A century of wage statistics: the BLS contribution Essay

Wage and salary rates of pay remain at the heart of the labor
bargain, although a new dimension has been created in recent decades by
the rise of various forms of supplements to employee compensation.
Information on the general movement of wage rates, and on the structure
of rates by such characteristics as occupation, industry, region, union
status, and sex, provides crucial insight on the status and well-being
of the working population. In a complex industrial society, the
development with limited resources of useful statistics in these areas,
and more recently in the area of supplementary compensation, has been a
formidable undertaking.



This article traces the work of the Bureau of Labor Statistics over
the past century in the field of wage statistics, including the
attention that has been given since World War II to the growth of wage
supplements. An effort has been made to place this work in broad
historical perspective. This account does not cover related Bureau
programs, including the extensive work on consumer prices and the
important series of average hourly and weekly earnings by industry
developed from employment statistics. 19th century beginnings



The years 1875 to 1900 were in many ways a period of extraordinary
economic growth and change in the United States. It was marked by the
closing of the geographic frontier. An impressive expansion of the
transportation network facilitated the settlement of the West, and
contributed to a large increase in farm output. At the same time,
manufacturing expanded at a rapid pace, accompanied in many industries
by larger scale operations, consolidations of firms, and the growth of
monopoly practices.


The last quarter of the century also saw substantial changes in the
size and industrial composition of the labor force. Between 1880 and
1900, the number of “gainful workers” increased by almost 12
million. In 1880, the gainfully employed work force was about equally
divided between agricultural and nonagricultural employments, but by
1900 the agricultural share, while still rising in absolute terms, had
declined to approximately 37 percent of the total. The proportion of
employment in manufacturing, transportation, and construction had
increased significantly over the same period.



There changes were accompanied by economic fluctuations of
considerable magnitude, including an unusually sever depression
beginning in 1893. As a result, there arose new currents of thought
with respect to wage determination, trade unionism, and the role of
government in relation to labor, and various comprehensive movements for
social reform emerged. In particular, the impulse toward collective
action to defend or improve labor standards began to acquire momentum
among the growing wage-earning population. Of major although
short-lived significance was the meteoric rise of the Knights of Labor,
the membership of which reached about 700,000 in 1886 but declined
precipitously thereafter. Of much greater long-run importance was the
formation of the American Federation of Labor (AFL) in 1881 as a
permanent trade union center. The Federation survived the long
depression of the 1890’s, and union membership began to climb
sharply toward the end of the decade–rising from 447,000 in 1897 to
868,500 in 1900, mainly in unions affiliated with the AFL. (These
figures include Canadian members of labor unions with headquarters in
the United States.)



It was during this period that the Bureau of Labor Statistics was
created, with a broad mandate to “. . . collect information upon
the subject of labor, its relation to capital, the hours of labor, and
the earnings of laboring men and women. . . .” The creation of the
new agency in 1884 reflected a growing demand for information on labor
conditions to provide a basis for improved labor standards. With respect
to earnings or wages, there were few guides for the work of the new
Bureau. Some experience with wage surveys had been accumulated by a few
State agencies, notably in Massachusetts. At the Federal level, the
only important previous effort to develop statistics of wages by
occupation was a special study conducted for the decennial census of
1880. This report, which was not published until 1886, gave annual
average wage rates by occupation for “typical” establishments
in 53 industries back, where possible, to 1860 or even earlier. The
data, which were collected by mailed questionnaires, were published in
great detail.


The new Federal Bureau was extraordinarily fortunate in its first
commissioner, Carroll D. Wright, who had headed the Massachusetts Bureau
of Statistics of Labor since 1873. Wright would prove an authentic
pioneer in the development of labor statistics both here and abroad.
With respect to the collection and presentation of wage survey data, a
number of general principles reflected his experience in Massachusetts
and at the Federal Bureau. These related basically to the compilation
and presentation of data designed to throw light on the structure of
wages, although, as we shall see, the new Bureau’s studies also
would provide the basis for most of our knowledge of the trend of wages
well into the present century.



With regard to the collection of wage data by occupation, Wright
felt strongly that the use of trained field agents rather than mailed
questionnaires was required to ensure the adequacy and quality of
response. In the absence of modern sampling procedures and
establishment universe information, he favored the coverage of
“typical” of “representative” establishments, which
in practice meant those that had been in business for some time.
Concerning the presentation of wage statistics, Wright was insistent upon the use of well-defined occupational classifications. In view of
the observed dispersion of wage rates within occupations, he tended to
favor presenting survey results in the form of wage distributions, where
feasible, rather than by occupational averages alone. “Of late
years,” he wrote in 1892, “the demand has been that employees
should be classified not only minutely as to occupations, but as to
rates of pay as well.



