenditures in cents per paid hour for all
establishments^:
Aggregate expenditures for the practice
Aggregate paid hours
2. Expenditures in cents per hour of work for all
establishments^
Aggregate expenditures for the practice
Aggregate paid hours minus
aggregate paid leave hours
8 See chapter 2, “Employment, Hours and Earnings” .
7 U.S. Department of Commerce, Bureau of the Census,
County Business Patterns (various years).

EMPLOYER EXPENDITURES FOR EMPLOYEE COMPENSATION
3. Expenditures in cents per paid hour for es­
tablishments reporting expenditures=
Aggregate expenditures for the practice
Aggregate paid hours in establishments
reporting expenditures for the practice.
4. Expenditures in cents per paid working hour
for establishments reporting expenditures=
Aggregate expenditures for the practice
Aggregate paid hours minus paid
aggregate leave hours in establish­
ments reporting expenditures for the
practice.

The distribution of workers by establishment
expenditure ratios and rates is published, as
well as the averages of ratios and rates. Ex­
penditures also are shown by selected es­
tablishment characteristics such as size,
compensation level, unionization, and, for cer­
tain surveys, by area.
Analysis

The expenditure data describing the payroll
or nonpayroll elements of compensation are
presented in summary by this characteristic.
However, the analysis of the data is related
to the benefit function of each element. Thus,
for analytical purposes, elements of compen­
sation that provide similar or interchangeable
benefits are grouped together. The following
groups of compensation elements are studied:1
1. Pay for working time; straight time pay, and
premiums for overtime, weekend, holiday, and
shift work.
2.

Pay for leave time; vacations, holidays, miscel­
laneous leave of absence, and payments to vaca­
tion and holiday funds.

3. Payments for retirement programs; social se­
curity and private retirement plans.
4.

Payments for health and related programs; life,
accident, and health insurance, sick leave, and
workmen’s compensation.

5.

Payments for unemployment benefit programs;
unemployment insurance, severance pay, and
severance pay funds and supplemental unem­
ployment benefit funds.

6. Nonproduction bonuses.
7.

Savings and thrift plans.

Data are presented on the importance of
various types of paid hours relative to all paid
hours. Information is also published on the



169

number of paid holidays and number of weeks
of paid vacation received by workers.
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 sim­
ilar or dissimilar characteristics (industry,
size, location, union status, and average earn­
ings levels of workers). Labor and manage­
ment use the data in collective bargaining;
and Government uses the statistics in the
formulation of public policy, in producing esti­
mates of industry output per man-hour, 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 re­
ceived by workers.
As indicated earlier, the expenditures studied
comprise the significant elements of employee
compensation in American industry. The ag­
gregate of the expenditures studied represents
total employee compensation. It does not, how­
ever, represent total labor cost which is a more
encompassing concept and includes factors such
as the cost of recruiting and training labor, the
administrative expenses incurred in admin­
istering benefit programs, and many other ex­
penditures resulting from the use of labor as
a factor of production. Some of these expendi­
tures may be important in particular estab­
lishments.
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 re­
porting errors, to have a material effect on the
accuracy of the results, would have to be in the
same direction in substantially all of the cases.
The omission of establishments 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
ments.

170

BLS HANDBOOK OF METHODS

Technical References
Number

1. U.S. Department of Labor, Bureau of Labor Statistics. Problems in Measurement of Expend­
itures on Selected Items of Supplementary Employee Remuneration, Manufacturing Estab­
lishments, 1953 (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.
2 . ______Employer Expenditures for Selected Supplementary Remuneration Practices for Pro­
duction Workers in Manufacturing Industries, 1959 (Bulletin 1308,1962).
3 . ______Production Workers in Mining Industries, 1960 (Bulletin 1332, 1963).
4 . ______Employer Expenditures for Selected Supplementary Remuneration Practices in F i­
nance, Insurance, and Real Estate Industries, 1961 (Bulletin 1419, 1964).
5 . ______Employee Compensation in the Private Nonfarm Economy, 1966 (Bulletin 1627, 1969).
6 . ______ 1967 Industry studies titled “ Employee Compensation and Payroll Hours.” Seven re­
ports (1 969); Banks (3 6 2 ); Research Laboratories (3 6 3 ); Confectionery (3 6 4 ); Struc­
tural Steel (3 6 5 ); Hotels & Motels (366) ; Laundries. Cleaning & Dyeing Plants (3 6 7 );
Men’s & Boys Shirts (368).
7 . ______Employee Compensation in the Private, Nonfarm Economy, 1968 to be published in
1971.
Each expenditure bulletin contains descriptive information on the detailed procedures
and techniques used in the study.
8 . ______Bauman, Alvin, “ Measuring Employee Compensation in U.S. Industry,” Monthly La­
bor Review, October 1970, pp. 17-23.




A

l v in

B

a u m a n

Chapter 19. Work Stoppages
Background

Work stoppage statistics are compiled by
the Bureau of Labor Statistics to provide a
quantitative measure of the extent to which
disputes between labor and management result
in strikes or lockouts and of the immediate
economic disruption resulting from such stop­
pages.1 When considered along with general
economic measures, these statistics also serve
at times as a broad indicator of the state of
industrial unrest.
The first attempt by any Federal agency to
compile statistics on strikes was made in
1880,1 when the Bureau of the Census sent
2
questionnaires to employers and workers in­
volved in all disputes which were noted in the
public press during the year. Information was
received on 762 situations. Some data were
obtained on the causes of strikes and their re­
sults, but not on the number of workers in­
volved or resultant man-days of idleness.
The next collection of strike statistics was
undertaken in 1887, when the Bureau of Labor,
then in the Department of the Interior, ex­
amined files of newspapers, trade journals, and
commercial periodicals 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 organi­
zations.
No Federal agency collected national infor­
mation on stoppages occurring during the
1906-13 period. In 1914, relying exclusively
on printed sources, the Bureau of Labor Sta­
tistics attempted to compile a record of all
strikes and lockouts during the year. In the
following year, the Bureau inaugurated a
method for the collection of strike and lockout



material which has been followed, with modifi­
cations, since that time. Briefly stated, the
procedure was to send questionnaires 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
particular the procurement of data on the
number of workers involved in all stoppages
and the computation of man-days of idleness,
inaugurated the modern series of monthly and
annual strike data.3
Description o f the Series

The present series on work stoppages covers
all strikes and lockouts known to the Bureau of
Labor Statistics and its cooperating agencies.
It covers all that 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 efficiency, and, in part,
because of the difficulty involved in defining,
identifying, and securing information on
strikes that last a few hours or less.
The Bureau defines a strike as a temporary
stoppage of work by a group of employees to
express a grievance or enforce a demand. A
lockout is defined as a temporary withholding
of work by an employer (or a group of employ­
ers) to enforce terms of employment on 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.”
1 Throughout this chapter, the terms “work stoppage” and
“ strike” are used interchangeably; both terms, unless other­
wise noted, also include lockouts. The definitions, terms, and
classifications used by the Bureau in compiling work stop­
page data were adopted for statistical and research purposes
and have no legal significance.
2 On the State level, the Bureau of Labor Statistics of Mas­
sachusetts, issued a report in 1880 on strikes in that State
from 1825. In 1881, the Bureau of Industrial Statistics of
Pennsylvania issued a report on strikes in that State from
1835.
3 For additional information on the early history of the
work stoppage statistics program, see BLS Bulletin 651,
Strikes in the United States, 1880 to 1936 (1938).

171

172

BLS HANDBOOK OF METHODS

Although an employer-employee dispute is
implicit in these definitions, some inclusions in
the Bureau’s series relate only indirectly to this
concept. For example, jurisdictional strikes and
rival union disputes between two unions or
more often have the employer on the sidelines.
In a sympathy strike, the issue of the stoppage
does not usually involve the immediate em­
ployer. Moreover, protest strikes against the
actions of governmental agencies are not the
result of a dispute 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 American seamen or other workers in for­
eign ports and strikes of foreign crews in
American ports. Also excluded are so-called
slowdowns, where employees continue at work
but at deliberately reduced 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 clos­
ing time to attend rallies or mass meetings.
The number of work stoppages occurring
during a given period provides a measure of
the frequency of disputes; the severity and
effect of such actions are measured by the
number of workers involved, duration, and the
resultant man-days of idleness. The basic sta­
tistical unit in the Bureau’s series is the indi­
vidual strike or lockout. If groups of employees
(regardless of their number and type and loca­
tion 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
involved in a strike or lockout is the maximum
number actually made idle in the establishment
or establishments directly involved. No dis­
tinction is made in arriving at this figure be­
tween 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 operation.
Man-days of idleness, like the number of
workers involved, are based on the idleness at
the establishment or establishments directly
involved. Workers involved multiplied by work­



days lost equal total man-days idle. In instances
where the number of workers idle varies dur­
ing the period of the stoppage, appropriate
adjustments are made in this calculation. Al­
lowance is made in these computations for
holidays and days not normally worked.
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 ob­
tain the necessary facts regarding each situa­
tion as quickly as possible.
Information about the existence of stoppages
currently is obtained from various sources, in­
cluding: (1) clippings from daily and weekly
newspapers throughout the country provided
by commercial clipping services; (2) notices
received directly from the Federal Mediation
and Conciliation Service; (3) a periodic com­
pilation by the local offices of the State employ­
ment security agencies, provided through the
U.S. Training and Employment Service 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) in­
ternational unions and their publications; (7)
firms under contract to the Atomic Energy
Commission; and (8) other Federal agencies
and commissions.
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 num­
ber of stoppages reported, but has little effect
on the total number of workers and man-days
of idleness, because the added stoppages tend
to be small.
After the receipt of notices regarding the
existence of work stoppages, the Bureau mails
questionnaires to the parties involved to secure
direct information on each stoppage. Should
a reply not be received within 3 weeks, a sec­
ond questionnaire is mailed, and, in the case of
continued nonresponse, a mailogram or tele­
gram may be sent, or an effort made to secure
the necessary data by telephone. In some in­

173

WORK STOPPAGES

stances of nonresponse, field representatives of
the Bureau secure the necessary data; in
others, cooperating State agencies may be
asked to contact the parties.
The types of information sought by the Bu­
reau through its questionnaire have changed
over the years, partly in response to changing
needs. The primary function of these reports
is to compile statistics, not to keep records on
the strike activity of individual firms and
unions. The separate questionnaires currently
used for private and public sector disputes are
shown on pages 174-176.4
Although strikes, by their very nature, are
usually matters of public knowledge and of
reporting by newspapers and other publica­
tions, the Bureau holds confidential the indi­
vidual reports submitted by employers and
unions, as well as supplementary data collected
through State or Federal agencies. The rules
of confidentiality observed here are similar to
those followed in other Bureau surveys.
Estimating Procedures

Since the Bureau is able to obtain informa­
tion 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 reason­
ably accurate data on the larger stoppages at
the time these estimates are prepared— ap­
proximately 4 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 num­
ber of stoppages, the number of workers in­
volved, and man-days of idleness. As there is a
lag between the occurrence and reporting of a
number of relatively small strikes, the number
of stoppages beginning during a given month
is estimated by increasing 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 ob­
tained by supplementing the latter estimate by



a percentage of the stoppages in effect during
the prior month.
In estimating the number of workers in­
volved and total idleness, efforts are made to
obtain as much preliminary information as pos­
sible 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 number of workers (or
man-days) that previous experience indicates
for such stoppages.
In its preliminary reports, as well as in its
final reports, the Bureau relates the man-days
of idleness to the total estimated working time
of all workers. The “ total economy” measure
of strike idleness, which was instituted in
1967, includes government and agricultural
employees as well as private nonfarm workers
in its employment count as well as in the
computation of idleness ratios. Before 1967, the
BLS series excluded government and agricul­
tural workers from employment totals, but ac­
counted for time lost by these workers while
on strike. This reevaluation of methods has
improved the calculations of idleness and made
the Bureau’s measurement of work stoppage
intensity national in scope.
Analysis and Interpretation

The data presented in the parties’ reports
are analyzed and classified according to a num­
ber of significant factors, briefly described
here:
(1) Each strike is assigned an industrial
classification in accordance with the Standard
Industrial Classification Manual prepared by
the Bureau of the Budget.5 In those cases in
which a stoppage affects workers in more than
one industry, one of two procedures may be
followed. If the stoppage is small, the strike
is classified in the industry in which it was
initiated; in large interindustry stoppages, a
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 question­
naire is sent to the parties periodically to determine the
status of the stoppage.
5 See appendix B.

174

WORK STOPPAGES

Form A pproved
B udget Bureau N o. 4 4-R 2 1 0

B L S 817
(R e v . 1963)

U. S. DEPARTMENT OF LABOR
BUREAU OF LABOR STATISTICS
W A S H I N G T O N , D . C . 20212

WORK STOPPAGE REPORT
Your report w ill b e held
in c o n fid e n ce
P le a s e answ er a ll q u estion s
T h is req u est for inform ation re la te s to:

r

L
i . E m ployer

Name and m ailing address

2 E s ta b lis h ­
ments
in v o lv e d in
stop p a ge

Number o f esta b lish m en ts d irectly in v olv e d or in w hich workers ob se rv e d p ick e t lin e s
If m ore than one esta b lish m en t, u se rev erse s id e ; if one, enter b e lo w ;
a. L o c a t i o n _______ __ ________________________________________________________ _________ _
b . Industry
(In d ica te m ajor type o f a ctiv ity and a ls o p rin cip a l products or s e r v ic e s , e .g ., M ining-bitum inous c o a l; C o n ­
stru ctio n -h ig h w a y s and stre e ts ; M a n u factu rin g-w ood en up h olstered furniture; W holesale trade - plum bing
s u p p lie s ; T ra n sp orta tion -m otor freig h t.)
Name

3. Union
in v olv ed

AFL
CIO
A d dress

L o c a l N o.

l D

Other

4. D ates o f
stop p a ge

S toppage began on

5. Number o f
workers
a ffe cte d

E m p loy ees returned to
work on

T O T A L workers id le d at lea st one full sh ift or day

Settlem ent was rea ch ed on

□

Settlem ent was ra tifie d on
( i f a p p lica b le)

workers

(IMPORTANT —
Include a ll w orkers d irectly in v o lv e d in the sto p p a g e and w orkers made id le by la ck o f work in the
same esta b lish m en ts or by o b se rv a n ce o f p ick e t lin e s . If exa ct figu res are not a v a ila b le , p le a s e p rovid e e s tim a te .)
D id the number id le ch an ge s ig n ifica n tly during the sto p p a g e ? Y e s Q
No Q
(If “ y e s , ” p le a s e en ter ch a n g es in number id le and d a tes o f ch a n g es on r e v e r s e s id e o f s h e e t .)
........................... ............................................. d a ys

6. Normal w orkw eek prior to s t o p p a g e .....................................................................................................
7. C ontract status (c h e c k o n e )
Stoppage occu rred
Q In n eg otia tion o f first con tract or in obtaining union re co gn itio n
Q I n ren egotiatin g con tract terms (expiration or reop en in g)
Q During term o f con tract (ch an ge in con tract terms not in v o lv e d )
□ Other ( s p e c i f y )

.

.......

- ...... - —

■■

8 . MAJOR is s u e s in d isp u te (lis t in order o f im portance)

9. D id the agreem ent to return to work in clu d e a p rocedu re for handling any u n se ttle d m ajor is s u e s in v o lv e d in the sto p p a ge ( e .g .,
by subm itting is s u e s to a rbitration)?

Y es

Q

No Q

If “ y e s ” , in d ica te the is s u e s and the p rocedu re agreed upon.

10. D id a F ed era l, State, or lo c a l governm ent a g e n c y , or a private m ediator, m ediate in this d isp u te or a s s is t in arranging the return
to work? (C h eck more than one if a p p lic a b le ).

F ed era l Q

State Q

L ocal

Q

P riv ate

Q

N one Q

P le a s e id entify a s s is t in g governm ent a g en cy , if any

(D ate)




(Signature and title o f person making report)

(Com pany or organ ization )
U se re v e rse s id e for any
cla rify in g rem arks.

BLS HANDBOOK OF METHODS

175

Supplementary Information for Items 2 and 5:

If the stoppage involved more than one establishment or if idleness varied from period to period during
the stoppage, please use the following space to indicate the number idle in each establishment and the
variation in idleness at different dates. Include both workers directly concerned and those made idle
because of dispute in the same establishment.
IF E X A C T FIGURES A R E NOT A V A IL A B L E , P L E A S E FU R N ISH ESTIM A TE S.

Industry or
principal
product

Establishment involved and location
(City, County, State)

Approximate number Dates this number
was idle
of workers idle
a full shift or more a full shift or more

'
REM
ARKS:




* U. S. GOVERNM ENT

PR INTING O F F IC E :

1 9 6 9 - 3 6 1 - 7 14

176

WORK STOPPAGES

Budget Bureau No. 44-R 1397
A pproval expires 6 /3 0 /7 1

BLS 3006

U.S. DEPARTMENT OF LABOR

WORK STOPPAGES REPORT

BUREAU OF LABOR STATISTICS

Government

W a sh in gton ,

D.C.

20212

No. _____________

T h i s r e q u e s t L r in f o r m a t io n r e l a t e s t o :

1. Government Department, Agency, or Installation
N am e:

Address:

F a c ility w here stoppage occu rred (n am e if d ifferen t from above):*

Level (c h e c k one):

[^ F e d e ra l
j | S tate

Function (check one):
□ A d m in istrativ e services
□ W elfare services
□ Law en fo rcem en t and c o rrectio n
1 | Fire pro tectio n

□
□

□
□
P
□

Address:

M u n ic ip a lity
C ounty

□

S a n ita tio n services
Education
Streets and highways
Parks, re c re a tio n , lib ra rie s, e tc .

School d istric t
O ther (specify) -------------------------------

□ H ospitals and h e a lth services
□ T ran sp o rtatio n and a llie d fa c ilitie s
I j O ther u tilitie s
j | O ther ( s p e c i f y ) -------------------- ---------

2. Union or Association
A ffiliatio n :
□ AFL-CIO
1 | O ther union
I I E m ployee asso ciatio n
3. Dates of Stoppage

D id th e organization
c a ll or support th e
work stoppa ge ?
□ Yes
□ No
□ No in fo rm atio n

Does this org an izatio n
h av e o ffic ia l re c o g n i­
tio n ?
□

Yes

□

No

and Workweek

Stoppage
b egan on:

S ettle m e n t
reach ed on:

S e ttle m e n t
ra tifie d on:

Em ployees retu rn ed
to work on:

Sch ed u led w orkw eek prior to
stoppage—
Days

Hours

4. Employees Affected
T o ta l em ployees id led a t lea st one fu ll
shift or day

(IM PORTANT - In clu d e a ll em ployees d ire c tly involved in th e stoppage and em ployees
m ad e id le by la c k of work in th e sam e fa c ilitie s or by observ an ce of p ick e t lin es. If
f r
t'
•/

D id the num ber id led ch ange sig n ific a n tly during the stoppage?
\ j Yes
| [ No
(If "yes" please e n te r changes in num ber id le and d ates o f changes on reverse side of this f o r m .)
O c c u p a tio n a l c la ssifica tio n (ch e c k one or m ore):
| | T eachers
| | Nurses
□ O ther professional and te c h n ic a l em p lo y ees
□ C le ric a l

□
P o lic e m en
□ F ire m e n
Q ] S a n ita tio n m en
| j C raftsm en (specify)

I I O ther blu e c o lla r and m anual
□ O ther (specify)

5. Agreement Information
Stoppage occurred (ch e c k on e):
□
□

In a tte m p tin g to o b tain
reco g n itio n
In ne g o tia tin g first
ag re e m en t

□
□

D uring a g re e m en t term
(ch an g e in term s not involved)
In ren e g o tia tin g ag re e m en t
(ex p iratio n or reopening)

□
□

No fo rm al a g re e m en t involved
O th er (specify)

M ajor issues in dispute in order of im p o rtan ce:

Did em ployees retu rn to work Q ] v o lu n ta rily , or Q

under th e term s of a court o rd er or in ju n ctio n ?

Did ag re e m en t to retu rn to work in clu d e a procedure for han d lin g any u n settled m ajo r issues involved in th e stoppage (e . g . , by
su b m itta l to a rb itra tio n or factfin d in g )?
| | Yes
Q No
If yes, note issues and procedures agreed upon on reverse side of this form .
Did a governm ent ag en cy , or p riv ate in d iv id u al, or org an izatio n assist in arran g in g th e retu rn to work?
(C heck one or m ore):
F ed eral
| ) S tate
Local
j j Private
Please id en tify gov ern m en t agency
Signature of person m ak in g report:

T itle :

D ep artm en t or organization:

* If m ore th an one fa c ility was in v o lv ed , p lease en ter in fo rm atio n on reverse side of this form .
c la rify in g rem a rk s, p a rtic u la rly on n atu re of stoppage (mass sick le a v e , or resig n atio n s, e tc ).




I | None
D ate:

Also use reverse side for

177

WORK STOPPAGES

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 be­
ginning to end. For stoppages which begin at
a definite time and are terminated by a formal
agreement, no problem arises in the determina­
tion of duration. However, some stoppages, for
a variety of reasons, are never settled formally.
These range from situations in which the
workers gradually 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 details of each individual
situation are studied before a stoppage is
terminated for statistical purposes; in the lat­
ter 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 rec­
ord 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
workplace, 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 employees 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 superseded city boundary lines.6 In
interarea stoppages, a stoppage is recorded in
each area affected, and workers and man-days
of idleness are allocated proportionately.
(5) The issues in dispute in most strikes are
many and varied, and do not always lend them­
selves readily to immediate and exact classifica­
tion. Stoppages are classified by major issue
into the following broad groupings: (a) wages,
hours, and supplementary benefits; (b) union
organization and security; (c) job security;
(d) plant administration; and (e) inter- or




intra-union matters. Each of these groups is
sub-divided into more specific categories.7
(6) Stoppages are classified by the con­
tractual relationship existing between the
parties involved. The following four situations
apply: (a) negotiation of the initial agreement;
(b) renegotiation of an agreement; (c) agree­
ment in effect (new contract terms not in­
volved) ; and (d) no contractual relationship.
(7) The union involved is another major
classification. For this purpose, the union is the
organization whose contract was involved or
which has taken active leadership in the stop­
page. Disputes involving more than one union
are classified as jurisdictional or rival union
disputes or as involving cooperating unions. If
unorganized workers strike, a separate classi­
fication is used. For publication purposes,
union information is presented by major affilia­
tion of the union, i.e., AFL-CIO, or non­
affiliation such as “ Independent,” “ single firm,”
or “ no union.”
(8) The assistance of mediators, either gov­
ernmental or private, in the resolution of in­
dustrial disputes is recorded.
(9) The manner in which stoppages are
settled involves classification into the following
categories: (a) those ending with a formal
settlement; (b) those terminating without a
formal settlement, and those 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 after the resumption
of normal operations.
Presentation

Publications in the area of work stoppages
include monthly preliminary estimates, annual
8 See appendix C.
7 When a major change in the classification of issues was
introduced in 1961, the Bureau included in its annual report
for that year a method of linking the new classifications with
the old.

178

BLS HANDBOOK OF METHODS

reports, and special reports which are issued
irregularly.
Monthly preliminary estimates are issued
approximately 30 days after the end of the
month of reference in the form of press re­
leases. Such reports presently are prepared for
the first 10 months of each year, with a pre­
liminary estimate of total strike activity dur­
ing 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 is­
sued in June. An annual bulletin,8 containing
detailed information on the characteristics of
work stoppages during the prior year, is pub­
lished each winter.
Special reports containing historical work
stoppage data by industry and area, or an
analysis of a particular aspect of strikes, are
issued irregularly. This latter category also
may include Bureau reports of a nonstatistical
nature, including 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 rea­
son for their compilation in the United States
and in other industrialized countries.9 Whether
they serve this elementary purpose today is
open to question.1 In any event, some qualifica­
0
tions must be taken into consideration in this
use. The willingness of workers to strike as a
protest against existing conditions may be en­
couraged or deterred by outside influences, such
as the employment situation, 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, em­
ployers, and employer associations use strike
data to assess their own experiences, and busi­



ness and civic organizations are concerned with
their community promotional possibilities.
Schools, particularly 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 newsworthy.
Although it is virtually certain that the Bu­
reau is able to locate, and obtain information
on, the larger work stoppages, some small
strikes undoubtedly escape notice each year.
While these omissions do affect data on number
of strikes, the statistics on workers and mandays of idleness are virtually complete. As has
been noted, the addition of new sources of in­
formation has not materially 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 narrower the
classification of strike data, the greater is the
chance of a significant omission. For example,
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 ommission 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 result­
ing from a work stoppage. At times, the idle­
ness of employees directly involved in a strike
may be considerably less than the idleness of
other workers brought about indirectly. No
satisfactory method, however, 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 pro­
duction and wages lost. The calculation of cost
8 Since 1949, these bulletins have been titled Analysis of
Work Stoppages (year).
9 For a discussion of the methods used in compiling strike
statistics in other countries, see Methods of Compiling Statis­
tics 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 In­
dustrial Conflict, New York, John Wiley & Sons (1960).
Data on work stoppages in other countries are published
annually in the Year Book of Labour Statistics, Geneva, In­
ternational Labour Office.
1 See article by Joseph W . Bloch, “ The Strike and Discon­
0
tent,” Monthly Labor Review, June 1963, pp. 645-651.

179

WORK STOPPAGES

involves many complex and interrelated factors
for which information is not readily available,
including such matters as production schedules
before and after the stoppage, diversion of out­
put or services to other plants or employers,
the flow of raw materials, and the amount of

overtime worked before and after the strike.1
1
The problem is magnified beyond statistical
control if secondary costs are to be accounted
for.
1
1 A framework for the measurement of the cost of strikes
is developed by Neil W. Chamberlain and Jane Metzger
Schilling in The Impact of Strikes: Their Social and Economic
Costs, New York, Harper & Bros. (1954).

Technical References
Number

1. International Labour Office, Methods of Compiling Statistics of Industrial Disputes, Studies
and Reports, Series N (Statistics), No. 9, Geneva (1926).
Compares the methods used to compile statistics on industrial disputes, and outlines
standards by which some degree of international comparability may be secured.
2. Peterson, Florence, “ Methods Used in Strike Statistics,” Journal of the American Statistical
Association, 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, Analysis of Work Stoppages (year).
The Bureau of Labor Statistics issues annually a detailed statistical bulletin analyzing
work stoppages in the United States.
4 . ______Strikes in the United States, 1880 to 1936 (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. Onanian




Chapter 20. Collective Bargaining Agreements
Background

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,
supplementary benefits, job and union security,
the timing of wage negotiations, the nature of
plant operations and working conditions, and
many of the day-to-day aspects of employeremployee and union-management 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 re­
produce representative agreement provisions
or the variety of provisions relating to similar
problems, or digests of selected identified
agreements or benefit plans; and (3) by pre­
paring 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
bargaining.
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 almost 70 years.1
Systematic efforts to collect agreements began
in 1912. The first of a number of BLS bulletins
devoted entirely to the subject of collective
bargaining agreements appeared in 1925. A
large number of reports and bulletins on a wide
variety of industrial relations subjects have
since been published.
The Bureau’s responsibility in the field of
agreement 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
interested representatives of employers, employees, and
the general public, the Bureau of Labor Statistics of
the Department 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 inspection under appropriate conditions pre­
scribed by the Secretary of Labor, except that no
specific information submitted in confidence shall be
disclosed.
(b) The Bureau of Labor Statistics in the Depart­
ment of Labor is authorized to furnish upon request of
the [Federal Mediation and Conciliation] Service, or
employers, employees, or their representatives, all
available data and factual information which may aid
in the settlement of any labor dispute, except that no
specific information submitted in confidence shall be
disclosed.

Concepts and Scope

Although the substance of collective bargain­
ing rests partly upon a foundation of unwritten
industry, company, and union practices, and
upon various legal requirements, the basic unit
in agreement collection and analysis is the writ­
ten 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 bargaining purposes. It may express
conditions of employment in simple terms, and
leave 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. They reflect the diver­
sity of employment conditions among indus­
tries and companies and of the scope of the
issues over which bargaining takes place, as
well as differences in the degree of precision
sought and the language used.
Estimates of the number of agreements cur­
rently in effect range up to 155,000. The num­
ber of workers covered by agreements is
estimated at about 20 million. The Bureau
presently maintains a file of approximately
1 A bulletin of the Department of Labor, Number 42, Sep­
tember 1902, included this note: “ It is the purpose of this
Department to publish from time to time important agree­
ments 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.

181

182

BLS HANDBOOK OF METHODS

6,500 current agreements in the private sector,
covering about 9.8 million workers.2 All in­
dustries are represented in the file except rail­
roads and airlines. Since railroads and airlines
are required to submit copies of agreements
to the National Mediation Board, the Bureau
does not attempt to collect these agreements.
As a result of the recent growth of collective
bargaining in the public sector, the file has
been expanded to include a variety of docu­
ments covering employees of the Federal Gov­
ernment, States, counties, cities, and special
jurisdictions. These documents range from
traditional collective bargaining agreements
through memoranda of agreement to executive
orders, administrative promulgations, and leg­
islative actions, which are clearly the result
of bilateral negotiations. The public file is
growing daily, presently exceeds 1,000 docu­
ments at all government levels, and involves
more than 1 million public employees.
The Bureau’s quantitative analysis of agree­
ment provisions covers virtually the entire
range of issues dealt with in collective bar­
gaining.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 dis­
tinct from special industry studies, the Bureau
is concerned with these major objectives: (1)
the presentation 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 uncom­
mon in collective bargaining agreements. For
agreements covering public employees, data
are presented by the type of government ac­
tivity (police department, veterans administra­
tion, etc.) and for cities, by population size.

Methods o f Collection and Analysis
Collection o f Agreements

The selection of agreements for the file is
currently based on two guides: To enlarge to
the fullest the opportunities for public and



governmental use of the file,4 and to provide a
diversified collection 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 con­
structing a file which truly represents all public
and private 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 agree­
ments 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 volun­
tary cooperation of employers and unions is of
utmost importance. Any restrictions imposed
by respondents on the public use of agreements
are observed scrupulously 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
identifying features, which include union, loca­
tion, number of workers covered, industry, and
effective and expiration dates. For public em­
ployee documents, the level of government and
the government activity also are coded.
Agreement Analysis

The Bureau’s utilization of the private sector
agreements it collects has moved through dif­
2 During the early postwar period, the number of agree­
ments 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 Relations. Agreements
submitted to the Bureau with a stipulated limitation on public
use are not available for inspection. Requests for informa­
tion concerning specific agreements or agreement clauses are
accommodated, depending upon the nature of the request,
within the limits of staff resources.

COLLECTIVE BARGAINING AGREEMENTS

ferent stages over the years, in pace with, or
controlled by, the increasing prevalence and
maturity of collective bargaining. During the
early years, significant agreements were re­
produced in their entirety. As collective bar­
gaining spread, and the size and representative
character of the Bureau’s file increased, atten­
tion was directed towards reproducing and
analyzing the variety of agreement clauses re­
lating to similar subjects, culled from a large
number of agreements. Although illustrative
clauses continue to be utilized in most of the
Bureau’s agreement studies, major emphasis
during recent years has been devoted to meas­
uring 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. The Bureau also has undertaken a
broader, more extensive rather than intensive
analysis, in which it searches the agreement to
measure the prevalence, but not the detailed
characteristics of more than 100 different col­
lective bargaining provisions. In these kinds
of analysis, problems relating to techniques of
coding and analysis come to the fore.
In 1948 and 1949, when the Bureau’s file
consisted of more than 12,000 agreements, it
was decided that a sample of 3,000 agreements
would be feasible. The selection of specific
agreements was based on a number of factors,
including industry, agreement coverage, loca­
tion, union representation, 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 recon­
stituting 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 advantageous alternative was to base
the private industry agreement studies on all
agreements covering 1,000 workers or more
and, thus, to avoid sampling. The Bureau’s file




183

already contained almost all of these; the Bu­
reau’s monthly report, Current Wage Develop­
ments, was a ready source of information on
those that were not included. The total number
of workers covered by agreements of this size
(now between 1,700 and 1,900) is about 7.7
million and represents a very substantial
worker coverage in agreement studies. The
number of establishments covered is not
known.5
A key analysis list containing all private in­
dustry agreements covering 1,000 or more
workers, although perhaps not the ideal cover­
age, has definite advantages: (1) It achieves
maximum worker coverage in the studies for
a given investment of staff resources; (2) it
provides a simple, objective measure of the
coverage of the studies; (3) it permits the
presentation of various combinations and
breakdowns of the data without the necessity
of complicated weighting schemes and without
the bias resulting from the lack of proper
weighting; (4) it is safeguarded against obso­
lescence, 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
analysis is that it deals mainly with legal lan­
guage, which requires interpretation, rather
than with numbers or other universal, sharply
defined attributes. The process of analysis con­
sists of interpreting 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 number of agreements and number
of workers covered. Since agreement provisions
on the whole are notable more for their variety
of expression and details than for their uni­
formity, 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
5 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 agree­
ments, 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.

184

BLS HANDBOOK OF METHODS

a special importance if significant differences
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 por­
tion. Long and legal 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), misinterpretations undoubtedly
occur. These are kept to a minimum by a staff
experienced in agreement analysis and by con­
tinuous efforts to assure consistency of in­
terpretations.

Uses and Limitations

The studies of agreement provisions are of
practical use to public and private employers
and unions engaged in collective bargaining, to
arbitrators and factfinding boards, to admin­
istrators of company wage and industrial
relations programs, and to legislators and Gov­
ernment officials. Persons not directly involved
in collective bargaining or in related admin­
istrative activities (e.g., teachers and students
of labor problems, writers for newspapers and
trade and technical journals, and foreign ob­
servers) find value in the broader aspects of
employer-employee relationships revealed in
these studies.




The limitations of these studies of agreement
provisions and employee-benefit plans are de­
termined, in large part, by their application.
For studies of paid holiday provisions or other
supplementary benefits, the fact that these
studies cover only the area of collective bar­
gaining may constitute a limitation on gen­
eralizations applying to all workers but not
necessarily on their uses in collective bargain­
ing or in wage and employee administration.
On the other hand, these studies do not show
locality practice, which may reduce their use­
fulness for some collective bargaining purposes
but not for broad generalizations relating to
workers under collective bargaining.
Additional limitations of agreement pro­
vision studies are inherent in the selection of
agreements for study— the exclusion of rail­
road and airline agreements and of agreements
covering fewer than 1,000 workers— and in the
technique of analysis, as indicated previously.
Limitations are also connected with the par­
ticular 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 beyond the agreement,
traditional industry policy, informal acceptance
by management and unions, or arbitrators’ de­
cisions, can be neither detected nor measured
in agreement analysis.

185

COLLECTIVE BARGAINING AGREEMENTS

Technical References
Number

1. Bloch, Joseph W . Union Contracts— A New Series of Studies. Monthly Labor Review, October
1964, p p . 1 1 8 4-1 1 8 5 .

2. U.S. Department of Labor, Bureau of Labor Statistics, Major Collective Bargaining Agree­
ments (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
Guarantees
1425-4, 1966: Deferred Wage Increase and Escalator Clauses
1425-5, 1966: Management Rights and Union-Management Cooperation
1425-6, 1966: Arbitration Procedures
1425-7, 1969: Training and Retraining Provisions
1425-8, 1969: Subcontracting
1425-9, 1969: Paid Vacation and Holiday Provisions
1425-10, 1969: Plant Movement, Transfer, and Relocation Allowances
1425-11, 1970: Seniority in Promotion and Transfer Provisions
1425-12, 1970: Administration of Negotiated Pension, Health, and Insurance Plans
3 . ______ Bulletin 1353, Major Union Contracts in the United States, 1961 (1962).
4 . ---------- Bulletin 1686, Characteristics of Agreements Covering 5,000 Workers or More (1970).




■Jo s e p h W . B

loch




Chapter 21. Union Membership
Background

The Bureau of Labor Statistics’ interest in
union membership as a significant social and
economic indicator is of long standing. The
first BLS publication listing membership fig­
ures for national and international labor unions
in the United States was probably the Hand­
book of American Trade-Unions, issued in 1926
(BLS Bulletin 420), followed by a similar pub­
lication in 1929 (BLS Bulletin 506) and an
extensive revision in 1936 (BLS 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 activities.
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’ ad­
dresses, etc., was supplemented with an entry
for each union on membership and number of
local unions, and a compilation of total mem­
bership. In subsequent years the information
sought from national and international unions
has expanded considerably. Data on women
members first appeared in the 1953 Directory,
and separate tabulations for areas outside the
United States (Canada, Puerto Rico, etc.), in
the 1955 edition. Since 1957, each Directory
has carried information on the number and
proportion 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 furnish estimates on the number en­
rolled by Federation affiliates. In its 1965 Di­
rectory, the Bureau showed State figures as
reported by national and international unions.
Many of the items referred to have been re­
fined since they were first introduced, and the
accumulated information now permits analysis
1 For a membership survey of these unions see Unaffiliated
Intrastate and Single-Employer Unions, 1967 (BLS Bulletin
1640, 1969).
2 The names of all reporting unions appear in the Depart­
ment’s Register of Reporting Labor Organizations, last issued
in 1968.




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 are asked to report the average number
of dues-paying members or the number of
members in good standing for the 2 most recent
years, including members outside the United
States; the proportion of white-collar and
women members; breakdowns by major in­
dustry and by State, and other membership
information. (See copy of questionnaire on pp.
188-191.) 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 un­
affiliated unions as the Teamsters, the Mine
Workers, and the Auto Workers. In addition,
the Directory accounts for all unions of Fed­
eral Government employees that have received
“ exclusive recognition” as specified in Execu­
tive Order 11491. Thus, by definition, the Bu­
reau excludes from its Directory those unions
whose activities are confined to a single locality
or to a single employer.1 Although the Bureau
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 Reports, as required by
the Labor-Management Reporting and Disclosuere Act,1 possibly some small unaffiliated
2
unions, interstate in scope, escape attention.
These inadvertent omissions do not affect mem­
bership 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
proceedings, financial statements, and collec­
tive bargaining agreements. No sampling pro­
cedures are used; the data are based on the
187

188

BLS HANDBOOK OF METHODS

B u d g e t B u re a u N o. 4 4 *S 7 0 0 2 7
A p p r o v a l E x p ir e s D e c . 31, 1971

B L S 2441

U. S. D EP A R T M EN T OF LABO R
Bureau of Labor Statistics
Washington, D. C. 20212

DIRECTORY OF NATIONAL AND INTERNATIONAL
LABOR UNIONS IN THE UNITED STATES, 1971
I.

T . I.phono number

U n io n a nd O ffic e r Id e n tific a tio n :

and area code:

1. Affiliation
□

( Check appropriate box

2. President:

(□

M
r.

3. Secretary-Treasurer:

□

AFL-CIO
Mrs.

1 1 Mr.

□

None

□

Other (s p e c ify )

1 | Miss

Year first elected to this
office:

1 1 Mrs.

4. Person in charge of organizing activities:

□

Miss

( □ Mr.

Year first elected to this
office:

[ □ Mrs.

□ ] Miss

___________________________________________________________________ Tide:.
Mailing address if different from headquarters (Street, City, State, Zip Code):

5. Research Director:

I I Mr.

I I Mrs.

|_ |Miss ______________________
_

Mailing address if different from headquarters ( Street, City, State, Zip Code):

6. Education Director:

Q M r.

□

Mrs.

□

M iss______________________

Mailing address if different from headquarters (Street, City, State, Zip Code):

7 . P e rs o n in ch a rge o f s o c ia l in su ra n ce (health, insurance, pension, e t c .) a c t iv i t ie s :

□

Mr.

□

Mrs.

□ M i s s _________________________________

T itle :___

Mailing address if different from headquarters (Street, City, State, Zip Code):

8. Person in charge of legal activities:

□ ] Mr.

□ ) Mrs.

( □ Miss

___________________________________________________________________ T id e :____
Mailing address if different from headquarters (Street, City, State, Zip Code):

9. Person in charge of legislative activities:

□ ] Mr.

□ ) Mrs.

[ □ Miss

------------------------------------------------------------------------------------------------------- Title: Mailing address if different from headquarters (Street, City, State, Zip Code):

10. Person in charge of public relations activities:

( □ Mr.

( □ Mrs.

___________________________________________________________________

□ ] Miss
T i d e :___

Mailing address if different from headquarters (Street, City, State, Zip Code):




189

UNION MEMBERSHIP

II.

Conventions and Publications:

1. Frequency of conventions:

[ / j Annual

Semi-annual

2. Next convention:______________________
Month
Day
Year

Other ( s p e c ify ) ___

____________________________________________________
(C ity, State)

3. Name of official publication(s)

III.

Biennial

How often published

Editor

Affiliated Bodies:
Indicate number of locals in operation at the end of 1970.

IV. Membership:
1. Indicate annual average dues-paying membership count for 1969 and 1970. If complete returns for 1970 are not available,
use 9 or 10 month average:
1969

members

2. If your records do not permit
an annual average dues-paying
good standing or those carried on your rolls:
1969___________________ members

1970___________________ members
membership count, please

indicate the number ofmembers in

1970___________________ members

3. Indicate if retired members are included in 1 or 2 above:
Q
Yes
Q No
If yes, indicate number of retired members included: _____________________members*
1

V.

Classification of 1970 membership:

1. Indicate approximate percentage of membership who are women:______________________________________ %
( i f none, enter zero)

2. Indicate approximate percentage of membership in the following “ white-collar” categories:
Professional and/ortechnical
C lerical------------------------------Sales-----------------------------------Total “ white-collar” - -------

__________ %
%
%
%

( i f none, enter zero)

3. Industry composition of union membership. Indicate the approximate percentage of all union members working in
establishments in each of the following industry groups:

Manufacturing:
Ordnance and a ccessories--------------------------------------------------------------------------------------------------Food and kindred products (incl. b e v e r a g e s ) --------------------------------------------------------- -----------Tobacco manufactures---------------------------------- r ------------------------------------------------------------------Textile mill products-------------------------------------- ------------------------------------------------------------------Apparel and other finished products made from fabrics and similar materials-------------------------Lumber and wood products, except furniture------------------------------. . . . . . ----------. . . . . . . . . ----Furniture and fixtures---------------------------------------------------------------------------------------------------------Paper and allied products-------------------------------------------------------------------------------------------------Printing, publishing and allied industries------- ------------------ . . . . . -----. . . . . . . . . . . . . . . . . . . . . .
Chemicals and allied products---------------------------------------------- ---------------------------------------------Petroleum refining and related industries----------------------------------------------------------------------------Rubber and misc. plastics products------------------------------------------------------------------------------------Leather and leather products------------------- --------------------------------------------------------------------------Stone, clay, glass, and concrete products ----------------- ----------- --------------------------------------------Primary metals industries-------------------------------------------------------- -------------- -------------Fabricated metal products, except ordnance, machinery and transportation equipment ----------Machinery, except electrical...........- ------------------------------------------------ ------------- r -------------------Electrical machinery, equipment and supplies---------------------------------------------------------------------Transportation equipment---------------------------------------------------------------------------------------------------Professional, scientific and controlling instruments; photographic and optical goods, watches
and clocks-------------- — ---------- ----------------------------------------------------------------------------------------Miscellaneous manufacturing industries-----------------------------------------------------------------------------




190

BLS HANDBOOK OF METHODS

- 3-

V.

C la s s if i c a t i o n o f 1 9 7 0 m e m b e rsh ip : (Continued)

3. Indicate industry composition of union membership
N o n m a n u fa c tu rin g :

Mining and quarrying (including crude petroleum and natural gas production) ---------------------------------Contract construction (building and sp e c ia l trade) ---------------------------------------------------------------------------Transportation services (including railroads, air, bus, truck and water transportation, and allied

------------------- %
------------------- %

s e r v i c e s ) -----------------------------------------------------------------------------------------------------------------------------------------

--------------

Telephone and telegraph---------------------------------------------------------------------------------------------------------Electric, gas, and sanitary services (including w ater) --------------------------------------------------------------------Wholesale and retail trad e--------- -------------------------- ---------------------------------------------------------------Finance, insurance and real estate-------------------------------------------------------------------------------------------Service industries (including hotels, laundries and other personal s e r v ic e s , repair s e r v ic e s ,

-------------

motion pictu res, amusements and related s e r v ic e s , hospitals, educational institutions, non­
profit membership organizations) ----------------------------------------------------------------------------------------------------

Agriculture and fishing------------------------------------------------------------------------------------------------------------Nonmanufacturing (classification not available) --------------------------------------------------------------------------------

--------------

-------------------------

--------------

--------------------------

G overnm ent:

F ed era l----------------------------------------------------------------------------------------------------------------------------------State and lo c a l------------------------------------------------------------------------------------------------------------------------Total.....................................

------------------------.........100

4. State distribution of union membership. Indicate the approximate number or percentage of members
in each of the 50 States.
Alabama----Alaska -----Arizona —
Arkansas - California Colorado -Connecticut
Delaware
Florida-----Georgia-----Hawaii-----Id aho------Illinois-----Indiana----Io w a -------■
K a n s a s--------

Kentucky---Louisiana -•

Maine----------------Maryland—District
of Columbia —
Massachusetts---Michigan-----------Minnesota - --------Mississippi--------M issouri-----------Montana-----------N ebraska----------Nevada -------------New Hampshire
New J e rs e y ------New M ex ico ------New York.............
North Carolina
North Dakota ---O h io-----------------

Oklahoma------------------Oregon----------------------Pennsylvania-----------Rhode Isla n d -----------South Carolina----------South D akota-----------Tennessee ---------------Texas .......................
U tah.............................
Vermont--------------------V irginia-------------------Washington---------------West Virginia------------Wisconsin-----------------Wyoming...................
No. or percentage
not accounted for in
the U .S ....................

5. 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 1970 or any other appropriate current period:
Location
Canada
Puerto Rico
Canal Zone
Other (s p e c ify )




Approximate number of union members

Number of local unions

191

UNION MEMBERSHIP
- 4V I.

1.

T e rm o f O ffic e r s:

* Are officers required to retire at a specified age?
National officers:

I I Yes

[_J No;

If yes, age___________

Local officers:

I I Yes

3

If yes, age___________

No;

2. Are the number of terms an officer may serve limited?
National officers:

1.

Yes

3]

No;

If yes, number of terms

Local officers:

V I I.

[31
3]

Yes

3 1 No;

If yes, number of terms.

C o lle c t iv e B a r g a in in g A g re e m e n ts:

Indicate the number of basic collective bargaining agreements with employers. Exclude supplemental and pension,
health and insurance agreements:
Total_______________ agreements.
United States_________________ agreements.

2. (a) Indicate number of different employers covered by collective bargaining
agreements---------------------- ------------------------------.employers
United States o n ly ------------------------------ -employers
(b) If more than one employer, are the employers located in at least two
States?------------------- ------------ 3 ] Yes
3 No
3. Indicate the number of workers covered by these agreements. Include nonmembers in the bargaining units:
Total________________ workers.
United States________________ workers.
4.

Indicate the number of agreements in the United States that:
Expired in 1970_________________agreements.
Will expire in 1971_________________ agreements.

5.

Indicate the number of agreements in the United States that provide a wage reopener in:
1970

_______________agreements.

1971

_______________agreements.

V I I I . U n :on Staff:

Indicate the number of full-time employees on the payroll of the national in the United States. Exclude elected
officials and employees on the payrolls of local unions and intermediate bodies:
Managers and administrators---------- ---------

---------------

Professionals-------------------------------------- --

---------------

Clerical and secretarial--------------------------

---------------

Organizers and representatives----------------

---------------

Others ( s p e c i f y ) .....................- .....................

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

May we have your comments regarding the present Directory and proposals for changes in future editions?

Name of person reporting




Title

Date

192

BLS HANDBOOK OF METHODS

entire universe of national and international
unions, as defined.
Presentation

The data for each union and summaries are
presented 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 ques­
tionnaire items such 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 distribu­
tion. Gains and losses over the past decade are
analyzed, and major unions are ranked accord­
ing to size at particular points in time. Of con­
siderable interest to users of the Directory are
the changes of union membership related to
changes in the total labor force and to em­
ployees in nonagricultural establishments. Data
by industry and on women and white-collar
members permit rough approximations on the
extent of union penetration compared with
earlier periods.
In addition to statistical summaries, the
Directory also contains appendixes which list,
for each union, membership in areas outside
the United States, number of women, propor­
tion of white-collar members, and proportion
of members in major industry groups.
Uses and Limitations

The Bureau’s membership figures are pub­
lished on a regular basis, and serve as the
principal indicator of gains and losses for par­
ticular 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 relations policies and
their economic effects can be made between
organized and unorganized sectors. They are
used by agencies of the Federal Government,
State and local governments, by management
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 organi­
zations and labor unions.



Difficulties in measuring union membership
arise from (1) the variety of concepts and
practices among unions as to the definition and
reporting of membership, (2) the availability
at union headquarters of the various data re­
quested, 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, uni­
form reporting practices do not result from
applying this criterion alone. Unions define
eligibility for membership in a variety of ways
and payment of dues is only one of several
criteria. Some unions set less than full dues
requirements or waive the payment for work­
ers 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 repeat­
edly has requested unions to indicate whether
they include or exclude from membership re­
ports five specified groups: the unemployed;
those involved on work stoppages; those in the
Armed Forces; apprentices; and the retired.
Moreover, unions were asked to furnish an
:! 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 indi­
cated that while a number of unions use a “ per capita”
receipt figure in their reports to the Bureau, such an ap­
proach was inappropriate in the case of other unions for
reasons which the unions explained in detail. Some unions
include a large number of seasonal employees whose dues
payments are limited to several months during the year.
Other unions questioned the use of a computed membership
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 indi­
cated 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 computa­
tions 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.

193

UNION MEMBERSHIP

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 work­
ers who, at least in some way, are still attached
to unions. The responses, however, have 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 num­
ber of workers involved.4
Thus, no uniform answer exists to the de­
ceptively simple question: Who is a union
member? The answer varies from union to
union, as determined by its own policies and
practices.
Although financial statements may be of
* 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 at A F L CIO conventions, based upon average paid membership to
the Federation. This, however, is not a reliable measure.
Some unions, for reasons 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 AFL-CIO.

some help in arriving at membership approxi­
mations, they cannot be used to obtain data on
various components of union membership, such
as the proportion of white-collar and women
members, those under contract in particular
industries, etc. For such data, the Bureau must
rely entirely on the cooperation of national
unions. These, however, are not always able to
furnish the information for the simple reason
that it is not compiled at union headquarters.
Many national unions are decentralized or­
ganizations and as a rule, do not 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
officials of national and international unions.
Perhaps more accurate data, and certainly
more detailed data, could be obtained by con­
tacting local unions, as is done in Canada, but
the task of soliciting responses from more than
70,000 organizations is beyond the present re­
sources of the Bureau.

Technical References
Number

1.

2.
3
4.

5.
6.

7.

Troy, Leo. Distribution of Union Membership Among the States, 1939 and 1953. New York,
National Bureau of Economic Research, 1957.
A study measuring union growth by State and region, analyzing geographic and indus­
trial shifts in membership. Includes a discussion of sources and methods of measurement.
U.S. Department of Labor, Bureau of Labor Statistics. “ Membership of Labor Unions, 18971950.” Handbook of Labor Statistics, 1950 ed., (Bulletin 1016, 1951), pp. 137-139.
. ______Directories of Labor Unions in the United States; (Bulletins 937, 1948; 980, 1950; and
1127, 1953).
--------- - Directories of National and International Labor Unions in the United States (Bulletins
1185, *1955; 1222, 1957; 1267, 1960; 1320, 1962; 1395, 1964; 1493, 1966; 1596, 1968; and
1665, 1970). In various articles in Monthly Labor Review. (Excerpted)
In addition to membership data, the Directories also include a brief description of the
methods used, and a copy of the questionnaire.
Paschell, William, “ Limitations of Union Membership Data,” Monthly Labor Review, Novem­
ber 1955, pp. 1265-1269.
Wolman, Leo. Ebb and Flow in Trade Unionism, New York, National Bureau of Economic
Research (1936).
Of invaluable aid to students in the field because of its extensive discussion of measure­
ment problems. Appendix tables present figures for the period 1900 to 1934 including data
on the extent of organization by industry.
U.S. Department of Commerce, Bureau of the Census, Current Population Reports, Series P 20, No. 216, “ Labor Union Membership in 1966” (1971).
The first comprehensive study giving economic and demographic data for union and non­
union private wage and salary workers.




— H arry P. Co h a n y




Chapter 22. Annual Earnings and Employment Patterns
of Private Nonagricultural Workers
Background

The purchasing power of workers depends
largely on their annual earnings which are
determined by the interaction of variables such
as straight-time rates of pay, number of hours
worked, and the hours worked at premium
rates which depend on other variables like occu­
pation, union status, industry, and area. More­
over, some workers move into and out of the
labor force during the year. In addition, a
substantial portion work for more than one
employer in the same industry and for one or
more employers in different industries. Con­
sequently, occupational wages or hourly and
weekly earnings data cannot be extrapolated to
annual earnings estimates with any degree of
precision.
The Bureau’s program of annual earnings
and employment patterns studies, initiated in
the late 1960’s, is designed to fill the gap in
our knowledge of annual wage and salary
earnings from private nonagricultural employ­
ment. The first study 1 for 1964, was limited to
wage and salary earnings covered by social
security; the second,1 for 1965, included data
2
on wage and salary earnings covered under
either the Social Security Act or the Railroad
Retirement Act. Subsequent studies, like that
for 1965, will include information on earnings
covered under either social security or railroad
retirement. Bulletins providing 2 years of data
are expected to be issued biennially. Special
analytical studies will be published intermittantly in the Monthly Labor Review.

ticipate in social security and function like
private firms, such as hospitals and schools.
The studies provide distributions of median
and mean earnings for the private nonagri­
cultural sector as a whole, for each industry di­
vision, for each major industry group at the
two-digit SIC level, and for selected industry
groups at the three-digit SIC level of industry
classification. The data, though available only
for white-collar and blue-collar workers com­
bined, are unique, because unlike annual earn­
ings data from other sources, they permit an
analysis of the distribution of wage and salary
earnings and employment patterns by industry
and quarters of employment.
The studies focus separately on earnings in
the industry in which workers had greater
earnings than in any other, and earnings in all
industries. Some of these data and some of
the employment patterns data, separately and
with earnings data, also are presented for
selected demographic characteristics.
Earnings and employment patterns of
workers who had covered wage and salary
each quarter of the year are emphasized par­
ticularly. Information about earnings and em­
ployment patterns of four quarter workers is
the closest to data for workers fully attached
to the private sector work force that can be
obtained from the source materials. However,
some workers who have earnings in each quar­
ter are attached to the work force to only a
limited extent.

Source o f Data
Description o f the Series

The series covers earnings and employment
patterns in the private, nonagricultural sector
which, broadly defined, includes individuals
who work for wages and salaries in employ­
ment covered by the Social Security Act and
the Railroad Retirement Act.3 Excluded are
earnings in agriculture, self-employment, and
in government units other than those that par­



The data are developed by the Bureau of
Labor Statistics from statistical information
1 Annual Earnings and Employment Patterns, Private Non­
agricultural Employment, 1964 (BLS Report 330 1969).
2 Annual Earnings and Employment Patterns of Private
Nonagricultural Employees, 1965 (BLS Bulletin 1675, 1970).
3 For a discussion of the interrelationship of, and joint
coverage under, the railroad retirement and social security
systems, see U.S. Department of Health, Education, and W el­
fare, Social Security Administration, Social Security Hand­
book, 4th ed. (1969), and U.S. Railroad Retirement Board,
Handbook on Railroad Retirement and Unemployment Insur­
ance Systems (1969).

195

196

BLS HANDBOOK OF METHODS

obtained from the Social Security Administra­
tion and the Railroad Retirement Board. To
preserve the confidentiality of the records, the
data are provided to the Bureau of Labor
Statistics without identification of individuals
or employers. However, to combine data from
various employments and to facilitate statistiscal processing, each individual and employer
is assigned a permanent control number, differ­
ent from his social security or employer identi­
fication number.
Each individual in the sample provides
demographic information (race, sex, and year
of birth) when he applies for a social security
number. Each employer under social security
from whom the individual receive any wages
or salaries during a calendar quarter reports
the amount of the wage payment in the quar­
ter; employers covered under railroad retire­
ment .report monthly. However, employers,
cease to report wage and salary earnings after
the worker has reached his taxable earnings
limit in that employment situation. Employers
report wages paid to the maximum annual
limit under social security and to the maximum
monthly limit under the railroad retirement
system. Employer reports also indicate the
industry and, except for employment covered
by the Railroad Retirement Act (R R A ), the
area in which the wages or salaries were
earned. Employers subject to the RRA, also
provide information about the occupational
category the worker was employed in.

Sample Design, and Sampling and
Nonsampling Variability

The sample, which includes 1 percent of all
social security numbers, was selected on the
basis of a multistage systematic cluster sam­
pling procedure. Social security numbers are
used as the individual’s identification number
in both the social security and railroad retire­
ment systems. An individual selected for the
sample remains in it permanently.4
Since estimates in this study are based on a
sample, they may differ from census figures.
Moreover, the sample data are not adjusted to
benchmark levels established by complete




counts. In addition, the data are subject to
nonsampling variability due to errors in re­
porting and classification and other possible
error sources, that would be present in a com­
plete enumeration as well as in a sample. As
a result, ratios established from the sample
data are considered to be reasonable estimates
of those existing within the population as a
whole. Nevertheless, particular care should be
exercised in interpreting medians and percents
based on relatively small numbers of cases as
well as small differences between figures.5 Ab­
solutes are subjet to large error which may
vary from year to year.

Estimating Procedures

To estimate total wages of individuals, the
Social Security Administration determines the
quarter in which the taxable limit is reached
( “ limit quarter” ). Wages in the prior quarter
equal to or greater than the “ limit quarter”
wages are substituted for those in the “ limit
quarter” and in all subsequent quarters. Limit
quarter earnings, however, are used to estimate
earnings in the limit and subsequent quarters if
limit quarter earnings were higher than earn­
ings in previous quarters. The summation of
the quarterly wages after these substitutions
then becomes the estimated annual total. When
the taxable limit is reached in the first quarter,
the Social Security Administration imputes an
estimated total.
Employers covered by the Railroad Retire­
ment Act are required to provide information
about the monthly earnings of each employee up
to the maximum creditable limit subject to
Railroad Retirement Act taxes. Hence, even
earnings reported at the maximum level for
each month aggregated to annual totals may be*
3
4 For a detailed discussion of the sampling procedure, re­
porting criteria, and coverage under the social security and
railroad retirement systems, see U.S. Department of Health,
Education, and Welfare, Social Security Administration,
Workers Under Social Security, 1960 (1968) and Social Se­
curity Handbook, op. cit.; also see Handbook on Railroad
Retirement and Unemployment Insurance Systems, op. cit.
3 For an indication of the order of magnitude of the sam­
pling errors and a fuller discussion of the sampling and
non sampling variability to which the series is subject, see
Workers Under Social Security, 1960, op. cit.

ANNUAL EARNINGS & EMPLOYMENT PATTERNS OF PRIVATE NONAGRICULTURAL WORKERS

substantially below the worker’s total earnings.
The Railroad Retirement Board, however, col­
lects information from employer records about
the total annual earnings of a sample of
workers covered by the act. Factors for raising
creditable compensation under the Railroad Re­
tirement Act to total railroad earnings are
derived by comparing the total earnings data
for individuals collected in the special study
with the aggregated monthly earnings data for
the same individuals. The incremental factors
for workers in the same broad occupational
categories then are averaged. The resulting
factors, developed by the Railroad Retirement
Board, are applied by the Bureau of Labor
Statistics to the credited monthly earnings of
each individual by taking into account his
occupational category.6

Analysis and Presentation

Distributions of workers by annual earnings,
and mean and median earnings are presented
for all major earners and for those with four
quarters of earnings. In addition, data are
provided showing the number of workers who
had wage and salary earnings in the various
industries and the proportion of these who had
greater (major) earnings in that industry than
in any other industry. Data on major earners
by industry showing the proportion with earn­
ings in each quarter, their demographic char­
acteristics, what proportion had earnings
within a single industry and the number of
employers whom they worked for also are
presented. Some earnings and employment pat­
terns data are presented for broad geographic
areas.
Terms used in the series and the methods
used to classify workers by industry and region
of major earnings and the industrial classifi­
cation scheme used in classifying nonpolicy
governmental units in scope of the study pro­
gram are described below.
Annual earnings are defined as gross wages,
salaries, and other payments (such as bonuses)
received by employees, before deductions, in
employment covered under the Social Security
Act or the Railroad Retirement Act. Such pay­
ments may be cash, cash equivalents, or other



197

media such as goods, clothing, board, or
lodging.
Self-employment earnings, payment for work
in employment excluded from the coverage of
the acts, and payment for work in agriculture,
in covered governmental units engaged in
public administration and for military service
have been excluded from this study. Most pay­
ments by employers to or on behalf of em­
ployees, or for employees and their dependents
for retirement, death, sickness or accidental
disability, or medical and hospitalization ex­
pense under provisions of a plan or system
meeting certain general criteria, and employer
payments to a trust fund, such as a pension
trust, exempt from tax under the Internal
Revenue Code, are not counted as wages in this
series.7
Workers with some earnings in the industry.
Each individual who earned $1 or more in an
industry during the year is counted in each
industry in which he had any earnings. A
worker who had some earnings in each of five
three-digit industries, as defined in the Stand­
ard Industrial Classification Manual8 for ex­
ample, is counted for each of these industries
as well as in each two-digit industry and in
each division of which the three-digit industres
are a part. Because a worker is counted in each
three-digit industry, each two-digit industry
and in each division in which he had $1 in cov­
ered wage and salary earnings or more, the
aggregate count at each level is greater than
the total number of workers at each broader
industry level (two-digit, division, private nonagricultural economy).
8 Although the Railroad Retirement Board collects annual
earnings data for a 1-percent sample of workers, the sampling
criteria differ from those used by the Social Security A d­
ministration. To permit the combination and integration of
data from the two systems, the Railroad Retirement Board
provides the Bureau of Labor Statistics with data for a
sample of workers selected according to the sample design
established by the Social Security Administration.
7 Under certain circumstances tips are counted as wages for
social security purposes and thus are included in the data
file used in this study series. Payments to workers from tax
exempt trust funds are not considered wages (except for
wages paid by the fund to its employees) and thus are
excluded. Employer payments to trust funds which are not
tax exempt also are excluded. Depending on their nature
and purpose, payments to workers from these funds may
be counted for social security purposes and thus are included
in the data file.— Such would be the case if the payment was
a bonus or vacation payment. For a more detailed discussion
of covered wages, see the Social Security Handbook, op. cit.
8 Issued by the U.S. Office of Management and Budget.

198

BLS HANDBOOK OF METHODS

Major earners and industry of major earn­
ings. A major earner in this study, is a worker
who earns more of his annual wages and
salaries in a specific industry than in any other
industry. This test to determine the worker’s
industry of major earnings is based on a
plurality earnings concept and is applied sepa­
rately at each industry level.
Data for major earners are included in only
one industry at each level in industry classifi­
cation (i.e., at the three-digit, the two-digit,
and at the divisional levels). Thus, the number
of workers who received the major proportion
of their earnings in each industry, at any level
of industry classification, is unduplicated.
Hence, data for any level of classification at
which all component industries are shown add
to the total for the private nonagricultural
sector. However, data at the three-digit level
(if all three-digit industries were shown)
might not add to the detail at the two-digit
level, and data at the two-digit level might not
add to the divisional level. Except for totals at
the private nonagricultural sector, this non­
additive relationship exists because some
workers shift employment several times during
the year and bring about patterns like that
shown in the following tabulation of a hypo­
thetical worker’s employment and quarterly
earnings experience.
Total
(any
quarter)
Private non­
agricultural
econom y_
_
Division A
2-digit group,
A -l
3-digit group,
A -ll _
..

$530

January
to to
March

$80

July
Aprilto
Septem Juneber

October
to
D ecem ber

$170

$130

$150

150

60

90

150

60

90

50

90

150

Division B
2-digit group,
B -l
3-digit group,
B -l l
3-digit group,
B-12 __ _

180

60

50

60

50

70

130

10

50

70

50

50

Division C
2-digit group,
C -l
_ ___
3-digit group,
C -l l _ - __
2-digit group,
C-2
______
3-digit group,
C-21 _
___

200

20

20

60

100

20

20

60

100

20

20

60

Quarters o f W ork 1 1
0

In this series, because workers are classified
both by quarters worked 1 in their major in­
0
dustry and in all wage and salary employment,
quarters of work of the hypothetical worker
whose employment and earnings were illus­
trated previously are as follows:
Major
industry
Private
Division
2 - digit
3 - digit

nonagricultural __
C ________________
group, B -l
____
group, A -l l

4
4
3
2

All wage and
salary em ploym ent
4
4
4
4

Employer. An employer in this study is an
individual, partnership, or corporation recog­
nized under the law as a separate entity meet­
ing certain criteria.1 However, a firm which
1
separately incorporates at each of its locations
may be considered a separate employer at each
location. Thus, a worker transferred from one
to another location that is incorporated sepa­
rately may have more than one employer in
the same year even though he continued to
work for the same firm.
Industrial classification. Employment and
earnings data based on the Social Security Ad­

70

180

In this tabulation, the worker had greater
earnings at the three-digit level in industry
A - ll than in any other three-digit industry; at
the two-digit industry he had greater earnings
in industry B-l than in either A -l, C-l, or
C-2; and at the divisional level he earned more
in division C than in either A or B. Therefore,
applying the plurality earnings test, this
worker is assigned to industry A - ll at the
three-digit level, B-l at the two-digit level, and
to C at the division level.9

100

100

100

100

100




9 An unpublished tabulation, prepared by the Social Security
Administration, indicated the industry of major earnings at
the three-digit level for about 12 percent of all covered
workers was part of a two-digit industry different from the
workers two-digit industry of major earnings.
1 “ Quarters of work” and “ quarters of coverage” are not
0
synonymous. A quarter of coverage based on covered non­
agricultural employment is a calendar quarter in which the
individual was paid $50 or more; a quarter of coverage also
may be credited on other bases such as agricultural employ­
ment or self-employment.
1 For details, see the Social Security Handbook, op. cit. In
1
addition, because some workers work for more than one
employer during the same week, data showing number of
employers should be used with caution.

ANNUAL EARNINGS & EMPLOYMENT PATTERNS OF PRIVATE NONAGRICULTURAL WORKERS

ministration’s (SSA) data file are classified
according’ to the Administration’s industrial
classification system. This system differs
slightly from that published in the Standard
Industrial Classification Manual (SIC), and
used in most statistical series in the assignment
of industry codes to governmental units. Most
statistical series classify governmental units
into SIC Division I-Government. The SSA,
however, classifies only policymaking govern­
mental units in Division I. All separable non­
policymaking units are assigned to nongovern­
mental SSA industry classification code appro­
priate to their activity. Thus, for example, all
employment with policymaking boards of edu­
cation (classified by the SSA and SIC as gov­
ernment) has been excluded from this study.
Schools, colleges, and other operating units cov­
ered under voluntary election provisions of the
act, however, were treated as service industry
employment, because the units were classified
by SSA into SIC 82, educational services.1
2
Employment and earnings data based on the
Railroad Retirement Board’s file are classified
into the following industries as defined in the
SIC Manual: Railroads, SIC 401; sleeping car
companies, SIC 402; express companies, SIC
404; rental of rail cars companies, SIC 474;
and other companies performing services rail­
road transportation and certain railway labor
organizations, SIC 861 and 863. In each case
the assignment is based on the industrial classi­
fication of the worker’s last employer who was
covered under the Railroad Retirement Act.
Data for workers who had earnings in more
than one industry, all of which were covered
1 For detailed information, see U.S. Department of Health,
2
Education, and Welfare, Social Security Administration, Com­
parison of Social Security Administration and Standard In­
dustrial Classification Systems, 1963 (undated) and the guide
prepared by the Administration in 1968, entitled “Industrial
Codes in the Social Security Administration Continuous
Work History Sample (CW HS), Data for 1957 through 1966.”
1 For all industries, except water transportation, employ­
3
ment outside the 50 states or the District of Columbia is only
a small fraction, if any, of the total number of the industry’s
major earners.
1 Employment under the Railroad Retirement Act makes up
4
nearly all the major earners in the railroad industry (SIC 40)
and a substantial proportion of those at the all transportation
level, but only a very small fraction of the total employment
in other industries or at the total private nonagricultural
level. Thus, the convention adopted has a serious effect on
the regional employment distributions and regional earnings
picture at the all transportation level but little effect on
other industries or at the total private nonagricultural level.




199

under the social security system or one of which
was covered under the railroad retirement
system, are classified and presented separately
and in combination.
Single and multi-industry workers-at each
level of industry classification the employment
experience of each sample member was ex­
amined to see if all of his earnings were in
one or more than one industry. Those with
earnings in more than one industry were classi­
fied as multi-industry workers. This conceptual
approach may be seen in the case of a worker
who was employed, as illustrated below, by an
employer in each of two three-digit industries
within the same two-digit industry.
Industry level

Number of employers

Division A _____________ _ _____
_

______

2

2-

digit group, A -l ...__________ ______

2

3-

digit group, A - l l

1

_________ ______

3-digit group, A-12 ______________

___

1

Region of major earnings. The region in which
workers had their major earnings is deter­
mined by a plurality earnings test similar to
that described in detail in “Major earners and
industry of major earnings.” The region in
which he had greater earnings than any other
is his region of major earnings.
In a few cases, earnings in the industry and
region of major earnings may not coincide.
All data for major earnings are classified first
by industry and then by region. A worker who
earned 40 percent of his annual wages in in­
dustry A in the Northeast, 30 percent in in­
dustry B in the South, and the remaining 30
percent in industry C also in the South would
be classified as a major earner in industry A
and as having had his major earnings in the
South.
In this series, workers employed under the
Social Security Act or the Railroad Retirement
Act are divided into five regions: Four cover
the 50 States and the District of Columbia; the
fifth includes all employment in U.S. territories,
on foreign soil, or aboard ocean-going vessels.1
3
Data in this series do not indicate where
wages and salaries covered by the Railroad Re­
tirement Act were earned.1 Therefore, a con­
4
vention was adopted ascribing all such em­
ployment to the North Central region where

200

BLS HANDBOOK OF METHODS

many railroads and railroad-related organiza­
tions have their headquarters.
Race. All workers in this series have been di­
vided into two groups, “ white” and “ Negro.”
The white category includes all workers except
Negroes. This convention, which is different
from that used in most statistical series, was
adopted to minimize the effect on those groups
for whom the sample was not sufficiently large
to present data separately and to maximize the
analytical usefulness of the data.

Uses and Limitations

The data provide an insight into the answer
to the question: “ How well do private nonagricultural workers do in their industry of
major earnings and to what extent do they
supplement these earnings by employment in
other private sector industries.” Data are used
in collective bargaining; in formulating public
policy and in making inter-industry and inter­
national comparisons; in analyzing the distri­
bution and diversity of earnings, and varia­
tions in employment patterns among industries,
regions, and between workers of different races
and sexes.
The data, however, have several substantial
limitations. Some workers whose annual earn­
ings are included in the series also have
earnings in self-employment, agriculture, or
employment in governmental units excluded
from the series. As a result, these workers
appear to have low annual earnings. Probably
most of these workers were attached to the
employed work force only a very limited extent.
The unavailability of data on hours or weeks
worked or paid for, or occupational group
(other than in railroading) seriously limits the
analytic potential of the series. The 3-5 years
lag between the reference period and the date
of publication introduces another limiting
factor. These delays result primarily from re­
porting requirements established by the law,
the administrative requirements of the collect­
ing agencies, and the processing required to
reduce the mass of micro data into statistical
summaries. Nevertheless, since employment



patterns usually change very slowly and rela­
tive earnings distributions generally are quite
comparable from year to year even though the
level of earnings moves upward, the relation­
ships shown are indicators of the current
situation.
Nothwithstanding the limitations, data from
this series, unlike those from other sources,1
5
permit an analysis of the distribution of wage
and salary earnings and employment patterns
of workers by demographic characteristics, in­
dustry, and quarters of employment. Thus, they
are uniquely useful to all concerned with the
annual wage and salary income of individuals
and the employment patterns of those in the
private nonagricultural work force.
1
5 Dissimilarities in concept or method between the BLS
annual earnings and employment pattern series and other
series may result in important differences in sampling and
nonsampling variability between series. Therefore, caution
must be exercised in using data from the BLS annual earn­
ings series in conjunction with data from other statistical
series. The major sources of other annual earnings data to­
gether with a brief explanation of the most important differ­
ence between them and the data in this series are noted
below.
The Social Security Administration (SSA) publishes some
annual earnings data by industry. Their most recent report
Workers Under Social Security, 1960 (1968), provides sta­
tistical information about employment, earnings, and insur­
ance status of workers under old-age, survivors, disability,
and health insurance. The SSA also publishes selected sum­
mary data in the Social Security Bulletin. The industry
attachment concept used by the SSA, however, is different
from that used in this study. Further, the Social Security
data do not include earnings in employment covered by the
Railroad Retirement Act.
The Railroad Retirement Board (RRB) annually publishes
a “ research and statistics note” which provides information
about the total railroad earnings of railroad employees. The
RRB data, however, exclude earnings in employment not
covered by the Railroad Retirement Act.
Some annual earnings data at the all-industry level by
occupational group are published by the Bureau of the
Census, U.S. Department of Commerce, in “ Consumer In­
come,” Series P-60 of the Current Population Reports. This
publication also provides a distribution of wage and salary
earnings, at the all-industry level, by sex and race. In addi­
tion average earnings by sex are presented for selected in­
dustry divisions and for a few major industry (two-digit
SIC) groups. The study, based on a household survey, does
not provide distributions of wage and salary earnings by
industry group and has different concepts of industry attach­
ment from those used in this study.
The Office of Business Economics of the U.S. Department
of Commerce also publishes estimates of the average annual
earnings of “full-time employees” in its Survey of Current
Business; these estimates do not reflect the effect on average
earnings of workers who work less than a full year.

— A

rnold

Strasser

Chapter 23. Measuring Collective Bargaining Settlements
Background

The Bureau’s program of measuring the
effects of collective bargaining settlements on
hourly labor compensation is a reflection of
two developments: One, the growing impor­
tance of fringe benefits as a proportion of
employee compensation, and two, increased
concern about the effects of collectively bar­
gained wage and benefit changes on the price
level.
Whereas in earlier years the economic terms
of negotiated settlements could be equated
largely with agreed-upon changes in wage
rates, today, possible changes in a host of pay
supplements must be considered— such as vari­
ous forms of premium pay, paid leave, cash
bonuses, and employer contributions to funds
providing pension or health and welfare bene­
fits. Although straight-time pay for working
hours is still the major element of compensa­
tion, supplements are now a significant portion,
accounting for about a fifth of total employer
outlays for worker compensation.1
Moreover, growing concern during the 1960’s
over the extent to which increased labor costs
1 See Alvin Bauman, “ Measuring Employee Compensation
in U.S. Industry,” Monthly Labor Review, October 1970, p.
23. It is difficult both to measure the growth of supplements
over the years and to quantify their current importance.
The national income accounts provide one pertinent source
of data. They show that supplements to wages and salaries
rose as a percentage of total employee compensation from
1 percent in 1929 to 10 percent in 1970. The National Income
and Product Accounts of the United States, 1929-1965: Sta­
tistical Tables, U.S. Department of Commerce, Office of Busi­
ness Economics (1966), p. 14; Survey of Current Business,
January 1971, p. 10. These figures, however, do not reveal
the relative importance of all supplements in either year,
since many— such as premium payments, leave payments, and
cash bonuses— appear as parts of wages and salaries. The
figures therefore are not comparable to those found in sur­
veys of employer expenditures for supplementary compen­
sation.
2 Publication of first-year changes is a recent innovation.
Before 1970, the Bureau published two measures of change
over the life of the contract, the so-called equal timing and
the time-weighted measures. The former assumed equal
spacing of changes during the term of the contract; the latter
took account of the actual effective dates of wage and benefit
changes. The time-weighted measure has been discontinued,
because it appeared to be of significance primarily for
the analysis of individual settlements rather than for overall
series of the type produced by the Bureau. Moreover, drop­
ping of the time-weighted measure and introduction of a
series on first-year changes provides parallel statistics both
on wage-rate changes alone and on wages and benefits
combined.




may contribute to inflation has heightened in­
terest in the size of collective bargaining
settlements.
Responding to these influences, the BLS be­
gan estimating the cost of wage and benefit
(i.e., “ package” ) changes in a limited number
of key settlements in 1964. The work was ex­
panded the following year and, since 1966, the
Bureau has attempted to determine the price
of all settlements affecting 5,000 workers or
more in the private nonfarm sector. In addi­
tion, a separate series has been developed for
the construction industry, covering settlements
for 1,000 workers or more.

Description o f Series

At present, the Bureau publishes two sets of
data on wage-benefit decisions. One shows the
annual rates of increases in settlements
reached in a given time period and scheduled to
go into effect at any time during the term of
the agreements. The other is limited to the
changes set for the first 12 months of the
agreements.1 Published data summarize settle­
2
ments reached during individual quarters of a
year, during full years, and during the first 6
and 9 months of each year.
Frequency distributions are shown for work­
ers grouped by the size of their settlements.
In these distributions, all workers affected by
a given action are entered at the average for
the bargaining unit. The sums of the individual
settlements are averaged— both means and
medians are presented— each settlement being
weighted by the number of workers affected.
However, the pricing of individual settlements
is not disclosed.
Averages for full years are available sep­
arately for manufacturing and nonmanufactur­
ing industries. Otherwise, no industry detail is
published, except for the separate construction
industry series.
As indicated in the preceding paragraphs,
these series relate to the pricing of decisions,
i.e., they measure the effect of changes agreed
on in a given period although, considering the
general practices of negotiating multiyear
201

202

BLS HANDBOOK OF METHODS

collective bargaining agreements containing
provisions for annual (and sometimes more
frequent) improvements, the changes may be
introduced only at a subsequent date. Work has
recently been completed on the development of
data on the wage and benefit changes actually
placed in effect in specified periods, whether
as a result of current bargains, changes agreed
upon earlier but with deferred effective dates,
or the operation of cost-of-living wage es­
calator clauses.

Data Sources and Collection Methods

The terms of the settlements to be priced are
obtained primarily from secondary sources,
such as general circulation newspapers and
periodicals and union, management, and trade
publications. Collective bargaining agreements
and documents on pension and health and wel­
fare plans also are consulted. When these
sources are inadequate, direct requests for in­
formation are made to the companies and
unions involved.
Large quantities of statistical data, as well
as the settlement terms, are required. These
are needed both to determine existing em­
ployer outlays and to assay the effect on these
expenditures of agreed wage and benefit
changes. Efforts are made to use existing data.
However, when these prove inadequate, the
parties are requested to furnish data. Such re­
quests, it must be emphasized, are made to
receive specific information from which the
Bureau can price settlements; the requests are
not made to receive the parties’ own evalua­
tions of the terms of their settlements.
One of the major sources of information on
current hourly earnings is the establishment
information which BLS obtains through its
monthly employment, payroll, and hours sur­
vey. Information on current outlays for pay
supplements may be available from BLS sur­
veys of expenditures for such benefits. Perti­
nent information for estimating expenditures
for some items may be found in industry wage
surveys, e.g., extent of late-shift work and
occupational employment distributions. Annual



financial reports filed with the Department of
Labor under the Welfare and Pension Plan
Disclosure Act provide useful material. Not all
the sources tapped are governmental; for in­
stance, information on workmen’s compensa­
tion insurance rates is reported by the National
Council on Compensation Insurance.

Sampling and Estimating Procedures

As was noted earlier, the Bureau attempts to
cover all settlements for 5,000 workers or more
(1,000 or more in construction) in the private
nonfarm sector. Substantially all such settle­
ments come to the attention of the Bureau and
are included in its series, it is believed.
Discussion of procedures for pricing individ­
ual settlements 3 may be centered around three
questions: (1) What items in a collective bar­
gaining settlement are to be priced? (2) How
are the costs of these items to be determined?
(3) How are these costs to be expressed?
Coverage. Little evidence is needed to demon­
strate that many terms of a union-management
agreement, not merely the wage and benefit
provisions, may affect directly or indirectly an
employer’s costs. For example, one of the socalled “ noneconomic” terms of an agreement—
seniority— may have a limited influence on costs
through its effect on employee efficiency.4 How­
ever, such an item is essentially not meas­
urable.
Consequently, the BLS program is confined
to the wage-benefit component of collective
bargaining agreements, i.e., to the effect of
settlements on employer outlays for employee
compensation. This component, clearly, is of
major significance in its own right. Included
are items such as changes in wage rates; modi­
3 Production of statistical series merely entails grouping
and averaging the pricing of individual settlements. Published
distributions show the number of workers under known
settlements for which data were insufficient to permit pricing.
4 Almost inevitably a tendency has developed to regard all
contract changes as serving to increase employer payments.
Although this undoubtedly is the common result, some
changes, even when proposed by unions, may lower costs.
Union-management cooperation schemes have at times pro­
vided examples.

MEASURING COLLECTIVE BARGAINING SETTLEMENTS

fications in premium pay, bonuses, paid leave,
and severance pay; and adjustments in em­
ployer payments for pension, for health and
welfare, and for supplemental unemployment
benefits, excluding the costs of administering
these benefits. Also included are changes in
formal contract provisions specifying paid
time for clothes change, washup, and lunch
periods.
Excluded are informal modifications of un­
written rest-period practices; items related to,
but not normally considered part of, compensa­
tion— such as per diem payments, moving ex­
pense reimbursements, and payments for safety
clothing; and provision of facilities or services
such as parking lots and health units, the costs
of which often are charged to capital rather
than labor accounts.
Determination of costs. Since a value is placed
on settlements at the time they are reached,
the costs attributed to them obviously are esti­
mates of outlays to be made in the future; they
cannot be taken from employers’ accounting
records. The estimates are made on the assump­
tion that conditions existing at the time the
contract is negotiated will not change. For ex­
ample, estimators assume that methods of
financing pensions will not change, and that
expenditures for insurance will not change ex­
cept as a result of altered benefit provisions or
modified participation because of changes in
company contributions. They also assume that
the composition of the labor force will not
change.
In this regard, except for any guaranteed
increases, which are treated as deferred adjust­
ments, possible wage-rate changes as a result
of cost-of-living escalator clauses are excluded
because of difficulties in predicting movements
of the Consumer Price Index for 3 years—the
time span of many collective bargaining agree­
ments. Thus, the Bureau prices the wage and
5 The series on wage and benefit changes actually placed
into effect includes escalator adjustments, since the issue of
prediction of CPI movements does not arise.
8 The terms “roll-up” and “bulge,” among others, also are
used to express the same idea.
7 It varies from this ratio to the extent that there are limits
on earnings that are subject to social security taxes.




203

benefit changes that would go into effect if the
price level were to remain stable.5
Nevertheless, package estimates do attempt
to measure the costs associated with actual
characteristics of the work force affected by
the settlements, not the costs for some hypo­
thetical employee group. Attempts to base esti­
mates on the actual age, length of service, sex,
and skill characteristics of the workers involved
recognize that the choice in incorporating
alternative benefit changes into contracts is
affected by their costs, which, in turn, are
affected by the character of the work force.
For example, an extra week of vacation after
15 years of service will cost very little when
only 10 percent of the workers have that much
service, but will cost about 1 percent of
straight-time hourly earnings when half of the
workers have been employed for 15 years or
more.
As a rule, indirect effects of settlements are
ignored: factors such as possible extension of
settlement terms to nonunion workers in the
same firm or to members of other bargaining
units. Similarly, the cost of providing length­
ened vacations is measured by the wages and
salaries paid for the additional time o ff; costs
of hiring vacation replacements, if necessary,
are not considered. Moreover, effects on unit
labor costs, which involve consideration of em­
ployee efficiency as well as employer payments,
are disregarded.
However, “ creep” 6 is accounted for. Creep
reflects the fact that an increase in wage rates
will have a secondary effect on employer costs
through its effect on outlays for benefits. A
20-cent-an-hour wage increase will affect not
only straight-time wage rates but also supple­
mentary payments governed by wage rates—
such as overtime premiums, leave payments,
pension benefits related to salary level, and
social security payroll taxes (if earnings are not
at or above the statutory maximum tax base).
Creep is taken into account by raising each
wage increase provided by the new contract by
a creep, or loading, factor. This factor is es­
sentially the ratio of current hourly expendi­
tures on benefits that vary with wage levels to
current straight-time hourly earnings.7

204

BLS HANDBOOK OF METHODS

Many items in a collective bargaining agree­
ment are priced without difficulty. This is
particularly true when settlement terms are
expressed as cents-per-hour adjustments, e.g.,
a 20-cent-an-hour general wage increase 8 or a
5-cent increase in employer contributions to a
health and welfare fund. These stipulated
cents-per-hour figures are utilized as the costs
of the settlement provisions.9 Percentage wage
adjustments are converted to cents-per-hour
figures on the basis of current average straighttime hourly earnings. Although less direct, the
cost of an additional holiday is estimated ade­
quately by prorating 8 hours’ average pay (if
the normal workday is 8 hours) over the num­
ber of annual working hours per employee.
The cost of an additional week of vacation for
25-year employees is estimated similarly, but
one must know the number of employees with
the required seniority.
Other settlement terms are more difficult to
price. For example, the cost of an unfunded
severance pay plan depends on the frequency
of layoffs as well as on plan provisions.
Estimates of such frequencies are at best
hazardous. Pension improvement costs are
particularly difficult to estimate because of the
considerable discretion employers often have
in funding their obligations.1 The general ap­
0
proach followed by the Bureau is to assume
that a given pension benefit increase will raise
existing expenditures for current service pro­
portionately. Since employer contributions for
pensions frequently vary widely from year
to year, outlays in several past years are
examined to develop a measure of current pay­
ments.
Under the BLS framework, estimates con­
cerning most provisions are of actual cash
outlays to be made by employers. However, in
the case of improved paid leave provisions, a
change may entail time off for workers, but
not additional cash payments by the employer.
8 Where appropriate, the effect of a general wage increase
on incentive earnings is included in the measurement of the
wage increase.
9 Since estimates are on a cents-per-hour-worked, rather
than per-hour-paid-for, basis, agreed-upon increases are ad­
justed if they are on an hours-paid-for basis.
1 Employers’ contractual obligations commonly are to pro­
0
vide given levels of pension benefits, rather than to expend
specified amounts of money.




However, payment per hour worked will rise
and this change is taken as the cost effect of
the settlement provision.
In case of a reduction in the basic workweek,
the increase in hourly rates needed to maintain
weekly pay is the major item BLS prices. To
some extent, a reduced basic workweek may be
accompanied by additional overtime work.
However, unless this overtime is provided spe­
cifically in the agreement, it is ignored in the
cost estimate.
Increases in hourly pay rates are not the
only cost effects considered in this instance.
Even if there is no change in total employer
outlays for particular pay supplements but the
contract provides for reduced hours, the out­
lays for them per hour worked will rise and
affect the cost of settlement.
Expressing the costs. The total cost of a given
settlement is obtained by adding up the centsper-hour-worked costs of each of the individual
wage or benefit changes. This sum is then
expressed as a percent of pay, as this ratio
facilitates inter-company comparisons by elimi­
nating influences of payroll size and wage level.
Furthermore, since economic studies generally
emphasize relative rates of change in statistical
series, percent-of-pay costs can be integrated
into broad economic analyses.
Expression of package costs as a percent of
pay requires estimation of an appropriate base
(the denominator of the ratio) as well as the
cost of the settlement terms (the numerator).
The base used by the Bureau consists of cur­
rent outlays per hour worked for all items of
employee compensation, as defined, plus em­
ployer expenditures for legally-required social
insurance. The latter is part of employee com­
pensation, although not subject to change
through collective bargaining.
Since collective bargaining agreements gen­
erally are for 2-year periods or longer, BLS
expresses the total percent increase over the
contract term at an annual rate to permit com­
parison among agreements for differing time
spans as well as to facilitate the use of the data
in conjunction with other statistical series.
These annual rates of increase take into ac­
count the compounding of successive changes.

MEASURING COLLECTIVE BARGAINING SETTLEMENTS

In addition, the Bureau computes first-year
changes, i.e., the total change scheduled for
the first 12 months of the agreement, expressed
as a percent of current hourly compensation.
As a general rule, the first-year increase is
larger than the average annual increase over
the full term of the agreement; contracts com­
monly are “ front-loaded.”
Contracts are considered to run from their
effective dates to their termination dates. How­
ever, where wage reopening clauses are found,
the reopening date is taken as the termination
date and any agreement under the reopening
clause is treated as a new settlement.
Pricing of a collective bargaining settlement
is illustrated on the sample worksheet. This
example assumes that at the time of the settle­
ment straight-time hourly earnings averaged
$3 and that total supplementary benefits were
$1 an hour worked, providing total compensa­
tion of $4 an hour worked (the base). Also
assumed is a creep factor of 20 percent, 2,000
annual working hours per employee, and a
3-year agreement effective January 1, 1971,
providing the immediate and deferred wage
and benefit improvements shown on the work­
sheet. The settlement provides a 7.1 percent
first-year package and a 6.5 percent annual
rate of increase over the 3-year term. The
worksheet also shows the wage and benefit
gains scheduled for each of the 3 calendar
years (1971, 72, and 73), from which material
the series on changes actually placed into effect
is developed,1 and the wage-rate changes apart
1
from benefit improvements. The latter data are
computed without reference to creep and relate
wage gains to average hourly earnings rather
than average hourly compensation.

Presentation and Analysis

Press releases covering wage changes and
wage and benefit changes in major collective
bargaining settlements are issued toward the
end of the month following the close of each
1
1 Thus, the series on effective package changes essentially
is based on the pricing of items at the time settlements are
reached. The only exception is the subsequent addition of
cost-of-living escalator wage changes.




205

quarter. These releases contain preliminary
data for the first 3, 6, and 9 months of a year
and for the full year. This material also ap­
pears in Current Wage Developments (C W D ).
Final quarterly material (both for individual
quarters and the cumulative quarterly material
appearing in the press releases) is included in
the CWD article for the full year. An annual
summary also appears in the Monthly Labor
Review.
Summary data for recent individual quarters
and 4-quarter periods are shown monthly in
CWD together with other statistical series
depicting aspects of change in employee com­
pensation. The presentation facilitates analysis
of the interrelations between the series and the
divergences in their movements.

Uses and Limitations

Package cost data are used extensively by
union and management officials, for whom data
on developments in other firms and industries
often provide an important criterion for their
own deliberations. In a different vein, the data
are examined by government officials and
private analysts, concerned with the economic
repercussions of collective bargaining on the
costs of individual employers and on wageprice-employment relations within the eco­
nomic system as a whole.
The user of the data should remember that
the series does not purport to measure all
changes in average hourly expenditures for
employee compensation. Estimates are derived
under the assumption that all factors affecting
employer outlays other than contract modifica­
tions are constant.
Nevertheless, changes in the volume of over­
time and shift work, in the composition of the
work force, in the level and stability of employ­
ment, in factors affecting incentive earnings,
etc., are not unusual, and will influence outlays
for employee compensation. In some instances,
these changes are introduced by management
specifically to offset costs of new labor agree­
ments. In other cases, changes are the result of
modified production schedules or of tech­
nological developments independent of collec-

Date Priced Out
Previous Expiration Date ----Date Negotiated: -------------Effective Date of Agreement -New Expiration Date ----------

3/15/71
1 2 /3 1 /7 0
1 /15/71
1 /1 /7 1

Escal.

Clau se I n f o r .

(if

rnsT

an a ly sts

206

parkaor

worksheet

ap p lic ab le)
Increases

12/31/73

STATISTICAL INFORMATION
MEASURE:

First-Year Wages & Benefits

First-Year Wages Alone
Wages Over Life of Contract
Wages and Benefits Effective in Wages and Benefits Effective in Wages and Benefits Effective in ---

Exclude •With
creep

Wages : 5% @ 3 . 0 0
S k i l l a d i u s t - 5d f o r 20%
Impr. s h i f t d i f f .
add. 2 d h r . on 2 nd s h i f t
30% = . 6C
ad d. 3d h r . on 3rd s h i f t
20% = . 6d
2% @ 3 . 1 7 2
5% @ 3 . 2 3 5
3% @ 3 . 3 9 7
2% @ 3 . 4 9 9

1 /1 /7 1
1/1 /7 1
1 /1 /7 1
for

creep

15.00

18.00

1 .0 0

1 .2 0

1 .2 0

1.44
7.61
19.42
12.23
8.40

for
10/1/71
7 /1 /7 2
1 /1 /7 3
7 /1 /7 3

6.34
16.18
10.19
7.00

H o l i d a v s : 1 ad d.
1 /1 /7 3
$3,499 @ 8 h r s . = 2 7 .9 9 t
1972 = h r s . =
Vacations:
Impr. a v e r a g i n g 20 add.
h r s . o f v a c . / y r . f o r a l l em pl s.
1 /1 /7 2
$ 3 , 2 3 5 @ 20 h r s . = 6 4 . 7 0 a 1980 hr s=
added impact in 1973
$ 3 , 4 9 9 @ 20 h r s . = 6 9 . 9 8 a 19 7 ? h r s .
= 3 . 55d
( 3 . 55d - 3 . 2 7 d 5 =
P e ns io ns add. 5d / h r . t o fund 1 / 1 / 7 8
5d @ 2 0 8 0 h r s . = $ 1 0 4 . 0 0 A i q 7 ? =

1.42

3.27

.28
5 97

Wages and Benefits Effective in --Wages and Benefits Effective in — -

Estimate avg. c o s t o f
$1 00 a 1972 =

$100/em p l.
5.07

Wages Alone Effective in ---------Wages Alone Effective in
Wages Alone Effective in
Wages Alone Effective in
Wages Alone Effective in
Other
TOTALS-

WaSes and f r i n g e s

83,61

--56.91

Remarks:
Note:

* * For wages and b e n e f i t s , d i v i d e by the b a s e ; f o r wages a l o n g ,
d i v i d e by the A . H . E .
For th e e f f e c t i v e m e a s u r e s , d i v i d e by the a d j u s t e d base or A . H . E .
as o f the s t a r t o f the a p p r o p r i a t e y e a r .

Priced Out By:




John Doe

Over l i f e t
or con tract
* No data should be supplied in these columns. When verification is
made, the data will be supplied.

BLS H ANDBOOK OF M ETHODS

Wages and Benefits
(over life of contract)

in c e n t s

P r o v is ion

N/A

MEASURING COLLECTIVE BARGAINING SETTLEMENTS

tive bargaining, and may either add to or
subtract from the cost of the union-manage­
ment settlement. In any event, an important
influence on the level of employee compensa­
tion, social insurance taxes, is essentially out­
side the scope of the package cost estimates.
Two other factors must be considered. First,
package costs are only estimates of future
changes in employer outlays. As already em­
phasized, completely accurate estimates should
not be expected. Secondly, the data apply pri­

marily to settlements for 5,000 workers or
more.
Although package cost estimates are ex­
tremely valuable as comprehensive measures of
change resulting from union-management ne­
gotiations, to use the estimates as precise,
unambiguous, and unfailing measures of the
economic effects of collective bargaining is
adding an assignment which the data are in­
capable of fulfilling.

Technical References
Number

1. David, Lily Mary and Sheifer, Victor J. "Estimating the Cost of Collective Bargaining Set­
tlements,” Monthly Labor Review, May 1969, pp. 16-26. Reprint No. 2617.
A more detailed description of the Bureau of Labor Statistics procedures for pricing
collective bargaining settlements.
2. Sheifer, Victor J. "Th e Relationship Between Changes in Wage Rates and in Hourly Earn­
ings,” Monthly Labor Review, August 1970, pp. 10-17. Reprint No. 2688.
Compares general wage rate and hourly earnings changes in 87 manufacturing estab­
lishments over a 7-month period. The issues raised are pertinent in any consideration of
the relation between package cost estimates and actual changes in hourly compensation
expenditures.




207

-V ic t o r J. S h e i f e r




Chapter 24. Wage Chronologies and Salary Trend Reports
Background

Most Bureau of Labor Statistics series show­
ing the movement of money wages— such as
data on average hourly earnings— apply to
large aggregates of workers, e.g., all produc­
tion workers in a given manufacturing indus­
try either nationwide or in a particular State
or local area. Wage chronologies and salary
trend reports, on the other hand, apply to
more narrowly defined employee groups.
Chronologies report on wage-rate changes
made by specific employers. Except for the
study on Federal Classification Act employees,
they deal with developments under collective
bargaining agreements. In all cases they report
on supplementary benefit as well as wage
changes.
Salary trend reports present and analyze
changes in salaries of selected categories of
government employees.1 In most instances,
however, they do not contain separate data for
individual employers.
Both programs date back to the early postWorld War II period. The wage chronology
program was instituted in 1948 and the first
salary trend report was issued in 1950. To the
extent possible, material for earlier years was
included in the initial reports.

Description o f Program

Chronologies. The following 32 chronologies
currently are being maintained :1
2
1.
2.
3.
4.
5.
6.
7.
8.

Aluminum Company of America
American Viscose Division of FMC Corp.
The Anaconda Company
Armour and Company
A .T. & T.— Long Lines Department
Berkshire Hathaway Inc.
Bethlehem Atlantic Shipyards
Bituminous Coal Mines

1 Although other BLS studies report on salary trends, they
do not have this specific title. The time series in these studies
are often by-products of repetitive Bureau survey activity.
2 Several chronologies have been discontinued, generally
either because of fragmentation of bargaining units, declining
importance, or because standardization of collective bargain­
ing eliminated the value of more than one chronology in a
given industry.




9. The Boeing Company
10. Commonwealth Edison Company
11. Dan River Mills
12. Federal Classification Act Employees
13. Firestone Tire and Rubber Company and B.F.
Goodrich Company
14. Ford Motor Company
15. International Harvester Company
16. International Paper Company (Southern Kraft
Division)
17. International Shoe Company
18. Lockheed-California Company (A division of
Lockheed Aircraft Corp.)
19. Martin-Marietta Corp.
20. Massachusetts Shoe Manufacturing
21. New York City Laundries
22. North America Rockwell Corp.
23. North Atlantic Longshoring
24. Pacific Coast Shipbuilding
25. Pacific Gas and Electric Company
26. Pacific Longshore Industry
27. Railroads— Nonoperating Employees
28. Sinclair Oil Companies
29. Swift & Company
30. United States Steel Corp.
31. Western Greyhound Lines
32. Western Union Telegraph Company

Each chronology covers either a single wagedetermination unit or a group of closely related
units. It may cover an individual company and
union (e.g., Ford Motor Company and the
United Automobile Workers), a single com­
pany and two or more unions (e.g., the Alumi­
num Company of America and the United
Steelworkers and the Aluminum Workers), a
group of employers and a single union (e.g.,
New York City Laundries and the Amalga­
mated Clothing Workers), a group of com­
panies and a group of unions (e.g., Pacific
Coast shipbuilding companies and a number of
craft unions), or a governmental body (e.g.,
the chronology covering Federal Classification
Act employees).
The program is designed to summarize long­
term wage-benefit developments in a variety
of industrial environments. Accordingly, chro­
nologies cover groups that (1) have existed
over a period of years; (2) are important in
their own industry; (3) employ a significant
number of workers; (4) are of general public
interest; and (5) are willing to cooperate with
the Bureau and for which appropriate infor­
mation is available.
209

210

BLS HANDBOOK OF METHODS

Each chronology is divided into a narrative
synopsis of the collective bargaining or legisla­
tive developments resulting in wage and benefit
changes and a tabular summary of the changes
themselves. Each chronology contains separate
tables showing general wage changes and
changes in supplementary benefits.
As used in the wage chronology program,
general wage changes are defined as upward
or downward changes that affect an entire unit
or a substantial group of workers at one time.
Excluded are adjustments in individual rates
(such as promotions, or merit or seniority in­
creases) and minor adjustments in wage struc­
ture (such as changes in individual job rates
or incentive rates) that do not have an im­
mediate and noticeable effect on the general
wage level. Because of the omission of non­
general wage changes, fluctuation in incentive
earnings, and other factors, the sum of the
wage changes listed in each chronology will
not coincide necessarily with the movement of
average hourly earnings over the same period.
The tables generally include benefits such as
guaranteed minimum earnings, shift premiums,
daily and weekly overtime, weekend premiums,
pay for holiday work, paid vacations and holi­
days, other paid leave provisions, reporting
time, waiting time, paid lunch periods, pay for
travel time, and health, welfare, and pension
benefits.
When minimum plant rates, common labor
rates, occupational wage rates, or rates for
labor grades are important in the wage struc­
ture, they are shown in chronological sequence,
parallel to the general wage changes.
Salary trend reports. These reports currently
are issued for Federal classified employees,
firefighters and police patrolmen, and urban
public school teachers. They all contain indexes
of long-term salary movements. Data on recent
and current salary levels also are provided.
Reports for firefighters, police patrolmen,
and teachers basically apply to cities of 100,000
population or more, and provide separate fig­
ures for regions and city-size groups, as well
as overall national data. Material on individual
cities, however, is not shown. Because it deals
with a single employee group, the report on



Federal classified employees contains consider­
able detail on developments in wage structure.

Data Sources

Both wage chronologies and salary trend
reports are developed primarily from published
data. Collective bargaining agreements, pen­
sion and welfare documents, and newspaper
and periodical articles provide the main sources
of chronology information. These are supple­
mented by direct requests to the parties for
information when available written records are
inadequate. Thus, most of the information used
to compile chronologies is, in one form or
another, already a matter of public knowledge.
In all instances, to avoid dissemination of
erroneous material, pre-publication drafts of
reports are submitted to the parties for their
review and comments.
Salary trend reports are prepared largely
from salary data collected by other groups.
That for Federal classified employees is based
on data published by the U.S. Civil Service
Commission in its annual report on Pay Struc­
ture of the Federal Civil Service.
Reports on teachers and firefighters and
police patrolmen are based mainly upon com­
pilations of data for individual cities made by
the National Education Association and the
International City Management Association,
respectively. These are supplemented by annual
surveys of salaries and working conditions
conducted by the International Association of
Fire Fighters and the Fraternal Order of
Police and, when necessary, by direct inquiries
by the Bureau of Labor Statistics.

Statistical Procedures

Statistical analysis in these programs is
confined largely to the preparation of indexes
of salary movements of groups of government
employees. Sampling problems do not arise
since in each case an effort is made to examine
the total universe, i.e., all Federal classified
employees, and all teachers, firefighters, and
police patrolmen in cities of 100,000 population
or more.

WAGE CHRONOLOGIES AND SALARY TREND REPORTS

Indexes generally are computed by a method
that minimizes the effect of year-to-year
changes in relative employments in the cities
or occupational categories covered. As a rule,
chain indexes are employed, i.e., the index for
the current year is obtained by adjusting the
index for the prior year by the percentage
change in average salaries over the intervening
period. Normally, to preserve a pure measure
of salary change, average salaries for each of
the two years are computed using current-year
employments as weights.

Presentation and Analysis

Wage chronologies. Wage chronologies are
published individually as BLS Bulletins and
revised periodically to incorporate material
resulting from new collective bargaining settle­
ments or legislative developments. Bulletins are
updated after every other contract settlement
or legislative enactment. Intervening develop­
ments are reported in supplements to existing
bulletins. Thus, when 3-year collective bar­
gaining agreements are negotiated, a revised
bulletin should be issued once in 6 years.
Salary trend reports. Articles covering develop­
ments for Federal classified employees and fire­
fighters and police patrolmen appear annually
in Current Wage Developments. Press releases
containing summary data for firefighters and
police patrolmen precede publication of the
articles. Salary changes for teachers are re­
ported on in biennial CWD articles, since the
basic data are issued at two-year intervals.
Reprints are available of all CWD articles. In
addition, all articles issued up to the mid-1960’s
have been collected and reprinted in the BLS
Bulletins listed in the technical references at
the end of this chapter.
Chronologies are primarily listings of wage
and benefit changes, with background material
limited to descriptions of the collective bar­
gaining or legislative processes leading up to
the changes. Greater effort is made in salary




211

trend reports to analyze the data. Background
factors are presented and the wage movements
are compared with wage changes for other
employee groups.

Uses and Limitations

Both wage chronologies and salary trend re­
ports serve two audiences; they are useful as
sources of comparative wage data for union,
management, and government officials engaged
in wage setting and as research tools for eco­
nomic analysts. Chronologies are particularly
useful for negotiators because the studies pre­
sent detailed information on developments in
units that not infrequently provide wage lead­
ership for their industries. Moreover, com­
parisons of wage and benefit changes in such
units provide valuable insights into wage set­
ting in the American economy. In addition, the
data help to explain the movements in aggrega­
tive statistics such as the Bureau series on
average hourly earnings.3 Salary trend reports
are noteworthy as one of the relatively few
sources of data on wage movements and levels
in the government sector.
Although wage chronologies describe changes
in supplementary benefits, they do not measure
the effect of these changes on employers’ hourly
labor costs. In this connection, decisions on
whether or not to adopt a given benefit change
may hinge on its cost, which may vary among
bargaining units with work forces of differing
composition.
The temptation may be to use salary trend
reports as indicators of salary movements for
government employees in general. However,
the particular groups covered by these reports
are by no means a representative sample for
this purpose.
3 For examples of the use of chronologies by economists,
see Harold M. Levinson, Postwar M ovem ent of Prices and
Wages in Manufacturing Industries, Study Paper No. 21, Study
of Employment, Growth, and Price Levels, Joint Economic
Committee, 86th Cong., 2d Sess. (Washington: 1960); and
John E. Maher, “ An Index of Wage Rates for Selected In­
dustries, 1946-1957,” Review of Economics and Statistics,
August 1961, pp. 277-282.

212

BLS HANDBOOK OF M ETHODS

Technical References
Num ber

1. Arnow, Philip, Bloch, Joseph W ., and Quant, Willis C. “ The New Wage Chronology Series,”
Monthly Labor Review, December 1948, pp. 581-583.
Describes the aims of the chronology series at the time it was introduced.
2. U.S. Department of Labor, Bureau of Labor Statistics. A Directory of Wage Chronologies,
1948-June 1969 (1969).
Lists chronologies that were maintained in 1969, the time span covered, and the place of
publication.
3 . ---------- Salary Trends: City Public School Teachers, 1925-65 (Bulletin 1504, 1966).
4 . ---------- Salary Trends: Federal Classified Employees, 1989-64 (Bulletin 1444, 1965).
5. --------- - Salary Trends: Firemen and Policemen, 1924-64 (Bulletin 1445, 1965).
Compilations of previously published salary trend articles.




— V

ic t o r

J. S h e i f e r

Productivity and Technology
Chapter 25. Output Per Man-Hour: Private Sector
Background and Description of
Measures

To provide information about the relation­
ship between productivity, prices, wages, em­
ployment, and economic growth, the Bureau
of Labor Statistics publishes indexes of output
per man-hour, compensation per man-hour,
unit labor costs, and related costs for broad
economic sectors. Measures of output per man­
hour have been developed for the total private
sector and the farm and nonfarm sectors an­
nually from 1909 to the present. Since 1947,
these data have been supplemented with com­
parable measures on hourly compensation and
related costs for these sectors as well as manu­
facturing. Post World War II indexes are avail­
able quarterly as well as annually. In addition,
to the farm, nonfarm, and manufacturing mea­
sures, annual information on productivity and
costs is available for the nonmanufacturing sec­
tor and its component major industrial sectors.
Productivity measures, first published in 1959,1
represent the culmination of a long series of
developments in productivity measurement in
the Bureau.1
2
The Bureau of Labor Statistics’ output per
man-hour measures specifically refer to the
ratio between dollar gross product (GNP)
originating3 in the private or individual sec­
tors and the corresponding hours of all persons
employed. Indexes of output per man-hour in­
dicate the relationship between output and la­
bor input (man-hours). Index changes through
time show the effectiveness of man-hours in
producing current levels of output.
Man-hours are based mainly on BLS estab­
1 Trends in Output per Man-Hour in the Private Economy,
1909-58 (BLS Bulletin 1249, 1959).
2 For a detailed description of the historical development of
the productivity measurement program of the BLS, see
Chapter 26.
3 Gross Domestic Product equals Gross National product less
the net return on foreign investments.




lishment payroll data on employment and
hours. These man-hours refer to hours of work
and paid time for vacation holidays and sick
leave.
The Bureau also develops an output per
man-hour series based on labor force man-hour
data from the Current Population Survey
(CPS). The concept underlying the labor force
man-hours is hours worked, rather than hours
paid. Theoretically, the difference between the
labor force man-hours series and the establish­
ment payroll data series is equal to paid vaca­
tion time, sick leave, and other paid leave.
However, actual differences in man-hours be­
tween the two series also reflect statistical
variation and differences in method.
Indexes of compensation per man-hour meas­
ure the hourly costs of wages and salaries, in
addition to supplemental payments, such as the
contributions of employers to social security,
unemployment insurance tax, and payment for
private health and pension plans. Measures of
real compensation per man-hour reflect the
adjustment of hourly compensation for changes
in the Consumer Price Index.
Unit labor costs measure the cost of labor
input required to produce one unit of output
and is derived by dividing compensation per
man-hour by output per man-hour. Unit non­
labor payments (costs) measure the cost of
nonlabor inputs such as depreciation, rents,
interest, indirect taxes, in addition to profit
income such as corporate profits and income
of proprietorships and partnerships.
Data Sources and Estimating Procedures

Output. The real gross national product orig­
inating in the private or individual sectors is
used to prepare output per man-hour estimates.
Gross national product is the market value of
final goods and services produced within a
certain time period. It includes purchases of
goods and services by consumers gross private
domestic investment, net foreign investment,
and purchases by Government. GNP is equal
to income received by labor and property for
213

214

BLS HANDBOOK OF METHODS

services rendered in the current production of
goods and services, in addition to capital con­
sumption allowances, indirect business taxes,
and several other minor items.
Gross national product in current dollars
cannot be used directly as the output measure
since it reflects price changes in addition to
changes in physical volume. The Office of Busi­
ness Economics (OBE), U.S. Department of
Commerce, prepares estimates of constantdollar GNP for the total private sector and
major sectors. These estimates indicate only
changes in the volume of production and are
used to develop output per man-hour measures.4
Compensation. The OBE develops compensa­
tion of employees data as part of the national
income accounts. These data include direct
payments to labor, such as wages and salaries
inclusive of executive compensation, commis­
sions, tips, and bonuses and payments in kind
which represent income to the recipients and
supplements to these direct payments. The last
item consists of employer contributions for
social insurance, private pension and health
and welfare plans, compensation for injuries,
doctors’ fees, pay for military leave, etc.
Since these compensation measures refer
only to wage and salary workers, they do not
reflect the cost of labor expended by proprietors
and unpaid family workers. Labor costs can be
seriously underestimated in sectors like the
farm where proprietor man-hours represent a
substantial portion of the labor input. The Bu­
reau, therefore, imputes a payment for labor
services of proprietors and family workers.
Compensation per man-hour of proprietors is
assumed to be the same as that of the average
employee in that sector.
Unit labor and nonlabor costs. The Bureau also
presents data on labor and nonlabor costs per
unit of output for the private sector and its
4 A detailed description of the methods and procedures of
estimating GNP and GDP in current and constant dollars is
given in the 1954 National Income Supplement to the Survey
to Current Business, U.S. Department of Commerce. Further
information on Gross Product Originating estimates for
major industry sectors is presented in the October 1962 issue
of Survey of Current Business.




major components. Unit labor costs relate
hourly compensation of all persons to output
per man-hour. Unit nonlabor costs are com­
puted by subtracting compensation of all per­
sons from current dollar GNP and dividing by
output.
Labor input (man-hours). The primary source
of man-hours and employment data is the BLS
Current Employment Statistics (CES) pro­
gram, which provides data on employment (all
employees and production or nonsupervisory
workers) and average weekly hours of produc­
tion workers in nonagricultural establishments.
Jobs rather than persons are counted, so that
multiple jobholders are counted more than
once. Weekly hours are hours paid rather than
plant hours. These statistics are based on pay­
roll records from a sample of establishments
where probability of sample selection is pro­
portionate to the establishment size; large
establishments (relative to the sector) fall
into the sample with certainty. Data on em­
ployment, hours, and earnings are collected
monthly; however, the reference period for
these data is the payroll period including the
12th of the month. Methods are described in
Chapter 2. Establishment data are published
monthly in Employment and Earnings and in
an annual summary.
Since the CES includes only nonfarm wage
and salary workers, data from other sources
(National income accounts or the CPS) are
used for the uncovered sectors (farm, proprie­
tors, unpaid family workers, and private
household workers) to develop employment and
man-hour estimates for the total private sector.
Separate estimates for employment and man­
hours paid are developed for each major indus­
trial sector and these are aggregated to total
private and private nonfarm levels. Each man­
hour is treated as a homogeneous unit; no dis­
tinction is made between workers who have
different skill levels or rates of pay.
In the manufacturing sector, separate esti­
mates for production and nonproduction
worker man-hours are derived and then aggre­
gated to the manufacturing total. Production
worker and nonproduction worker employment
and production worker average weekly hours

OUTPUT PER MAN-HOUR: PRIVATE SECTOR

are taken directly from CES data. Average
weekly hours for nonproduction workers are
developed from BLS studies of wages and sup­
plements in the manufacturing sector which
provide data on the regularly scheduled work­
week of white-collar employees.5
For nonmanufacturing sectors, employment
and weekly hours paid are taken from the pay­
roll series. Although payroll average weekly
hours data refer to nonsupervisory workers for
man-hours computation, the assumption is that
the length of the workweek in each nonmanu­
facturing industry is the same for all wage
and salary workers.
When establishment employment data are
not available (proprietors, unpaid family
workers, and private household workers),
either labor force data (CPS) or national in­
come employment data are used. On the other
hand, average weekly hours are based on labor
force data for hours worked rather than hours
paid. However, persons who have a job but are
not at work are assumed to have been paid for
an average workweek in the sector where the
job was located.

Analysis and Presentation

Indexes of output per man-hour show
changes in the ratio of output to labor input
(man-hours). These indexes relate output to
man-hours but should not be interpreted as
representing labor’s sole contribution to pro­
duction. Rather, they reflect the interaction of
many forces in addition to labor and skill, such
as changes in technology and increased capital
investment per worker.
BFor manufacturing nonproduction workers, average hours
are estimated as follows: Estimates of vacation time, holidays,
paid sick leave, and personal time off are subtracted from
an estimate of scheduled annual hours paid nonproduction
workers: scheduled annual hours are derived by extrapolat­
ing the 1959 level of scheduled weekly hours with data from
BLS Area Wage Surveys, and then multiplying by the num­
ber of workweeks in each year; the level of scheduled
weekly hours for nonproduction workers was calculated from
data collected by BLS for the study, Employer Expenditures
for Selected Supplementary Remuneration Practices for
Production Workers in Manufacturing Industries, 1962, BLS
Bulletin 1428 (1965): estimates of vacation time, holidays,
sick leave, and personal time off are derived primarily from
data from Area Wage Surveys and Social Security A d ­
ministration studies.




215

For economic aggregates like productivity
in the private sector, changes over time reflect
movements within the various component in­
dustries as well as shifts in the relative im­
portance of each of the industries. For example,
increases in output per man-hour are influenced
not only by the increments or decrements in
the component sectors but also by the shift
from low to high productivity industries.
Within industries, other forms of shifts also
take place and are not accounted for ade­
quately. In output measures, for example,
changes in income and taste may be reflected
in shifting consumption patterns to higher
quality goods or to services rather than goods.
Short-term movements in productivity and
unit labor costs often result from cyclical varia­
tion which tends to distort the underlying
relationship between the numerator and de­
nominator of the ratio. For example, because of
market imperfections, an employer often has
difficulty in adjusting his staff immediately to
changes in production. Sharp increases or de­
creases in demand usually are followed by
erratic movements in productivity and unit
labor costs. When longer periods are analyzed,
however, these fluctuations tend to iron out
and a secular trend is more apparent.
Long-term productivity trends tend to move
more smoothly over time and present a more
stable picture of historical patterns than short­
term movements. The latter tend to reflect
temporary changes in level of demand, utiliza­
tion of productive resources, and other short­
term phenomena, such as strikes, which affect
current conditions but should not be expected
to continue into the future. The period selected
for analysis may affect these long-term trends,
so that some element of judgment is present in
any analysis of trend.
Indexes of output per man-hour, compensa­
tion per man-hour, and related cost data are
published quarterly in the BLS press release,
Productivity and Costs. Historical indexes of
these data are available on request. Detailed
procedures for developing output per man-hour
indexes appear in Trends in Output per ManHour in the Private Economy, 1909-58 (BLS

BLS HANDBOOK OF METHODS

216

Bulletin 1249, 1959). Indexes of output per
man-hour and related cost data appear in Em­
ployment and Earnings, the Monthly Labor
Review, the Handbook of Labor Statistics, The
Economic Report of the President, the Man­
power Report of the President, and the Sta­
tistical Abstract of the United States.

in hourly compensation tends to increase unit
labor costs, and an increase in output per man­
hour tends to reduce these costs. Therefore,
through its relationship to unit labor costs, out­
put per man-hour is a crucial element in the
wage-price relationship. It indicates the extent
to which compensation gains can occur without
putting pressure on prices or reducing profits.

Uses and Limitations

Certain characateristics of the output per
man-hour and related cost data should be
recognized when applying them to specific
situations. First, the data reflect not only
changes in various component industries but
also changes in the relative importance of these
industries. Second, these measures represent
the culmination of a chain of economic events.
Data for a single time period often do not
show direct causal relations and should be
interpreted in light of previous economic events
as well as current conditions. Third, underlying
concepts and data available for estimation limit
to some extent measures of productivity, out­
put, compensation, and employment. In fact, in
several sectors where output data are difficult
to obtain, output changes are equal to employ­
ment changes. This definition understates pro­
ductivity growth. Consequently, the meaning
of these measures should be interpreted with
caution because of conceptual and practical
limitations and statistical errors which can
arise in any numerical series.

Data on output per man-hour, compensation
per man-hour, and related costs are designed
for use in economic analysis and public and
private policy planning. These data apply in
areas such as wage determination and analysis
of prices and living conditions.
An especially relevant use of output per
man-hour and related costs is the relationship
of productivity, wages, prices, profits, and costs
of production. Within the framework of na­
tional income and product accounting, gross
national product for a specific time period
represents the market value of all final goods
and services produced or the sum of all costs
of production— compensation, profits, deprecia­
tion, interest, etc. Unit labor costs, or compen­
sation per unit of output, represent a major
portion of total unit costs and reflect the com­
bined effect of changes in output per man-hour
and compensation per man-hour. An increase

Technical References
Number

1. Edward Denison, W hy Growth Rates Differ; Sources of Economic Growth, Washington, D.C.,
The Brookings Institution (1967).
A study of output and productivity growth in 9 Western countries. Includes a discus­
sion of factors affecting productivity growth and the way varying effects of these factors
can be attributed to differential growth rates between countries.
2. Kendrick, John W ., Productivity Trends in the United States (National Bureau of Eco­
nomic Research, 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 measure­
ment. 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, Input, and Productivity Measurement, (Studies in Income and Wealth, Vol. 25),
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.




OUTPUT PER M A N -H O U R : PRIVA TE SECTOR

Technical References— Continued
4.

5.

6.

7.

8.

9.

National Bureau of Economic Research, Conference on Research in Income and Wealth, “ In­
dustrial Composition of Income and Product,” (Studies in Income and Wealth, Vol. 32),
New York, Columbia University Press (1968).
A selection of papers concerning the development of output, input, productivity and
cost measures for individual industrial sectors.
National Bureau of Economic Research, Conference on Income and Wealth, “ Production and
Productivity in the Service Economy,” (Studies in Income and Wealth, Vol. 34) New York,
Columbia University Press (1969).
A collection of papers concerning the concepts, definitions, procedures, and data limita­
tions of measuring output, input and productivity in service producing industries.
U.S. Department of Labor, Bureau of Labor Statistics, Trends in Output per Man-Hour in
the Private Economy, 1909-1958, (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.
U.S. Department of Labor, Bureau of Labor Statistics, Productivity: A Bibliography (Bul­
letin 1514,1966).
A collection of nearly 600 references concerning productivity and productivity measure­
ment. Each reference includes a brief annotation giving the gist of the subjects covered.
Mark, Jerome A ., Wage-Price Guidepost Statistics: Problems of Measurement, American Sta­
tistical Association, Proceedings of the Business and Economic Statistics Section (1968).
A paper describing some of the problems of developing the measures which were used
in the specification of the guideposts.
Ziegler, Martin, “ Productivity in Manufacturing,” Monthly Labor Review, October 1967.
An analysis of the effect of changes in capacity utilization on productivity.




— Shelby A . H erm an

217




Chapter 26. Output Per Man-Hour Measures: Industries
Background

Studies of output per man-hour for individ­
ual industries have long been a part of the BLS
program. 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 re­
mained an important focus of the BLS program
throughout 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 Manufactures and employ­
ment statistics collected by the BLS.
In 1940, Congress authorized the Bureau of
Labor Statistics to undertake continuing stud­
ies of productivity and technological changes.
The Bureau extended earlier indexes of output
per man-hour developed by the National Re­
search Project of the Works Projects Admin­
istration, and published measures for selected
industries. This work was reduced in volume
during World War II, owing to the lack of
meaningful production and man-hour data for
many manufacturing industries.
The advent of World War II also caused a
change in program emphasis, from problems of
unemployment to concern with the most effi­
cient utilization of scarce manpower. The BLS
undertook a number of studies of labor require­
ments for defense industries, such as synthetic
rubber and shipbuilding. After the war, the
industry studies program resumed on a regular
basis, and was supplemented by a number of
industry studies based on the direct collection
of data from employers. Budget restrictions
after 1952 have prevented the continuation of
direct collection of data. Consequently, 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 produc­
tivity has grown, and there has been greater
recognition that increases in output per man­
hour are important indicators of economic
progress and the means to higher levels of in­
come, rather than merely a threat to job oppor­
tunities.
The Industry Studies program covers a
variety of manufacturing and nonmanufactur­
ing industries. For these industries, indexes of
output per man-hour, output per employee and
the related data on output, employment, and
man-hours are prepared and published on an
annual basis. The indexes are generally avail­
able for most years from 1947 to the most
recent year for which data are available, and
for many industries also for 1939.

Concepts

Industry indexes 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 physical 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 following 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 /„ represents the unit man-hour index,
7P represents the output per man-hour index,
and li and l0 denote unit man-hours expended
in the current and base periods, respectively.
For an industry producing a number of prod­
ucts— the more typical case— the unit man­
hours index is the ratio for two periods of the
219

220

BLS HANDBOOK OF M ETH ODS

total hours required for the output of a given
composite of products. Indexes of such indus­
tries vary with the composite of products spe­
cified and can take many forms. Letting q0 and
Q represent base period and current period
i
quantities of a given product, respectively, two
of these forms are:
a. Using a current period composite

Methods and Sources

/

ZqJ„
Output Per Man-Hour

b. Using a base period composite
SqJ*

Zqn
l„
An index constructed according to (a) com­
pares the man-hours expended in the produc­
tion of the current composite with man-hours
which would have been required to produce
the current composite in the base period. An
index constructed according to (b) compares
the man-hours required in both periods to pro­
duce the base period composite. 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
index
j _

ZLqi
Zhq,,

Output index _ Unit man-hours
’ (Laspeyres)
index (Paasche)
SLffi

_

2 l<qi

SJotfi
Man-hours Output index Unit man-hours index
index -s- (Paasche)
= (Laspeyres)
u~

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 weights.

'

2.l„q0

~

~ tL q 7 ~

tL q 0

~

~ tlq 7

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 establish­
ments classified in the industry, in producing
the base period and current period composites.
As can be seen in the formulas, the appro­
priate output index is one which compares the
quantities of the various products in the cur­
rent and the base periods, each weighted by
the man-hours expended per unit produced in
a given period. A current period weighted



The Bureau of Labor Statistics computes an
index of output per man-hour by dividing an
output index by an index of aggregate man­
hours. Measures are prepared separately re­
lating output to (a) all employee man-hours,
(b) production worker man-hours, and (c)
nonproduction worker man-hours. (The stan­
dard definitions of production workers and
nonproduction workers are used.) Three cor­
responding measures also are computed relat­
ing output to the number of employees.
Output

BLS industry output indexes are based
primarily on the physical output of the prod­
ucts of the industry combined with fixed period
weights. However, the availability of quantity
data on physical output varies among indus­
tries, and, for manufacturing and mining in­
dustries, may vary depending on whether the
data are for a year when a Census was con­
ducted or for a noncensus year. For manu­
facturing and mining industries, quantity data
on physical output are usually most compre­
hensive 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 in­
dexes 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 subse­
quently are adjusted to the benchmark levels
for the census years.

OUTPUT

PER

M A N -H O U R

Weights. The mathematical form of the output
index implies use of unit man-hour weights,
and such weights are used whenever possible.
These weights are derived from special surveys
or from data published for specialized estab­
lishments in the Census of Manufactures. In
some industries, however, unit man-hour infor­
mation is not available for individual products.
Consequently, the BLS uses substitute weights
which are assumed to be proportional to unit
man-hour weights. Usually these are unit value
weights. Unit value weights are computed from
Census or survey data on the quantity and
value of shipments of the primary products of
the industry. The introduction of these substi­
tute weights results in an industry output per
man-hour index which reflects shifts in value
per man-hour of the various products 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
introduced by the use of unit value weights
is not known. The index is equivalent to one
weighted with unit man-hours if the unit man­
hours and unit values among the products are
proportional or if there is no correlation be­
tween the relative change in quantity and value
per man-hour1 There is evidence that unit
values are fairly reliable 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 considerably smaller
than the error in the weights themselves.
In some industries, unit value weights for
specific products and unit man-hour weights
for product groups are used at different stages
in constructing the industry output indexes.
When this procedure is used, the individual
products are first aggregated into primary
product group indexes with unit value weights.
1 See Irving H. Siegel, “Further Notes on the Difference
Between Index Number Formulas,” Journal of the American
Statistical Association, December 1941, pp. 519-524.
2 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 shipments of the primary products made by
plants classified in the industry as a percent of the total ship­
ments of the industry’s primary products made by all pro­
ducers, both in and out of the specified industry.




M EASURES:

IN D U S T R IE S

221

These indexes in turn are combined into an
industry output index with primary product
group man-hours. The primary product group
man-hours relate to a base period, as do value
weights.
To obtain primary product group man-hour
weights, total man-hours for plants specializing
in each primary product class are derived from
published census data on production worker
man-hours and nonproduction worker employ­
ment. These are supplemented by unpublished
BLS estimates of nonproduction worker man­
hours. (See page 223 for the procedures used
to estimate nonproduction worker man-hours.)
Ratios of man-hours to value of shipments are
multiplied by the corresponding value of pri­
mary products shipped by the entire industry
to yield the estimated primary product group
man-hour weights. This procedure assumes
that the man-hours per dollar for each product
class shipped by the whole industry are the
same as those for plants specializing in the
product group. This procedure is used only
when the “ specialization” and “ coverage”
ratios of the industry are high and specializa­
tion data for all or most of the product groups
are available.1
2
Most published industry indexes have used
1947 weights for 1947-58, 1958 weights for
1959-63, 1963 weights for 1964-67, and 1967
weights for years after 1967. The Bureau
policy is to revise the weights as more current
data become available from the periodic
censuses.
Benchmark indexes. For most manufacturing
and all mining industries, indexes reflecting
changes in output between census years are
constructed. These are called benchmark in­
dexes.
For manufacturing industries, benchmark
indexes are developed using the following pro­
cedure : Price indexes for each primary product
class are developed from data on the value of
each individual product within the class
whether made in the industry or elsewhere.
Wholesale price indexes are used wherever
possible to convert the product values to con­
stant dollar estimates. If a wholesale price
index is not available, a price index is de­

222

BLS H ANDBOOK OF M ETH ODS

veloped using both the quantity and value data
reported for the product in the Census of
Manufactures. The primary product class
price indexes are derived from the sum of the
current dollar values and the sum of the con­
stant dollar values.
These “ wherever made” primary product
class price indexes are used to deflate the value
of primary products produced only by the in­
dustry. This procedure assumes that the price
movements of the primary products within the
industry are the same as the price movements
for all primary products wherever made. These
constant dollar values are related to cor­
responding base year values in order to derive
separate primary product indexes within the
industry.
These separate primary product indexes in
turn are combined with man-hour weights to
derive the total industry primary product out­
put index. The index of primary products of
the industry is multiplied by a “ coverage” ad­
justment to represent the total output of the
industry. This “ coverage” adjustment is the
ratio of the index of value of industry ship­
ments (after inclusion of net additions to in­
ventories) to the index of value of shipments of
primary products. The final industry output
index thus reflects inventory buildups and
changing proportions of secondary products.
For the mining industries, benchmark in­
dexes are computed from unweighted tonnage
data as reported in the Census of Mineral
Industries.
Annual Indexes. Annual output indexes are
*
constructed by the following described pro­
cedures. For manufacturing and mining in­
dustries, the annual indexes are adjusted, if
necessary, to the levels of the benchmark in­
dexes previously described. The adjustment
factors for 2 census years are used to de­
termine the adjustment factors for the inter­
vening years by linear interpolation.
1. Physical output. Most annual output
indexes are based on physical quantities of
products combined with fixed-period unit man­
hour or unit value weights. The basic quantity
data are generally primary products of an in­
dustry classified into product groups; the



greatest amount of detail available is used. The
quantity data relate to primary products
“ wherever made” and in some cases to ship­
ments of the products.
The Bureau’s annual measures of production
are constructed from data on physical quanti­
ties 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 well-defined industries with
a relatively homogeneous output.
2.
Deflated value. When adequate annual
physical quantity data are not available, in­
dexes are derived from data on the value of
industry output, adjusted for price change.
Since the adjustment 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 com­
bined 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 current periods.3
Often data on value of production are not
available and data on value of shipments must
be used. In this case, data on value of ship­
ments for each year are divided by an industry
price index representing the average annual
price for the year. Beginning- and end-of-year
finished goods and work in process inventories
are also deflated. The estimated value of ship­
ments in constant dollars is then adjusted by
the net change in inventories, also in constant
"F o r example:
Value Index _i_ Price index (Paasche) = Output index
(Laspeyres)

SPiQ,

.

ZP.Qi

_

ZPoQt

2P„Q„

'

2 P „Q<

_

S P 0Q0

where p , 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 industries where
this measure is used, and quantity data are usually available
for the base year only, so that the deflated value indexes em ­
ployed usually take the following form:
Value Index

SPiQi
2 P„q„

Price index

^

(Laspeyres)

2P(q„
2P„q0

_
~
~

=

Output index
(Paasche)

SP«3i
2PiQ„

OUTPUT

PER

M A N -H O U R

dollars, to yield an estimate of the constant
dollar value of production.
Sources. 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
the mining, cement, coke, and metals indus­
tries. Other important Government sources
include the U.S. Department of Agriculture,
the Fish and Wildlife Service, U.S. Department
of the Interior, the Interstate Commerce Com­
mission, the Internal Revenue Service, and the
Civil Aeronautics Board. Important sources of
trade association data include the Textile Eco­
nomics Bureau, Inc., National Association of
Hosiery Manufacturers, Inc., National Canners
Association, Rubber Manufacturers Associa­
tion, and the American Iron and Steel Institute.
For deflated value series, industry price in­
dexes are derived from wholesale price indexes
published by the Bureau of Labor Statistics.
Man-Hours

An index of man-hours is computed by divid­
ing the aggregate man-hours for each year by
the base-period aggregate. Man-hours are
treated as homogeneous and additive with no
distinction made between hours of different
groups of employees. Data on changes in qual­
itative aspects of man-hours, such as skill,
efficiency, health, experience, age, and sex of
persons comprising the aggregate, are not used
and generally not available. Man-hours indexes
are developed for all employees, production
workers, and nonproduction workers.
Sources. Industry employment and man-hours
indexes are developed from basic data compiled
by the Bureau of Labor Statistics or the Bu­
reau of the Census. For most private nonagricultural industries (including manufacturing),
the Bureau of Labor Statistics publishes em­
ployment and average weekly hours data for



MEASURES:

IN D U S T R IE S

223

production or nonsupervisory workers and em­
ployment data for all employees. For manu­
facturing industries, the Bureau of the Census
publishes employment and aggregate man­
hours data for production workers and em­
ployment 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. Over­
time and other premium pay hours are included
on the basis of actual time at the plant. In
contrast, the Bureau of Labor Statistics data
include time for paid vacations, holidays, and
sick leave, as well as plant man-hours. Differ­
ences 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
available 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 preferred.
Nonproduction Worker Hours. While both the
Bureau of the Census and the BLS provide
data on production worker man-hours, neither
source provides annual data by industry on
nonproduction worker nor all-employee man­
hours. Therefore, these measures are esti­
mated.
The estimates of aggregate nonproduction
worker man-hours for the manufacturing in­
dustries are derived from published employ­
ment 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 num­
ber of work-weeks in the year times the sched­
uled weekly hours. This produces an estimate
of average annual hours paid. Estimated hours
for vacations, holidays, disability, and personal

224

BLS HAN DBOOK OF M ETH ODS

time off are subtracted from average annual
hours paid, to obtain an estimate for average
annual hours worked.
Vacation and holiday trends are based on
data from various BLS surveys. Estimates of
disability time are based on studies of the De­
partment of Health, Education, and Welfare,
and data from BLS surveys. Personal time off
has been estimated as a constant from refer­
ences in relevant publications.
All employee man-hours estimates for manu­
facturing industries are derived by summing
the aggregate man-hours for production work­
ers, and the estimated aggregate man-hours
for nonproduction 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 produc­
tion of primary or secondary products. Annual
physical output data, on the other hand, usually
include only primary products of an industry.
In addition, 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 considerable variation
from year to year in the proportion 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 industries.
The comparability of the man-hours and out­
put data is indicated by the specialization and
coverage ratios which the Bureau of the Census
publishes. All industries in the BLS industry
measurement program have high specialization
and coverage ratios.
In selecting industries for the measurement
program, attention is also given to changes in
the degree of vertical integration. Man-hours
relate to all operations performed by establish­
ments of an industry, while output usually is
measured in terms of the final product. If es­
tablishments undertake additional operations,
such as the manufacture 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 BLS ex­
amines data such as the ratio of cost of materi­
als to value of shipments for any indication of
a change in the degree of vertical integration.

Presentation

BLS indexes are published annually in the
form of a bulletin, Indexes of Out-put Per ManHour, Selected Industries. As new industry
indexes are developed, they are presented as
articles in the Monthly Labor Review. The
articles contain an analysis of productivity,
output and employment trends in the industry.
Technical notes describing the methodology
used to develop the indexes are available on
request.
Indexes of output per man-hour also are
published in the Statistical Abstract of the
United States and in the Handbook of Labor
Statistics. Some indexes for earlier years are
published in Historical Statistics of the United
States.

Uses and Limitations

Industry measures of output per man-hour
are particularly useful for studying changes in
manpower utilization, projecting future man­
power requirements, analyzing trends in labor
costs, comparing productivity progress among
countries, examining the effects of technolog­
ical improvements on employment and unem­
ployment, and analyzing related economic and
industrial activities. Such analysis usually re­
quires 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 expenditures are
necessary.
Although the measures relate output to one
input— labor time— they do not measure the
specific contribution of labor, capital, or any
other factor of production. Rather, they reflect

OUTPUT

PER

M A N -H O U R

the joint effect of a number of interrelated in­
fluences 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 labormanagement relations. Also, indexes which re­
late 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,
existing techniques cannot fully take into ac­
count changes in the quality of goods and
services produced. Second, although efforts
have been made to maintain consistency of
coverage between the output and labor input
estimates, some statistical differences may re­
main. Third, changes in the degree of plant
integration and specialization often are not
reflected adequately in the production statis­

M EASURES:

IN D U S T R IE S

tics. This may result in overstatement of pro­
ductivity gains in some years, understatement
in others. Fourth, indexes involving nonpro­
duction 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 nonproductibn work­
ers, however, have a relatively insignificant
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 necessarily applicable to any one year
or to any period in the future. Because of these
and other statistical limitations, these indexes
cannot be considered precise measures; instead
they should be interpreted as general indicators
of movements of output per man-hour.

Technical References
Num ber

1. Dunlop, John T. and Diatchenko, V. L a b o r P r o d u c tiv ity (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, P ro d u c ­
t iv i t y M e a s u r e m e n t, 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 Corol­
laries ; Alternative Productivity Concepts; Aspects of Productivity Measurement and Mean­
ing; 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,” Jou rn a l o f the
A m e ric a n S ta tistic a l A sso c ia tio n , March 1942, pp. 103-111.
The nature, use, and limitations of productivity indexes.
4. Fabricant, Solomon. “Of Productivity Statistics: An Admonition,” R e v ie w o f E c o n o m ic s and
S ta tistic s, November 1949, pp. 309-311.
The deficiencies and limitations in productivity measurement.
5. Greenberg, Leon. D a ta f o r M e a s u r e m e n t o f In d u stria l P r o d u c tiv ity in the U n ited S ta te 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 O
ffice. M e a su r in g L a b o r P r o d u c tiv ity , Geneva., 1969. Discusses meth­
odology and problems in the measurement of productivity, analysis of national series and
international comparisons of productivity measures.
7. Kendrick, John W., P o s t w a r P r o d u c tiv ity T r e n d s in the U n ited S ta te s, 1 9 ^ 8 -6 9 , National
Bureau of Economic Research, forthcoming.



225

226

BLS HANDBOOK

OF M ETH ODS

Technical References— Continued

8.

9.
10.

11.

12.

13.

14.
15.

Presents trends in productivity by industry groupings from 1948-66 with preliminary
estimates through 1969. Long term trends, patterns of productivity growth and inter­
relationships among variables are analyzed. Also included are descriptions of concepts,
methods and sources.
Klotz, Benjamin P., P r o d u c tiv ity A n a ly s is in M a n u fa c tu r in g P la n ts, U.S. Department of
Labor, Bureau of Labor Statistics, 1970.
An econometric analysis of industry production functions derived from data for over
1700 plants.
Klotz, Benjamin P. and Herman, Shelby A., P r o d u c tiv ity in the R a ilroa d h id u s tr y , U .S . De­
partment of Labor, Bureau of Labor Statistics, March 1970.
An analysis of productivity in the railroad industry using econometric techniques.
Mark, Jerome A. “Industry Indexes of Output Per Man-Hour,” M o n th ly L a b o r R e v ie w , No­
vember 1962, pp.1269-1273.
A description of the methodology used in the construction of Bureau of Labor Statis­
tics indexes of output per man-hour. Covers methods and sources, construction of produc­
tion and man-hour indexes, and limitations.
National Bureau of Economic Research, Conference on Research in Income and Wealth. O u t­
p u t, In p u t, and P ro d u c tiv ity M e a s u r e m e n t, Volume 25 of S tu d ies in In co m e and W e a lth ,
Princeton, N.J., 1961.
A collection of papers and comments devoted to an appraisal of the measurement of
output, input, and productivity.
Siegel, Irving H. C o n cep ts and M e a s u r e m e n t o f P ro d u ction and P r o d u c tiv ity , Working Paper
of the National Conference on Productivity, 1952.
The rationale and techniques of measurement of changes in the physical volume of pro­
duction and the level of productivity. Includes an extensive bibliography on production
and productivity measurement.
U.S. Department of Labor, Bureau of Labor Statistics. In d e x e s o f O u tp u t p e r M a n -H o u r ,
M o t o r V eh ic les and E q u ip m e n t in d u str y, 1957-66, December 1968.
A study of productivity trends in the motor vehicles and equipment industry. Contains a
detailed technical note describing the methodology used to derive the measures.
United Nations. In d e x N u m b e r s o f In d u s tria l P ro d u ction , Series F, No. 1 of Studies in
Methods, New York: 1961.
Deals with the more technical aspects of compiling production indexes.
P r o d u c t i v it y : A B ib lio g ra p h y (BLS Bulletin 1226, 1958) and (BLS Bulletin 1514, 1966).
Two collections of annotated references concerning productivity and productivity mea­
surement.




— Charles W. A rdolini, A rthur S. H erman , and John L. Carey

Chapter 27. Technological Change
Background

Studies of technological changes and their
labor implications have been undertaken by the
BLS over the years for a variety of purposes.
During the 1930’s, public interest focused on
the unemployed and reports were prepared
on technological changes and displacement of
workers in various industries. During World
War II, emerging technologies were studied for
purposes of improving manpower utilization.
Beginning in the mid-1950’s, nationwide at­
tention was focused on the implications of new
developments classified under the general term
“ automation.” The BLS made a series of stud­
ies on a plant basis, in the insurance, petroleum
refining, bakery, air lines, and electronics in­
dustries, to explore 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 in­
dustries.
These studies formed the basis, beginning in
the early 1960’s, for a more systematic investi­
gation of future changes. Research now under­
way pinpoints technologies which will become
increasingly important over the next decade
in key industries and attempts to provide
advance information about their manpower
implications.

Description o f Studies

The Bureau’s research program on tech­
nological change involves a variety of reports
and studies of different degrees of detail and
approach. The current program thus provides
detailed case studies of changes within a single
plant or office; summary reports surveying
trends in major industries; detailed industry
studies; and studies of major technological
innovations, such as computers, that affect
workers in different industries.1
1 For discussion of various research methods used in study­
ing automation, see technical references 1 and 2.




Case Studies

BLS case studies provide detailed informa­
tion on various aspects of adjustments made to
technological change. The framework of such
studies is a single plant or office. A typical case
study covers such topics a s: management’s
objectives and results of introducing technolog­
ical change at the plant or office; exent of
displacement and reassignment of employees;
practices regarding transferring, retraining,
and selecting employees for new occupations;
characteristics of employees whose jobs were
eliminated and who were assigned to new posi­
tions ; and implications of automation for older
employees. Both qualitative and quantitative
data are presented.
The case study approach has also been used
to investigate in detail special aspects of ad­
justments to technological change. Among the
topics covered by case studies are: the per­
formance of older workers in industry retrain­
ing programs, as shown in four plants in
different industries; manpower planning pro­
cedures followed in connection with a series of
technological changes at a public utility and
in telephone offices; the adjustments to elec­
tronic data processing in a government agency;
experience under an adjustment program in
the railroad industry; post-layoff work ex­
perience of displaced workers in plants in dif­
ferent industries and regions; and job redesign
for older workers at different plants. In these
studies, information about single plants is
presented, as illustrative, rather than as rep­
resentative cases.
Reviews of Technological
Industries

Trends

in

Major

To provide a broad overview of significant
trends in the economy, the Bureau prepares a
summary report, applying to key industries on
new types of machinery, processes and prod­
ucts which are believed likely to have an im­
portant effect over the next 5 to 10 years. The
industries covered comprise a cross section of
the economy and include those where the pace
may be slow as well as those where change is
227

228

BLS H ANDBOOK OF M ETHODS

rapid. The first version of this report, entitled
Technological Trends in 36 Major American
Industries, was issued by the President’s
Advisory Committee on Labor-Management
Policy in 1964. A revised edition covering 40
industries was published in 1966.2
The emphasis of the report is on technolog­
ical 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. In­
ventions 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 technolog­
ical developments, indicating insofar as prac­
ticable some economic advantages of various
types of new equipment, processes or products;
their importance in terms of the man-hours
engaged in the operations affected; estimated
extent of use currently and in 5 to 10 years,
and some factors affecting adoption such as
the volume of investment and expenditures for
research and development. The advantages de­
scribed include not only labor savings 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
Trends

on

Industry

Technological

Intensive studies are made of selected major
industries where far reaching changes, on a
large scale, are taking place, such as coal, rail­
roads, and textiles. These studies involve de­
tailed analysis of the economic implications of
major technological developments within in­
dividual industries. Factors analyzed include
investment trends and factors affecting the
* See technical reference 12.
3 See technical references 9, 10, and 12.




prospects for the diffusion of recent technolog­
ical advances, such as the structure of the
industry. Estimates are developed of the dis­
placement of present by new methods over the
next 10 years. Unit labor requirements under
new and old technologies are compared,
wherever possible. Since the focus of the study
is on the industry as a whole, data on recent
industry trends in output per man-hour, pro­
duction, 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 de­
velopments as computers, numerical control of
machine tools, materials handling equipment,
and control instruments.3 Because of their far
reaching impact, special studies have been
made of the nature, status, prospects for adop­
tion, and implications for unit labor require­
ments, occupational change training needs, and
problems of industrial relations. In analyzing
their impact in different industries, differences
as well as similarities are revealed.

Data Sources and Collection Methods

A variety of data sources and collection
methods are utilized in making studies of tech­
nological change and its impact.
Personal Interviews

In making case studies, analysts personally
conduct intensive interviews with plant man­
agers, personnel directors, and other officials
who have direct knowledge of changes at their
plant. Union officials at the plant, and in some
cases, individual workers are interviewed. The
analyst uses a checklist of questions in con­
ducting informal interviews 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 meth­
ods of production.

T E C H N O L O G IC A L

CHANGE

229

Personal interviews also are utilized to help
determine industry trends. Informal inter­
views are conducted with engineers, scientists,
economists, and other experts in companies
which produce and use new technology, unions,
trade associations, government agencies, uni­
versities, etc., who have specialized knowledge
of particular technological development of in­
dustry trends. One objective in these cases is
to obtain their expert judgment about the
nature, pace of introduction, and possible im­
pact 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.

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 stud­
ies 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 equipment;
and the National Science Foundation’s esti­
mates of research and development.

Trade and Technical Publications

Because of the complexity of the subject,
relatively little use has been made of mail
surveys in studying the impact of technological
change. This technique, however, is useful in
obtaining information on a broad scale not
otherwise available, to supplement detailed in­
formation 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 employ­
ment and extent of computer use was sent to a
group consisting of all large offices, and a sam­
ple of small companies. From information re­
ceived from this screening questionnaire a
number of companies were selected to whom a
more detailed questionnaire was sent, asking
for information about computer uses, number
of employees engaged in electronic data proc­
essing, planned applications, etc.
The mail survey technique also was used in
followup surveys of workers who have been
laid off as a result of technological and eco­
nomic change.5 The names and addresses of
such workers were obtained from plant person­
nel 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 occurred, in
order to allow a period of time for some adjust-

Important sources of information concerning
technical trends are trade journals, technical
magazines and books, conference proceedings,
government hearings, and company reports.
Annual reports of leading corporations and
company house organs often contain useful
information on current technical development
in some industries. In making studies of indus­
tries, 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 publications 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
specific technological developments is fragmen­
tary and scarce. The Bureau makes use of
available data from many public and private
sources. These sources include, for example:
Office of Management and Budget; annual
inventory of computers in the Federal Gov­
ernment; Business and Defense Services Ad­
ministration : annual survey of numerically
controlled machine tools; McGraw-Hill Com­
pany : survey of computers in industry; Ameri­
can Bankers’ Association: survey of banking
automation; American Machinist Magazine:



Mail Surveys

4 See technical reference 5
r See technical reference 4.
<

230

BLS HAN DBOOK

ment 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
technological change on individual workers
within a plant, analysts sometimes can obtain
from employers’ files, data on such aspects as
the age, sex, and related personal charac­
teristics of employees whose jobs are elimi­
nated and the jobs in the plant held by each
individual affected before and after the change;
similar data are collected on individuals who
are selected for the positions created in con­
nection with automated equipment.6
Expert Review

In preparing forecasts of future technolog­
ical 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 re­
search directors, trade association officials,
technical journal editors, and university and
government specialists for their assessment 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
visited personally to review draft statements in
detail. Through this means, reports on tech­
nological 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 re­
sults in this field, it is important to keep in
mind the meaning of certain key ideas and



OF M ETH ODS

concepts. Some of the key problems of interpre­
tation and analysis in this type of research are
therefore set forth, briefly.
Definition of Technological Change

Technological change is defined broadly in
the BLS 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 popularly used as a synonym for
“ technological change,” designates, strictly
speaking, a particular type of current de­
velopment. It has been variously defined, for
example, as “ automatic operation,” “ the mech­
anization of sensory, control and thought
processes,” and “ a concern with production
processes as a system.”
While BLS studies have been concerned with
developments in “ automation,” particularly in
anticipating long-term trends in the future,
they are not the only technological changes
taking place that affect labor requirements and
industrial relations. For example, new ways of
generating power, piggybacking in transporta­
tion, use of synthetic materials in manufactur­
ing, mechanized methods of materials handling,
and faster steelmaking processes are impor­
tant technological developments, not usually
covered by technical definitions of “ automa­
tion,” but having significant manpower im­
plications.
Impact on Productivity

Since one of the principal consequences of
technological change, so far as manpower
utilization is concerned, is an increase in pro­
ductivity— i.e., output per man-hour, special
attention is given in BLS studies to analyzing
changes in industrial productivity. Such trend
analysis is a useful method of measuring the
pace of technological change. Changes in pro­
ductivity, however, also reflect changes in
capacity utilization and many other nontech­
nical factors; it is important to recognize that
BSee technical references 1 and 6 for use of plant records.
7 See technical reference 12 for further detail.

T E C H N O L O G IC A L

the productivity trend is only a partial meas­
ure of the rate of technological change.
In determining the impact of a specific
technology, BLS 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 comparisons with the estimated
average technology of the industry under
study; in others, with the best equipment that
is available; or in actual plant studies, with
the technology that is actually displaced.
It is also important to distinguish between
the impact on productivity of the operation
directly affected and on productivity of the
plant as a whole. An advanced machine tool,
for example, may result in a relatively large
reduction in unit labor requirements in the
machining operation, but would have little im­
pact on finishing, and assembling, and may
even require additional labor in engineering
and maintenance work.8 The impact on plant
productivity, therefore, would be considerably
less than the effect on productivity of any de­
partment or operation directly affected.
Impact on Employment

In assessing the impact of technological
change on employment, it is necessary to con­
sider the implications of plant manpower
policies and the effects of economic changes,
with which technical changes interact. Analysis
of the impact of technological change purely
in terms of machinery is incomplete.
At the plant level, for example, the substitu­
tion of machinery for labor may substantially
reduce job opportunities in operations directly
affected. If efforts are made, however, to elimi­
nate 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 work­
ers affected from the plant.
Moreover, the employment impact of tech­
nical change is also interrelated with the effects
of the business cycle. Thus, workers whose jobs
are eliminated by technical changes may not
be displaced from a plant until a decline in



CHANGE

231

demand results in layoffs— a long time after
the change has been made in some cases. In
the subsequent recovery, however, they may
not be hired back because their jobs no longer
exist.
Since many changes exert their effects on
employment through the competitive market,
the employment trend for the industry as a
whole must also be examined. The plant which
reduces its unit costs through technological
improvement may be able to gain a larger share
of the market and increase its employment, but
at the expense of the less technically advanced
competing plants, which may be forced to shut
down, displacing workers far from the location
of the change.
Because of the whole complex of economic
factors that operate through the market, in­
cluding changes in demand, location, foreign
competition, merger, and consumer taste, it is
very difficult’ to isolate the expanding and dis­
placing effects of technological change.
Impact on Occupations

Two aspects of occupational change result­
ing from technological changes are examined.
Changes in job structure— the distribution of
the plant or office work force by function or
broad skill grouping— are studied to determine
the extent of upgrading or downgrading. Since
the content of jobs may be altered as a result
of changes in equipment or processes, attention
also is directed to intensive before-and-after
analysis of job duties and the knowledge and
abilities required to perform these duties as
indicated by job descriptions and observation.
The content of newly created jobs, such as
programmer, also is studied and the qualifica­
tions 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
bargaining within plants. The introduction of
8 See technical reference 10 for further discussion.
" See technical references 1, 6, and 10.

232

BLS HAN DBOOK

new machinery, products, or processes often
requires movement of workers among jobs
within the plant or office by transfer or promo­
tion, the setting of wage rates, and selection of
persons for new jobs. Often the adjustment
proceeds according to rules established in ad­
vance through collective bargaining. Provisions
to assist workers whose Jobs are eliminated
include severance pay, retraining, and early
retirement. Besides analyzing the operation of
formal provisions under collective bargaining,
Bureau studies describe informal efforts to pro­
vide training, to utilize attrition, and to obtain
jobs for displaced workers elsewhere.1 The
0
limitations of these measures as well as their
advantages are important matters studied.

Uses and Limitations

BLS studies of technological change are pre­
pared as part of the U.S. Department of La­
bor’s program for carrying out the objectives
and responsibilities of the Manpower Develop­
ment and Training Act. Under this act, the
Secretary of Labor is required “ to establish
techniques and methods for detecting in ad­
vance the potential manpower impact of auto­
mation, technological progress, and other
changes in the structure of production.” As
part of such an early warning system, BLS
studies and reports of technological change are
useful to managers, union leaders, educators,
economists, government officials, and others in
planning policies to cushion the impact of
1 See technical references 1 and 6 for further discussion.
0




OF M ETH ODS

change. The study of emerging technological
trends and possible implications, moreover,
provides a basis for more valid projections of
productivity and economic growth. They also
are useful 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
assessing their appropriateness for particular
uses. In general, it is important to recognize
that judgments about the future direction and
pace of technological change and its implica­
tions are necessarily complex and difficult. The
rate of introduction 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 avail­
ability of trained workers, all of which are sub­
ject to significant variations. Moreover, since
the period of introduction generally spans a
number of years, new developments are con­
stantly 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 technolog­
ical change deal primarily with changes within
individual industries. But these changes often
involve changes in the type and amount of
goods and services purchased from other in­
dustries and could therefore have important
implications for production and employment in
industries supplying inputs. The accumulation
of information on interindustry relationships,
through the Bureau’s economic growth studies,
will provide a quantitative basis for analyzing
this aspect of technological change.

T E C H N O L O G IC A L

233

CHANGE

Technical References
Number
1. A d ju s tm e n t to the In trod u ction o f Office A u to m a tio n ,

2.

3.

4.

5.
6.
7.

8.

9.
10.

11.
12.

(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 re­
quirements.
A u t o m a t io n : A D isc u ssion o f R esea rch M e th o d s, Labour and Automation Bulletin No. 1,
Geneva, International Labour Office, 1964.
Fourteen papers on problems of research methods in studying manpower impact of
automation at the plant and industry level.
A u to m a tio n and T ech n ologica l C h a n ge (edited by John T. Dunlop). The American Assem­
bly, Columbia University, Englewood, N.J., Prentice-Hall, 1962.
Case studies of psychological impact are summarized in Chapter 3, by Floyd C. Mann.
Chapter 7, “Employment,” by Ewan Clague and Leon Greenberg, discusses problems of
measuring employment impact.
C a se S tu d ies o f D isp la ced W o r k e r s , 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 a c t o f Office A u to m a tio n in the In su ra n ce I n d u s tr y , (BLS Bulletin 1468, 1966).
Mail survey of over 400 insurance companies. Covers extent, pace, and employment im­
plications of electronic data processing.
Im p a c t o f Office A u to m a tio n in the In te r n a l R e v e n u e S e rv ic e, (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 etra in in g P ro g r a m s f o r T ech n ologica l C h a n g e, BLS Bulletin 1368 (1963).
Four case studies of experience of older and younger workers in industrial retraining
programs, based on plant records. Discusses some problems of measuring comparative per­
formance.
M a n p o w e r P la n n in g f o r T ech n ologica l C h a n g e — C a se S tu d ies o f T elep h on e O p era to rs, (BLS
Bulletin 1574, 1968).
Case studies of the manpower policies and experiences of several telephone companies in
cushioning the impact of technological change on their employees.
O u tlook f o r C o m p u te r P ro c e ss C on trol, (BLS Bulletin 1658, 1970).
A study of manpower implications of the introduction of computers to control produc­
tion processes in six process industries.
O u tlook f o r N u m erica l C o n tro l o f M a ch in e T o o ls , (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 change.
T ech n o lo g y and M a n p o w e r in the T e x tile In d u s tr y o f the 1 9 7 0 ’s, (BLS Bulletin 1578, 1968).
A study of the changes in technology in this major industry, and the impact of pro­
ductivity, employment, and occupational requirements.
T ech n ologica l T ren d s in M a jo r A m e ric a n In d u stries, (BLS Bulletin 1474, 1966).
Description of outlook for major technological developments based on a variety of data
sources and expert review.




— E dgar Weinberg




Chapter 28. Construction Labor Requirements
Background

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 employ­
ment, but also in the many manufacturing,
trade, transportation, and service industries
which furnish the materials and services re­
quired in the construction process. Because of
this far-reaching employment impact, the crea­
tion of new construction projects often is
regarded as a means of counteracting cyclical
unemployment.
To assist in assessing the extent of the impact
of construction expenditures on employment, a
series of labor and material requirements stud­
ies for different types of construction was
started in 1959. The program was established
as a result of Congressional legislation, requir­
ing 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
intermittently 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 Ad­
ministration studies published in 1940, cover­
ing Federal public works projects constructed
in the mid-thirties; and a few individual stud­
ies of specific types of construction.
The present studies include the major types
of building construction (schools, hospitals,
public and private housing, etc.) and also heavy
construction (highways, dams, etc.). However,
only one or two selected types of construction
are surveyed in any given year. Selected types
of construction are resurveyed periodically.
These resurveys may, in addition to providing
current information on labor requirements,
contribute information useful in preparing con­




struction cost indexes and estimates of changes
in productivity of on-site construction labor.

Description o f Survey

The surveys are designed primarily to deter­
mine the number of man-hours represented by
a fixed dollar volume ($1,000) of contract con­
struction. Man-hours, as defined by the surveys,
include both on-site construction employment
and the off-site employment required to pro­
duce and deliver materials used in the con­
struction. 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 construc­
tion 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 ma­
terials from raw materials to the final manu­
factured product. Man-hours are also estimated
for the employment created by overhead ex­
penditures of contractors.
Certain types of employment are not covered
by the survey. For example, no estimate is
made of the employment used in the planning
and designing of the projects studied. Also
excluded are estimates of employment required
in government and public utility agencies
which might be affected by the construction
being studied. Employment created by the re­
spending 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 cover construction which is
nonfederally assisted and/or that which may
be totally or partially financed through Federal
235

236

BLS H AN D BO O K OF M ETH ODS

funds. Although the type of construction labor
and material requirements data sought are
similar for both federally and nonfederally
aided projects, the sources for the data are
different, particularly for the on-site man-hour
information.
For the construction of nonfederally aided
projects— 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 representa­
tives from local authorities, architects, con­
tractors, and other direct participants in the
projects. These individuals provide the desired
man-hour information from payroll records,
daily work force reports, or summary time
reports.
For those projects financed wholly or in part
by Federal funds, on-site employment informa­
tion 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 estimat­
ing on-site man-hour requirements, as well as
data on wages for all hourly rated workers on
the projects. Data for on-site salaried employ­
ees, not accounted for on the payrolls, are ob­
tained 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 operations, 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 collected by the field representatives from
the individual contractors and subcontractors
engaged in the construction of the sample
projects.

Sampling and Estimating Procedures

Sampling procedures vary with the type of
construction being studied. The “ universe” of
projects for a specified study generally repre­



sents all of the projects known to have been
completed during a selected 1-year period.
The projects are then stratified into cells
having similar characateristics which may
affect man-hour requirements. Factors con­
sidered important in cell stratification include
(a) regional location, (b) metropolitan or non­
metropolitan 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
volume of each cell.
As indicated previously, the off-site employ­
ment estimates are derived from the materials
and equipment cost information obtained from
the contractors 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 a 4-digit Standard Industrial Classifica­
tion code as used in Census of Manufactures
product groups. For each of these product
groups, average amounts of material (in dol­
lars) required for each $1,000 of contract
construction cost are determined. The value of
materials is reduced by a ratio representing
the difference between valuation by the pur­
chaser and valuation 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 re­
quired because all data reported by contractors
are in purchasers’ value, and reduction to pro­
ducers’ value is necessary to obtain figures
consistent with Census data which are used
to calculate manufacturing employment.
Primary employment in manufacturing is
considered to be that required to produce the
construction bill of materials in their final
stage of fabrication. In this stage, man-hours
are developed by multiplying average pro­
1 For example, elementary or secondary schools.

C O N ST R U C T IO N

LABOR

ducers’ value of each construction material by
the ratio of manufacturing man-hours per
$1,000 of production. (This ratio is established
by using the Census of Manufactures.)
Primary man-hours in the trade, transporta­
tion, and service industries are estimated from
the difference between producers’ and pur­
chasers’ value for each construction material.
The value differences are alloted to trade and
transportation, and primary man-hours for
each component of transportation and trade
are then estimated from labor factors provided
by BLS.
Secondary employment is defined as the em­
ployment in all industries involved in produc­
tion and transportation of building materials
and equipment from basic extraction to, but
not including, the final manufacturing stage.
The Interagency Input-output Study is used to
obtain these estimates.2 For each group of ma­
terials, the interindustry study provides infor­
mation on the amount of secondary products
required from each of its 78 industry sectors.
The product data are converted to man-hours
by use of output per man-hour ratios for each
of the sectors. Adjustments for price and pro­
ductivity are made to provide estimates con­
sistent with the year of construction and bill of
materials.
For each off-site stage (primary and secon­
dary), a man-hour figure per $1,000 for the
construction being studied is obtained. When
these man-hours, plus the builders’ off-site em­
ployment, are combined 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
ers’ off-site employment is occupied not only
with the sample projects studied, but also with
other current or future projects. The estimate
of contractors’ off-site man-hours for each
$1,000 of construction contract is based on the
difference between construction worker em­
ployment and total employment in the COn­
^ e e "Input-output Structure of the U.S. Economy: 1963,”
Survey of Current Business, November 1969, pp. 16-47.




R E Q U IR E M E N T S

237

struction industry, with adjustments for
on-site supervisory and administrative employ­
ment.
Analysis and Presentation

The construction labor and materials re­
quirements studies are published in Bureau
bulletins. Summary articles, based on the sur­
vey findings, 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 data, the statistical
tabulations are supplemented with an analysis
of the various factors which apparently
affected the man-hour requirements for the
specific types of construction studied. The bul­
letins 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 re­
quirements and costs arising from differences
in design, type of structure, and regional
or metropolitan versus nonmetropolitan in­
fluences.
The reports include information and analysis
o f: on-site man-hour requirements by occupa­
tion ; the employment share of the general and
special trades contractors engaged in the con­
struction work; direct on-site wage cost; the
distribution of employment by periods of con­
struction tim e; and the cost of major materials.
When feasible, comparisons of unit man-hour
and material requirements for earlier periods
are made and analyzed.
Uses and Limitations

The results of the labor requirements sur­
veys are used by other offices of the Bureau,
other Bureaus of the Department of Labor,
other governmental agencies, congressional
committees, and industry research and trade
organizations, to assist in gauging the impact
of planned expenditures for construction on
employment, and the economy, generally. Of
special interest to market research analysts

238

BLS H ANDBOOK OF M ETHODS

and companies manufacturing materials for
use in construction, are the materials listings
per $1,000 of construction contract.
While the overall estimates of employment
are believed to be reasonably accurate, the de­
tailed data would have a wider margin of
sampling error and may be subject to other

limitations. Man-hour and material require­
ments are affected by a number of factors such
as location, size of project, type of structure,
architectural design, availability of certain ma­
terials or equipment, labor skills, and local
building codes and customs. The effects of these
separate factors cannot be isolated.

Technical References
Labor and Material Requirements Studies
Number

1.
2.

3.

4.

5.

6.
7.
8.

9.

10.

11.

(BLS Bulletin 1390, 1964).
A study of on-site and off-site man-hour and wage requirements for dredging and landtype projects in the U.S. Corps of Engineers’ civil works program from 1959 to 1960.
C o lleg e H o u sin g C o n stru ctio n (BLS Bulletin 1441, May 1965).
A report based on findings in a survey of 43 college housing projects which were ad­
ministered by the Community Facilities Administration. The survey is designed primarily
to determine the man-hours required for $1,000 of college housing construction.
C o n stru c tio n o f F e d er a lly A id e d H ig h w a y s , 1958, 1961, and 1964 (BLS Bulletin No. 299).
This study provides measures for 1958, 1961, and 1964 of the labor and material re­
quirements for federally aided highways, and separate measures of the requirements for on­
site and off-site construction. For on-site construction, the study also provides a comparison
of annual man-hour requirements for 1947-64.
F e d e r a l Office B u ild in g C o n stru c tio n (BLS Bulletin 1331, 1962).
A statistical study of on-site and off-site labor requirements for constructing 22 Fed­
eral office building projects in various localities of the United States over a 3-year period
from the fall of 1957 to 1960.
H o s p ita l C o n stru c tio n (BLS Bulletin 1340, 1962).
A statistical study of on-site and off-site labor requirements for construction of selected
public and private, profit and non-profit, general hospitals in various localities of the
United States between mid-1958 and mid-1959.
H o s p ita l and N u r sin g H o m e C o n stru c tio n (BLS Bulletin 1691, 1971).
A study similar to the 1962 study but with data shown per square foot as well as per
$1,000 of construction contract.
P r iv a te O n e -F a m ily H o u sin g C o n stru ctio n (BLS Bulletin 1404, 1964).
A statistical study of on-site and off-site labor requirements for constructing a sample
of one-family houses built in 1962 in various localities of the United States.
P u blic H o u sin g C o n stru ctio n (BLS Bulletin 1402, May 1964).
A report based on findings of a survey of 31 public housing projects which were
administered by the Public Housing Administration. Projects were selected in various
States to be representative of four broad geographical regions of the conterminous United
States.
S ch ool C o n stru c tio n (BLS Bulletin 1299, 1961).
A study of primary and secondary man-hours required per $1,000 of new school con­
struction based on contracts awarded throughout the United States for 85 elementary and
43 junior and senior high schools.
S ch ool C o n stru c tio n (BLS Bulletin 1586, 1968).
A survey of selected elementary and secondary public schools constructed primarily dur­
ing the period of 1964-65. In addition to providing information on man-hours, the study
also includes data on the types and value of materials used, wages paid, occupations dis­
tributed and use of apprentices.
S e w e r W o r k s C o n stru ctio n (BLS Bulletin 1490, 1966).
This study was designed to measure the total man-hours of labor required for each
$1,000 of new sewer facilities construction contract. The basis for this study was 138 con­
tracts for new sewer works in the years 1962-63.
C ivil W o r k s C o n stru c tio n b y th e C o rp s o f E n g in e e r s




— Martin Ziegler

Occupational Safety and Health

Chapter 29. Occupational Safety and Health Statistics
Background

Passage of the Occupational Safety and
Health Act of 1970 (Public Law 91-596)
marked a major departure in the collection of
work-injury statistics. The act is comprehen­
sive; it provides that every place of employ­
ment subject to the act shall be free from
recognized hazards which are likely to cause
death or serious physical harm. To assist in
attaining this goal, the act provides for record­
keeping and reporting procedures which will
identify the seriousness of on-the-job accident
and job related illnesses.
The Bureau of Labor Statistics has been
assigned responsibility for developing and im­
plementing the statistical program which is
required to achieve the objectives of the act.
The Bureau has been concerned for many
years with standardizing the methods for com­
piling work-injury statistics. As early as 1911
the Bureau called a formal conference to dis­
cuss the matter. The work of the conference
was continued by the International Association
of Industrial Accident Boards and Commissions,
resulting in publication of the first standardized
procedures in 1920.1 In 1937 the first workinjury standard was published by the American
Standards Association, now the American Na­
tional Standards Institute. The most recent
revision is the Standard Method of Recording
and Measuring Work Injury Experience
(Z16.1), 1967. ~
In December 1969, while Congress was con­
sidering comprehensive safety and health leg­
islation, Secretary of Labor J. D. Hodgson
(then Under Secretary of Labor) noted in a
letter to the American National Standards In­
stitute that the proposed legislation included
a national system for the collection of safety
1 Standardization of Industrial Accident Statistics (BLS
Bulletin 276, 1920).
2 Proposed National System for Uniform Recording and
Reporting of Occupational Injuries and Illnesses, American
National Standards Institute, New York, New York, 1970.




and health statistics. He requested that the
Institute evaluate whether the standard method
contained in Z16.1-1967 was appropriate for
the broad universe of employers who would be
subject to the proposed legislation. An informal
conference called by the Institute concluded
that there was cause to believe that the Z16.1
standard was inappropriate for the proposed
use and recommended that a study group be
formed to consider the matter further and, if
necessary, develop a simple method of report­
ing injuries. This recommendation was based
upon a belief that the Z16.1 standard: (1) had
grown too complex, through efforts to make
it equitable, to form a basis for a mandatory
national reporting system, (2) did not ade­
quately reflect trends in injury experience for
employers with good safety records, (3) was
not adequate for recording health experience.
A study group, formed by the Institute,
concluded that a new simple method for record­
ing and reporting occupational injuries and
illnesses was needed. The study group’s pro­
posal was published by the Institute in Decem­
ber 1970.1 The recordkeeping regulations which
2
have been issued by the Secretary draw heavily
on the proposals of the study group.
Records Required to be Kept

The Occupational Safety and Health Act of
1970 directs the Secretary of Labor to issue
regulations which require all employers sub­
ject to the act to maintain accurate records
of work related deaths, injuries, and illnesses.
However, records do not have to be maintained
for minor injuries requiring only first aid
treatment and which do not involve medical
treatment, loss of consciousness, restriction of
work or motion, or transfer to another job.
The regulations concerned with recording of
occupational injuries and illnesses (29 CFR
Part 1904) became effective on July 1, 1971.
They require employers to maintain in each
establishment a log of occupational injuries
239

240

BLS H ANDBOOK OF M ETHODS

and illness. Within 6 working days of receiv­
ing information that a recordable case has oc­
curred, the employer must enter the facts
called for in the log. At the end of each year
the employer is required to complete an an­
nual summary of occupational injuries and ill­
nesses. Information for the summary is drawn
from the log. The summary is to be prepared
no later than one month after the close of the
year and is to be posted prominently in each
establishment in a place accessible to the em­
ployees.
The log and summary must be kept available
in the establishment for a period of 5 years
following the end of the year to which they
refer. In addition, a supplementary record must
be maintained for recordable injury or illness.
A form is provided for that purpose (OSHA
Form 101) but in most cases a workmen’s com­
pensation, insurance, or internal form which
must be completed for other purposes will be
an acceptable record in place of form 101.
All logs, summaries, and supplementary rec­
ords must be available in the establishment for
inspection and copying by compliance officers
or statistical agents of the Federal or State
Governments.
Collection Methods

Collection methods are still being formu­
lated. Tentative plans call for an initial collec­
tion early in 1972 for the last half of 1971.
After that initial period, collection will be on
a calendar year basis. Collection will be pri­
marily by mail questionnaire. The respondent
will merely transcribe the injury and illness
data from the summary form to the question­
naire. The questionnaire will also seek informa­
tion about the numbers of employee-hours
worked and a statement of the product of the
establishment.
Coverage

Data will be collected from a representative
sample of establishments. The sample for the
first report period (July through December
1971) is designed to produce injury data at
most of the 2-digit SIC industry levels. The



sample for the first full calendar vear (1972)
of oneration under the act will be approxi­
mately 200,000 establishments, providing data
at the 4-digit SIC industry level in manu­
facturing and at the 3-digit level in nonmanu­
facturing.
Stale Participation

The act encourages the States to assume
responsibility for the administration and en­
forcement of occupational safety and health
laws, and for carrying out an occupational
safety and health educational program.
The Secretary of Labor is authorized to
make grants to the States to assist them in
developing and maintaining an effective pro­
gram of collection, compilation, and analysis
of work-injury and illness statistics. A con­
siderable number of States are expected to
participate in these grant programs.
Existing Work-Injury Data

Work-injury data have been collected by the
Bureau for many years. These data, which have
been collected through 1970, will not be com­
parable to data compiled under the Occupa­
tional Safety and Health Act. Methods used in
the collection of these data are described in the
previous edition of this Handbook. (See Chap­
ter 26 of the Handbook of Methods for Surveys
and Studies, U.S. Department of Labor, Bureau
of Labor Statistics, Bulletin 1458, 1966).
Description of Surveys and Methods

Programs for the collection, tabulation,
analysis, and publication of occupational safety
and health statistics are being formulated. De­
tails of these programs will be published as
supplements to the BLS Handbook of Methods
probably about mid-year 1973. These supple­
ments can be obtained after mid-1973 from
the Bureau of Labor Statistics, U.S. Depart­
ment of Labor, Washington, D.C., 20212, or
from any of its regional offices listed on the
inside back cover of the Handbook.

—Lyle R. S chauer

Due to budget cuts and personnel ceilings, Labor Developments Abroad, basic country
studies, labor digests, and country or area bibliographies will be discontinued by June
30,1972.

Economic Trends and Labor Conditions
Chapter 30. Foreign Labor Conditions, International
Comparisons, and Trade Research
Background

Almost from its inception, the Bureau of
Labor Statistics (BLS) conducted research on
labor conditions and developments abroad. The
Bureau carries on two research activities of
particular interest to users of research reports:
analysis of research data on labor in foreign
countries and the preparation of international
comparisons.1 The Bureau undertook these re­
search programs because (1) summary and
detailed information on labor conditions pub­
lished by a majority of foreign countries is not
available in English (or in any form readily
usable by U.S. employers, labor unions, Gov­
ernment officials, and others); (2) users cannot
keep pace with the variety of source materials
from nearly 200 independent countries and de­
pendencies; and (3) often, none but an expert
can judge the quality of source materials.
Description o f Reports

The foreign labor research reports published
by the Bureau are in part general and in part
statistical. The principal types may be listed
as follows:
1. Labor Developments Abroad, a monthly
publication which covers important develop­
ments in labor in foreign countries and pro­
vides data on living costs abroad, furnished by
the Allowances Program of the U.S. Depart­
ment of State.
1 Other major BLS activities in the international field in­
clude providing orientation and factual advice to U.S. policy
and program officers on labor in foreign countries, and pro­
viding training or orientation to foreign statisticians and
other foreign visitors, especially participants in exchange
programs and technical cooperation programs, on U.S. labor
statistics methods and on the economic conditions of U.S.
workers.




2. Basic country studies, primarily mono­
graphs on labor law and practice in selected
countries. These reports are factual, but not
primarily statistical.
3. Brief labor digests on single countries,
published either separately or with directories
of labor organizations in major geographic
areas.
4. Bibliographies, chiefly for a country or
an area, often with annotations or brief
summaries.
5. International comparisons presented in
feature articles and bulletins. These are mainly
statistical in nature and contain explanations
of the source data and, where possible, adjust­
ments for differences in definitions and meth­
ods. So far, principal comparisons published
have concerned hourly compensation, output
per man-hour, unit labor costs, unemployment,
and price trends.
6. Reports on trade and labor developments
in selected industries, including the relation­
ship of imports to domestic employment.
The Bureau also provides data in response
to individual inquiries from Government of­
ficials, U.S. businessmen, labor unions, and
students.
Data Sources and Collection Methods

The Bureau receives much material— de­
scriptive, statistical, and analytical— on labor
developments abroad, which it maintains in
files classified by country and by subject. The
material includes (1) current reports from
labor attaches and other officers of the U.S.
Foreign Service throughout the world; and
(2) foreign data from various sources. The
latter include periodicals and other publica­
tions issued by agencies of foreign countries
(labor ministries, bureaus of statistics, and
24 1

242

BLS HAN DBOOK

others); international organizations— govern­
mental, business, and labor— as well as the In­
ternational Labour Organization; and private
agencies like banks and economic research
organizations in the United States and abroad.
When specific supplementary information is
needed, the Bureau often requests it through
Foreign Service reporting channels.
Staff members keep informed by contact with
research institutions, with individual scholars
specializing in international labor, and with
U.S. firms employing labor abroad. In addi­
tion, they are briefed by labor attaches and
other Foreign Service officers when they
visit Washington. Also, from time to time,
members of the research staff study foreign
labor conditions at first hand by travel abroad.

Analysis and Uses

With the exception of international com­
parisons, the major focus in the presentation
of reports is geographic, although an occasional
summary by topic is used. Within an individual
country study, the organization is by subject.
For the labor law and practice series, subjects
include: (1) social, economic, and political
background; (2) manpower (including labor
force, employment, unemployment, manpower
planning and utilization); (3) labor law and
administration; (4) wages, supplementary
benefits, hours and working conditions; and
(5) labor organizations, management organiza­
tions, and labor-management relations; (6)
industrial safety and workmen’s compensation;
and (7) social insurance.
Most reports are concerned either with 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 attention is given to historical de­
velopment of long-term trends. The extent of
material covered is so great that often only the
most summary information for a given country
can be written more often than every 10 years;
many countries have never been covered in
individual monographs. Reports often contain
an annotated bibliography for readers in­
terested in more detail.



OF M ETH ODS

International comparison studies were begun
in 1960 on a regular basis. They deal with in­
dividual subjects and are primarily statistical
requiring specific information about details of
definition, and information is seldom covered
fully by regularly published sources. Therefore,
great reliance is placed by the Bureau on re­
quests to the Foreign Service and statistical
agencies abroad for information specifically
required.
Among subjects covered are international
comparisons of labor costs and productivity.
For instance, articles containing basic infor­
mation on “ The Role of Labor Cost in Foreign
Trade” and on “ International Comparisons of
Unit Labor Cost: Concepts and Methods” have
been published. Absolute comparisons have
been issued also for the iron and steel industry
in four countries, as well as trend data in index
form for all manufacturing in ten principal
trading natio*ns and in the United States. Rela­
tive labor cost trends are useful for analysis
of trade competitiveness, the balance of pay­
ments, and inflationary developments. Other
topics studied have included unemployment,
work stoppages, job vacancies, and price
trends.
The greatest problem involved in comparing
unit labor costs among countries arises from
the fact that labor cost data are available by
industry, but output is reported by product.
Since nearly all industries produce a number
of products, the question of labor cost alloca­
tion to specific products (or product synthesis)
arises.
The trend data published for all manufactur­
ing are taken, in most cases, from official statis­
tical reports issued by the individual countries.
Because the data are not entirely comparable
from country to country, and because more
than one set of data exists for most free indus­
trial countries, considerable effort is spent in
analyzing the individual series and selecting
those most consistent and valid. In most coun­
tries where it was available, the Bureau de­
cided (1) to use as the measure of labor cost
the compensation of employees in manufactur­
ing, from the national accounts, and (2) to use
as the measure of total output the real gross

F O R E IG N L A B O R C O N D IT IO N S, IN T E R N A T IO N A L C O M P A R IS O N S , A N D

in approach. The methods are often intricate,
national product originating in manufacturing.
The chief reason for this decision has been the
greater comparability and comprehensiveness
of these measures.
As regards unemployment, the interest in
international comparison centers on the per­
centage of the labor force unemployed. Data
published by the individual countries on a cur­
rent basis are of different types. However,
most industrial countries of the free world
have made one or more sample surveys of their
labor forces, using definitions and methods
fairly comparable to those in the United States.
The Bureau uses these latter data (adjusted
as carefully as feasible to uniform definitions
and brought up to date by the best data cur­




TRADE

RESEARCH

243

rently available) as the basis for international
comparisons of unemployment rates.
A program is under development for con­
ducting studies of the relationship of foreign
trade to U.S. labor. Analyses will be prepared
for selected industries in order to understand
the factors contributing to employment changes
and the effects of rising imports on employ­
ment. Since the relationship of trade to em­
ployment is usually indirect and difficult to
measure, a multiple approach will be taken,
including review of plant closings, mass layoffs
of employees, and the effects of changes in
demand, imports, and productivity on employ­
ment.
— Juliet F. K idney and John Chandler




Chapter 31. Economic Growth Studies
Background

The Bureau of Labor Statistics has de­
veloped a program of economic growth studies
aimed at providing a more comprehensive and
integrated framework for analyzing the prob­
lems of long-run economic growth in relation
to employment opportunities. Because this
program has important implications 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 Eco­
nomic Advisers; other members come from the
Bureau of Labor Statistics, U.S. Department
of Labor; Office of Business Economics, U.S.
Department of Commerce; and the Office of
Management and Budget.
A primary objective of the program is to
develop projections, under alternative assump­
tions, of the rate and pattern of economic
growth in the United States. These provide a
framework for assessing a number of impor­
tant economic problems, including problems of
manpower utilization over the next decade.

Methods

An economic growth model has been de­
veloped to serve as a tool in making economic
projections. This model begins by developing
a potential economic growth rate for the United
States. To do this, it must project the labor
force to the target year, assuming a specified
unemployment rate, and project the rates of
change in productivity and average hours paid.
This potential growth in gross national prod­
uct (GNP) is distributed among the four major
components (or demand categories) of GNP:
consumer expenditures; domestic investment;
government expenditures— Federal, State, and
local; and net foreign demand.
The next stage is to develop a projection of
the industry composition of demand for each
of the four major demand categories. For ex­
ample, the consumer expenditures category
includes the amounts spent on rents, automo­




biles, medical expenses, and other goods and
services purchased by consumers. For each of
the four major demand categories, a different
procedure is followed in allocating demand to
the producing industry. The industry detail to
which the categories of demand are allocated
matches the input-output classification.
Demand as used in the national income ac­
counts refers only to final demand, i.e., that of
the ultimate consumer. To place a value on the
output of an industry whose products are not
sold to ultimate consumers, but are used in­
stead by other industries in the course of their
own production, an addition set of calculations
is necessary using input-output. The inputoutput system translates final demand for a
given product into the output that is required
from all other industries to produce the ma­
terials needed to manufacture that product or
service.
The third stage is to project the input-output
coefficients. The input-output tables used as a
base in the economic growth model are pub­
lished by the Office of Business Economics, U.S.
Department of Commerce. However, these
input-output tables incorporate the technology
and product mix for a base year, and may not
reflect adequately the technology and product
mix which may prevail during the period for
which the projection is being made. To account
for this difference, two methods are used: De­
tailed analyses are made of the changes taking
place in the technology of various industries
as well as the changes expected in product mix
due to differing growth rates of product groups
within industries. For industries for which
detailed studies are not made, analyses are
developed to determine the direction and mag­
nitude of change in the use of its products by
other industries.
Next, employment estimates by industry are
developed. This is accomplished by use of a set
of industry productivity projections. The final
stage is to balance the model. Projections con­
tain many complex relationships among eco­
nomic variables that were developed through
a lengthy sequence of operations. It is neces­
sary to have a set of checks and balances to
insure that the various states of the projections
245

246

BLS H AN DBOOK OF M ETH ODS

make up an internally consistent model. The
economic growth model is designed to provide
a feed-back and balancing procedure with re­
spect to three of its elements: imports, invest­
ment, and employment. In practice, all three of
these elements must be brought into balance
simultaneously.
Uses

The projections developed in the economic
growth program serve a number of uses. The
employment projections by industry are used
in developing occupational outlook projections.
The projections developed by the Bureau form
an important part of the U.S. Government’s
report to OECD on long-term economic outlook
for the United States. In addition, other Gov­

ernment agencies use various facets of the eco­
nomic growth projection to develop projections
in their specific areas of responsibility. Outside
of government, the projections of GNP and in­
dustry growth patterns are important sources
of information for industry analysts. Their use
of these projections are primarily in diversifi­
cation studies, market analysis, and long-term
capital planning.
The structure of the economic growth model
is developed so that other analytical use is made
of the program. Specifically, the program is
used to analyze the industry by industry effect
of changing levels and patterns of various types
of demand such as defense expenditures. Simi­
lar types of analysis also is performed for ex­
ports and imports to provide information on the
employment requirements for foreign trade.

Technical References
Number

1. U.S. Department of Labor, Bureau of Labor Statistics,

P a tte r n s o f U .S . E c o n o m ic G r o w t h :

19 8 0 P r o je c tio n s o f F in a l D em a n d , I n te r -in d u s tr y R ela tion sh ip s, O u tp u t, P r o d u c tiv ity , and
E m p lo y m e n t (Bulletin 1672, 1970).
2 . _____ P r o je c tio n s 1 9 7 0 : In te r in d u str y R ela tio n sh ip s, P o te n tia l D em a n d , E m p lo y m e n t (Bul­
letin 1536,1966).
3. Oliver, Richard P., “Increase in Defense-Related Employment During Vietnam Buildup,”
M o n th ly L a b o r R e v ie w , February 1970, pp. 3-10.
4. Bowman, Charles T., “Report on Employment Related to Exports,” M o n th ly L a b o r R e v ie w ,
June 1969, pp. 16-20.




Ronald E. K utscher

Appendix A. The BLS Seasonal Factor Method
Background

An economic time series may be affected by
regular 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 re­
moving the effect of such intra-yearly repeti­
tive movements, the current evaluation of a
series may be made more perceptive.
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. Because these intra-yearly patterns are
combined with the underlying growth or de­
cline 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 judgment.
Other formal approaches were periodogram
analysis, regression analysis, and correlation
analysis. Because these methods involved a
large amount of work, relatively little applica­
tion of seasonal factor adjustment procedures
was carried out.
In the mid-1950’s, new electronic equipment
made more elaborate approaches feasible in
seasonal factor methods as well as in other
areas. Using a computer, the Bureau of the
Census developed seasonal factors based on a
ratio-to-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 Sub­
sequent improvements in methods and in com­
puter technology have led to more refined
procedures which are both faster and cheaper
than the original technique.
The Bureau of Labor Statistics began its
work in seasonal factor methods in 1959, pri­
marily 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 difference could affect any portion
of the series. It was difficult to accept a process
by which the addition of recent information
could affect significantly the seasonal factors
for periods as much as 15 years earlier, es­
pecially since this meant that factors could
never become final. The first method de­
veloped by BLS and introduced in 1960 had
two goals: first, to stabilize the seasonal fac­
tors 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
research, has made numerous changes and im­
provements in its techniques and in methods of
applying them. These changes were described 1
2
as they were introduced. The method intro­
duced in May 1966 is described in the sections
which follow.

Characteristics o f BLS Seasonal
Factor Method

The BLS method is a ratio-to-movingaverage method. It assumes that the three
component parts— trend-cycle, seasonal, and
irregular— are multiplied together (multiplica­
tive assumption) to give the original observa­
tions.3 (See illustration in charts 1 and 2.) The
BLS method differs from other similar methods
in the following respects:
1. The initial trend cycle is improved by
restoring to it any residual trend-cycle which
may have found its way into the irregular com­
ponent. This adjustment for the deficiency in
trend-cycle is developed explicitly and is avail­
1 Julius Shiskin, Electronic Computers and Business Indica­
tors. New York, National Bureau of Economic Research,
(1957). Occasional Paper No. 57.
2 See U.S. Department of Labor items in the list of Tech­
nical References at the end of this appendix.
3 A parallel method assuming an additive relationship
among the components also has been prepared for experi­
mental use.
247

248

BLS H ANDBOOK OF M ETHODS

able for review in the form of a table of trendcycle correction values.
2. The BLS method provides for changes
over time in the producing mechanism (changes
in samples, method of collection, method of
estimation) or in economic factors (introduc­
tion of guaranteed annual wage plans, etc.).4
3. The BLS method calibrates each observa­
tion in order to assign a supplementary weight
which is used in the various averaging proc­
esses. These “ credence factors” reduce the
effect of observations having large irregulari­
ties. They increase the smoothness with which
the seasonal factors change over time, and they
also keep large irregularities out of the final
trend-cycle.
4. A second way in which the BLS method
attempts to protect the final trend-cycle from
large irregular fluctuations is by using modified
original values for computing the centered
12-month moving average in the later stages of
the procedure. The credence factors are used in
obtaining the modified original values.
The BLS method is very complex and re­
quires an enormous amount of arithmetical
computations. Therefore, the method has been
adopted for use on electronic computers. At
present, two program decks are available: one
written in FORTRAN II and another in
FORTRAN IV. Special emphasis has been
placed on keeping the handling and clerical
requirements to an absolute minimum, on pro­
viding as many aids as possible for analysts or
others using the results, and on making the
application of the computer program as simple*
6
* 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 ex­
ample, indicates that in the early 1960’s the standard devia­
tion 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 increase 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.
6 Upon request, the Bureau of Labor Statistics will make its
seasonal factor computer program available in the following
form : (1) A source deck of punched cards containing the
program instructions: (2) A deck of punched cards contain­
ing a text problem which will enable the user to verify the
output: (3) Two copies of a document containing operating
and reference material.




and efficient as possible. Only the data and one
specification card are needed to produce a set
of completely labelled tables for each series.
Any number of series with varying character­
istics (length of series, starting and ending
dates, magnitude of original observations, final
table patterns desired) can be processed in a
single running without manual intervention.
The computer program for applying the BLS
procedure is available on request.5

Basic Approach

The BLS method attempts to separate an
economic 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 com­
ponents, when multiplied together, completely
and exactly exhaust the original observations
( 0) . (See charts 1 and 2.) The exact allocation
among the components is somewhat arbitrary,
because there are no simple criteria or gen­
erally accepted techniques for separating them.
The BLS 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; each successive
iteration provides an improved estimate for
each of the components of the original series.
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 irregular components. The second phase in
each iteration starts with an improved trendcycle in which the centered 12-month moving
average has been improved by recovering the
residual trend-cycle from the irregulars of the
first phase. The seasonal and irregular com­
ponents then are developed 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 effect of large

Chart 1.

United States Unemployment, Males, Age 20 and Over, 1948-70
Actual Data and Trend-Cycle

Millions

Millions

THE BLS SEASONAL FACTOR METHOD

0 .5

1948

49

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

1970

249




50

250

Chart 2.

U n ite d S tates U n e m p lo y m e n t, M ales, A g e 2 0 and O v e r, 1 9 4 8 - 7 0
Seasonal and Irregular Com ponents

Percent
140

Percent

140

£Seasonal

130

i on
lO U

120

1 120 S H | H i a 5 i 5 ! S i i 5 i S i 5 i S i S i 5

no

100

1AA
lU U

90

90

80

on
oU

70^
V

jjjw.liiijmiipww

\0 70
\

smmuumiiiimi....

130

130

jrregular
i
|

i^ U
l

lCi\J

I A

no

|

H

lyrnf^n^^

100

1 94 8 4 9



50

51

52

53

54

55

HQ

56

57

58

59

60

61

62

63

64

1
0
0
65

66

67

68

v 90
^0
6 9 1970

BLS HANDBOOK OF METHODS

110

251

THE BLS SEASONAL FACTOR METHOD

irregularities in the original observations. These
are supplementary weights in which observa­
tions with small6 irregulars are given more
weight than observations with large irregulars
in calculating the trend-cycle and the seasonal
factors.

Detailed Procedure o f 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 cor­
responding original value (0 ) to produce a
series of seasonal-irregular
(SI)
ratios
(0 / T = S I) . Treating each month separately
(i.e., all January’s, all February’s), the SI
ratios are arrayed by years and moving 7-term
weighted averages 7 (S') are secured (in per­
centage 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 seasonal-irregular ratio
(SI) is then divided by its forced seasonal
factor (S) to yield the random or irregular
ratio (SI/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 ir­
regulars, a 9-term weighted moving average 1
0
(trend-cycle correction) of the first phase ir­
regulars 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
observations) by the trend-cycle correction
(weighted moving average of first phase ir­
regulars). Using this improved trend-cycle, the
second phase repeats the computational steps
of the first phase to develop new SI ratios, new



unforced seasonals (S '), new forced seasonals
(S), and new irregulars (I). At this point, the
generated trend-cycle and seasonal components
represent the components of the original series
fairly well except for the effect of highly
0
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

Seasonal
1st year
...
2nd year ________________
3rd year __ ____________
Middle years _____ __
3rd from end ___________
2nd from end ___________
End year _______________
The underlined value
weighted average applies.

.281
.213
.160
.120

.270
.221
.179
.141
.137
___

_
_
_ _

indicates

_ _
.242 .207 _
.213 .191 .162 _
___
.185 .179 .160 .137 _
.157 .164 .157 .141 .120
.160 .179 .185 .179 .160
.162 .191 .213 .221 .213
_
_ .207 .242 .270 .281
the year to which the

8
The forcing is performed in two stages: If the unforced
seasonals do not start in January, the first 12 unforced sea­
sonals are summed and the total divided into 1200 to provide
a forcing factor. This factor is then multiplied by the un­
forced seasonals for the partial year only (through the first
December value) to provide the forced seasonals for the
incomplete at the beginning of the series. A similar pro­
cedure is followed at the end of the series if the unforced
seasonals do not end in December. 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.
0 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 completely 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 estimator of trend-cycle,
appears mainly in the irregular component. The periods
where the 12-month moving average is a poor estimator of
trend-cycle, usually have a run of consecutive irregular
ratio all on the same side of the base reference ratio of 1.000
instead of being scattered above and below this base.
1 The weight patterns used are:
0
Month

Weight pattern assigned to nine consecutive irregulars
____
____
____
.370 .341 .256 .115— .082 ____
_
1st month _
____
____
.254 .276 .254 .191 .086— .061 ____
2nd month
____
.160 .214 .231 .214 .160 .072--.051 ____
3rd month
4th month _ . .067 .150 .199 .216 .199 .150 .067—-.048
Middle
months ___ - -.050 .071 .157 .209 .226 .209 .157 .071--.050
4th month
-.048 .067 .150 .199 .216 .199 .150 .067
from end
3rd month
from end
— --.051 .072 .160 .214 .231 .214 .160
-2nd month
_ _ _ ___
__
-.061 .086 .191 .254 .276 .254
from end __
—
—
-.082 .115 .256 .341 .370
End month __
—
The underlined weight indicates the month to which the
weighted average applies.

252

BLS HANDBOOK OF METHODS

deviant original values. In the BLS 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 irregu­
lar components) and relatively more to the
neighboring values with smaller irregular com­
ponents.1
1
With the preliminary credence factors, the
SI ratios of the second phase are repartitioned
into seasonals and irregulars. The adjusted un­
forced seasonal is a 7-term moving average of
the SI 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 un­
forced 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 seasonals (S)
into the SI values.
This repartitioning of the SI ratios removes
the large irregular variation from the seasonal
component and puts it in the irregular com­
ponent instead. However, the trend-cycle is still
contaminated by deviant original values, be­
cause it was calculated before the credence
factors were developed.
The second iteration is designed to provide
an improved 12-month average for use in de­
veloping revised components. In preparation
for this, the adjusted irregulars of the first
iteration are used to develop revised (inter­
mediate) credence factors. These intermediate
credence factors are computed from the ad­
justed irregulars in exactly the same way (de­
scribed in footnote 11) as the preliminary
credence factors are obtained from the earlier
second phase irregulars.
The first use of the intermediate credence
factors is to create a modified original series
having no large irregularities. Each original
value having a credence factor less than 1.000
is replaced by a modified value in which the
irregularity has been reduced.1 The creation
1
2
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
iteration, like the first, has two phases. The
first phase begins with the centered 12-month
moving average of the modified original series
previously described. SI ratios are secured by
dividing the actual original series (not modi­
fied) by this 12-month average. The SI ratios
are arrayed by month and moving 7-term aver­
ages, using the weights of footnote 7 and the
intermediate credence factors, are taken to
yield unforced seasonals (S '). (The credence
factors prevent extreme observations from
affecting the seasonals.) The forcing process is
then applied to yield seasons (S) which aver­
age 100.0 for the calendar year. Irregulars (I)
are secured by dividing the latest SI ratios by
their corresponding forced seasonals (S).
These irregulars may include some residual
trend-cycle because of the failure of the 12month moving average to fully penetrate the
1 The credence factors are computed as follows. First, a
1
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 calibrat­
ing 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 Ir­
regular 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 devia­
tion.
A preliminary “ credence factor” is assigned to each value,
based on its standardized irregular, as follows: For a stan­
dardized irregular of 1.000 or less, the credence factor is
1.000. For a standardized irregullar of 2.800 or more, the
credence factor is 0.000. For a standardized irregular between
1.000. For a standardized irregular of 2.800 or more, the
standardized irregular. The 2.8 sigma limit makes it e x­
tremely 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 distribu­
tion lie beyond this limit. On the other hand, “bad” values
which deserve to be disregarded have a much higher prob­
ability of falling outside the limit.
1 The amount of reduction for each observation is such
2
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 trend-cycle and seasonal components de­
veloped 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 start the second
iteration: seasonal-irregulars are always calculated from the
actual original values.

THE BLS SEASONAL FACTOR METHOD

peaks and troughs of the modified original
series. A trend-cycle correction is computed by
arraying the irregulars in normal time se­
quence and taking moving 9-term averages,
using the weights of footnote 10 and the inter­
mediate credence factors. This completes the
first phase of the second iteration.
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 pro­
cedures of the first phase to develop new SI
ratios, new unforced seasonals (S') making
use of the intermediate credence factors, new
forced seasonals (S), and new irregulars (I).
These second phase irregulars are used to
calculate final credence factors in the same way
as before (see footnote 11). Then a new modi­
fied original series is secured in the same man­
ner as before (see footnote 12), using the final
credence factors. This completes the second
iteration.
Third iteration

The third iteration carries the refinement
process still further. It follows the same steps
as in the second iteration, from the centered
12-month moving average of the newly modi­
fied original series up to the development of the
irregular component near the end of the second
phase. This completes the partitioning of the
series into the final trend-cycle, seasonal, and
1 Additional iterations yield little further modification. The
3
decision to stop with 3 iterations was based on the very
small changes occurring after the third iteration, the rea­
sonable fit of the trend-cycle to the original data, and the
cost of additional processing.




253

irregular components.1 As supplementary in­
3
formation to aid 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 di­
viding the original series (0 ) by the final sea­
sonal 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 seasonally adjusted labor force will not be
quite the same as when the unemployment rate
is adjusted directly. Similarly, the sum of sea­
sonally adjusted unemployment and seasonally
adjusted employment will not quite match the
directly adjusted labor force. Separate adjust­
ment of components is usually preferable if
their seasonal patterns are different, provided
the increased measurement errors in the com­
ponents are not excessive and that the amount
of work does not proliferate unduly.
Finally, it is worth noting that the avail­
ability 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 procedure can take the place of care­
ful review and evaluation by a skilled analyst.
A subjective review of all results is strongly
recommended. The computer program for ap­
plying the BLS method facilitates such review
by providing the needed materials in a logical
and easily used format.

254

BLS HANDBOOK OF METHODS

Technical References
Number

1. Barton, H. C., Jr., “Adjustment for Seasonal Variation,” F ed er a l R e s e r v e B u lletin , June
1941, pp. 518-528.
The classic account of the FRB 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., T h e S m oo th in g o f T im e S e ries . New York, National Bureau of Eco­
nomic Research (1931). Publication of NBER 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 ea son a l A d ju s t m e n t on E le c ­
tron ic C o m p u te r s, Paris, 1961.
The report and proceedings of an international conference held in November 1960. De­
scribes experience in the United States, Canada, and several European countries. Includes
theoretical sections relating to calendar (trading day) variation and general properites of
moving averages.
4. Shiskin, Julius, E lec tro n ic C o m p u te r s and B u sin e ss In d ica to rs, New York, National Bureau
of Economic Research (1957). Occasional Paper No. 57. Also published in Jou rn a l o f B u si­
n e ss, Vol 30, October 1957, pp. 219-267.
Describes applications of the first widely used computer program for making seasonal
adjustments.
5. U.S. Department of Labor, Bureau of Labor Statistics, T h e B L S M e th o d o f D e r iv in g S e a ­
son al F a c to r s . Paper presented at the Interstate Conference on Labor Statistics, Newport,
R.I., June 16, 1960.
6. U.S. Department of Labor, T h e B L S S ea so n a l F a c to r M e t h o d ; paper presented at the 1960
Annual Meeting of the American Statistical Association, Palo Alto, Calif., Aug. 23, 1960.
7. U.S. Department of Labor, N e w S ea son a l A d ju s t m e n t F a c to r s f o r L a b o r F o r c e C o m p o n e n ts ;
Special Labor Force Report No. 8 (1960). Includes article by Morton S. Ralf and Robert L.
Stein published in the M o n th ly L a b o r R e v ie w , August 1960, pp. 822-827.
8. U.S. Department of Labor, T h e B u r ea u o f L a b o r S ta tistic s S ea son a l F a c to r M e t h o d ; Paper
presented at the Interstate Conference on Labor Statistics, Atlantic City, N.J., June 21,
1962.
9. U.S. Department of Labor, The Bureau of Labor Statistics Seasonal FactorMethod (1963
R e v i s i o n ) ; paper presented at the Interstate Conference on Labor Statistics,
San Fran­
cisco, Calif., June 27, 1963.
10. U.S. Department of Labor, T h e B L S S ea so n a l F a c to r M e th o d , I t s A p p lica tio n b y an E le c ­
tron ic C o m p u te r (1963).
11. U.S. Department of Labor, The BLS Seasonal Factor Method (1961,.). (1964).
12. U.S. Department of Labor, The BLS Seasonal Factor Method (1966). (1966).




Appendix B. Industrial Classification
Much of the usefulness of BLS 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
BLS and other Federal and State agencies fol­
low as closely as possible a single system to
define and classify industries in the U.S. eco­
nomy. The Government publishes a Standard
Industrial Classification Manual (SIC) of in­
dustries based on principles set forth by a
technical group made up of government and
industry experts.1 The Bureau of Labor Sta­
tistics took part in the development of the SIC
over a long period of years and continues to
work actively with the Office of Management
and Budget and other agencies in seeking to
improve the system.
Four basic principles were followed in de­
veloping the classification :1
2
(1) The Classification should conform to the
existing structure of American industry. (2)
The reporting units to be classified are es­
tablishments, rather than legal entities or
companies. (3) Each establishment is to be
classified according to its major activity. (4)
To be recognized as an industry, each group
of establishments must have significance from
the standpoint of the number of persons em­
ployed, volume of business, and other impor­
tant economic features, such as the number
of establishments.
As there are thousands of products and ac­
tivities, the system provides for grouping these
into categories, both narrow and broad, to en­
hance the value of industrial statistics for
users interested in different levels of detail.3
Using the SIC as a guide, the Bureau classi­
fies the reports received from each factory,
shop, or store according to major product or
activity. The SIC is used in the same way by
the agencies supplying the Bureau with its
universe lists and benchmark data. Hence, a
high degree of orderliness and consistency is
attained, which benefits not only the users of
all BLS establishment statistics, but also the
users of all Government figures.
Certain operational problems make it im­
practicable, however, to secure complete uni­
formity 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 ac­
tivity” is determined; or changes in the major
activity of individual establishments which
occur over time may be handled in statistical
time series in different ways. Consequently, the
use of the same manual and following a com­
mon set of principles of application do not
always result in identical industry classifica­
tions of a given establishment by all agencies,
or even by all programs within BLS. There­
fore, any major deviations from the normal
method of handling industrial classification
will be described in the chapters on BLS estab­
lishment statistics, such as those on employ­
ment, work injuries, and the like.
The standard definition of establishment is
stated as follows:
An ‘establishment’ is an economic unit which pro­
duces 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 predominantly one, type of economic ac­
tivity for which an industry code is applicable.
Where a single physical location encompasses two or
more distinct and separate economic activities for which
1 Office of Management and Budget, Standard Industrial
Classification Manual, 1967.
2 Ibid. pp. IX, X .
3 The SIC provides for different levels of aggregation. The
broadest level divides the economy into 10 Divisions: A.
Agriculture, forestry, and fisheries; B. Mining; C. Contract
construction; D. Manufacturing; E. Transportation, com­
munication, electric, gas, and sanitary services: F. Wholesale
and retail trade; G. Finance, insurance, and real estate; H.
Services; I. Government; and J. Nonclassifiable establish­
ments. At the 2-digit level all products and services are
combined into 99 “major groups.” Thus, in the Manufacturing
Division, establishments engaged in manufacturing machinery,
apparatus, and supplies for the generation, storage, trans­
mission, transformation, 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 major group the SIC provides 8
groups of industries: 361. Electric transmission and distribu­
tion equipment; 362. Electrical industrial apparatus; 363.
Household appliances; 364. Electric lighting and wiring equip­
ment; 365. Radio and television receiving sets, except
communication types; 366. Communication equipment; 367.
Electronic components and accessories; 369. Miscellaneous
electrical machinery, equipment, and supplies.
Thousands of products and activities are distinguished at
the 4-digit level. For example, in Group 367, five industries
are defined: 3671. Radio and television receiving type electron
tubes, except cathode ray; 3672. Cathode ray picture tubes;
3673. Transmitting, industrial, and special purpose electron
tubes; 3674. Semiconductors and related devices; 3679. Elec­
tronic components and accessories, not elsewhere classified.
255

256

BLS HANDBOOK OF METHODS

different industrial classification codes seem applicable,
such activities should be treated as separate establish­
ments and classified in separate industries, provided it
is determined that: (1) such activities are not ordi­
narily associated with one another at common physical
locations; (2) no one industry description in the Stan­
dard Industrial Classification includes such combined
activities; (3) the employment in each such econom
ic
activity is significant; and (4) reports can be prepared
on the number of employees, their wages and salaries,




and other establishment 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 distinguished from organizational sub­
units, departments, or divisions within an establish­
ment. Supplemental interpretations of the definition of
an establishment are included in the industry descrip­
tions of the Standard Industrial Classification.4
4 Ibid., p. X .

Appendix C. Geographic Classification
United States and States

All statistical series of the Bureau were ex­
panded in 1959 to include Alaska and Hawaii.
Due to the relatively 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, BLS assigns an estab­
lishment to the State in which its employees
are reported for payroll tax purposes, since
these sources are used both for deriving bench­
mark 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 statistics among Federal agencies is main­
tained by means of a set of published standard
definitions.1 Defining metropolitan areas for
statistical purposes is done on the basis of in­
formation about population and about economic
and social ties among cities. To qualify as a
Standard Metropolitan 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 cri­
teria.1 Which counties are to be included in the
2
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 repre­
sented 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 Office of Management
and Budget.
State Economic Areas

State economic areas are relatively homo­
geneous subdivisions of States developed by the
U.S. Department of Commerce’s Bureau of the
Census, in cooperation with the X / . Depart­
J&
ment of Agriculture’s Bureau of Agricultural
Economics, and several State and private agen­
cies. 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 char­
acteristics which distinguish it from adjoining
areas.4

— Robert B. Steffes
1 U.S. Office of Management and Budget, Standard M etro­
politan Statistical Areas, 1967.
2 The population criteria are:
1. Each standard metropolitan statistical area must include
at least:
(a) One city with 50,000 or more inhabitants, or
(b) Two cities having contiguous boundaries and consti­
tuting, 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 limits), they will be included in the same area unless
there is definite evidence that the two cities are not eco­
nomically and socially integrated.
3 Ibid., pp. 1-2.
* U.S. Bureau of the Census, Census of Population 1960. Re­
port PC (1) 1A. Number of Inhabitants. U.S. Summary, p.
X X V II.

☆

U.S . GO VER NM EN T PRINTING OFFICE: 1972 0— 430—780

257

BUREAU OF LA B O R STA TISTIC S
RE G IO N A L OFFICES

Region I
1603-JFK Federal Building
Government Center
Boston, Mass. 02203
Phone: 223-6762 (Area Code 617

Region V
8th Floor, 300 South Wacker Drive
Chicago, III, 60606
Phone: 353-1 880 (Area Code 31 2)

Region 1
1
341 Ninth Ave., Rm. 1025
New York, N.Y. 10001
Phone: 971-5405 (Area Code 212)

Region VI
1100 Commerce St., Rm. 6B7
Dallas, Tex. 75202
Phone: 749-3516 (Area Code 214)

Region III
406 Penn Square Building
1317 Filbert St.
Philadelphia, Pa. 19107
Phone: 597-7796 (Area Code 215)

Regions V II and V III
Federal Office Building
911 Walnut St., 10th Floor
Kansas City, Mo. 64106
Phone: 374-2481 (Area Code 816)

Region IV
Suite 540
1371 Peachtree St. NE.
Atlanta, Ga. 30309
Phone: 526-5418 (Area Code 404)

Regions IX and X
450 Golden Gate Ave.
Box 36017
San Francisco, Calif. 94102
Phone: 556-4678 (Area Code 415)




Regions V II and V III will be serviced by Kansas City.
Regions IX and X will be serviced by San Francisco.


Federal Reserve Bank of St. Louis, One Federal Reserve Bank Plaza, St. Louis, MO 63102