Appropriately, in view of the times, the first annual report of the
Commissioner dealt with industrial depressions. It included the results
of the first occupational wage survey conducted bys the Bureau. The
data related to 1885 and were taken directly by Bureau agents from the
payroll records of 582 establishments in about 40 industries
overwhelmingly in the manufacturing sector. The results were published
in the form of daily average wage rates by occupation, industry, and
State; estimates were presented separately for men, women, and children
and youths.



The Commissioner’s fifth annual report (1889) on railroad labor developed occupational wage statistics in great detail for 60
carriers. The Bureau’s field agents found more than a thousand job
titles in the payroll records of these railroads. Many of these
involved similar duties. The most detailed wage statistics were shown
in table 11 of the report, where distributions of daily rates or
earnings by occupation were presented, together with distributions of
annual earnings. The claim was made that “. . . the chief value of
this report, so far as time and wages are concerned, is to be found in a
thoroughly scientify classification, not only of the time employed of
each individual employee of the roads considered, but of rates by day
and by year,” as shown by the payrolls.



These studies, together with two other extensive wage surveys
conducted during the early 1890’s, provided the Bureau with
invaluable experience in occupational classification, data collection,
and the presentation of survey results. They prepared the way for a
highly fruitful survey stemming from a Senate resolution of March 3,
1981, which instructed the Committee on Finance to study the effects of
tariff legislation on wages and prices. In accordance with the
resolution, Senator, Nelson W. Aldrich asked Wright to undertake the
task of developing wage and price statistics for the years 1840-91.



The problem of locating establishments with payroll records for all
or most of this long period obviously was difficult. The study
ultimately provided data for more than 500 occupational series in 22
industries, mainly in manufacturing. Comparatively few of the
establishments and usable payroll records prior to 1860. The
publication of the data for two payroll periods in each year by
industry, establishment, and occupation, took up almost 1,300 pages of
the Senate Committee’s report. The Committee observed that “.
. . no other investigation has been made with so with a scope, such a
variety of detail, and covering so extensive a period.”



The significance of this study for the Bureau’s work in the
field of wage statistics was twofold. It gave experience in the
collection of data over an extended period in the past, and it provided
the basis for measurement of changes in the level of wages through the
construction of wage indexes. The latter represented an innovation of
signal importance for which credit must go to Professor Roland P.
Falkner of the University of Pennsylvania, who was employed by the
Senate Committee to analyze the survey data. Using 1860 as the base
year, Falkner prepared indexes for each occupation by establishment, for
each of the 22 industries included in the survey, and for the 22
industries combined.



Despite its many limitations, the Bureau’s work for the
Aldrich Committee is the major source of information on the structure
and course of wages in this country from 1860 to 1890, and yields some
insight for the years back to 1840. The study’s wage trend
estimates have been analyzed and reworked a number of times, most
recently by E. H. Phelps Brown and Sheila V. Hopkins and by Clarence D.
Long.



The next major wage study undertaken by the Bureau undoubtedly was
influenced by experience gained in surveys for the Aldrich Report, and
foreshadowed the nature of work to be done into the 20th century. This
study provided average daily wage rates by year for each of 25 selected
occupations in 12 major cities for the period 1870-98. The occupations
selected were those “susceptible of accurate definition,” and
the data were taken directly from the payroll records of at least two
establishments in each city for the occupations covered. The
report’s text tables showed average daily wages by year for all
occupations combined, and the percentage change since 1870 for each year
from 1871 to 1898. The latter estimates were deemed to be “quite
indicative” of the movement of wages generally. Similar data were
shown for three cities in Great Britain and for one city each in France
and Belgium. The foreign data were compiled by the authorities in the
countries concerned at the request of the Bureau.



In short, during the last 15 years of the 19th century, the Bureau
accumulated considerable experience in the planning and conduct of
surveys of wages and standard hours of work. The merits of data
collection from payroll records by personal visit were established. Much
insight was gained into the difficult problem of job classification at a
time when formal job descriptions were uncommon and titles for the same
job could vary widely among establishments. Problems in the presentation
of occupational wage data related to the observed dispersion of rates of
pay were recognized. Finally, the introduction of index numbers in the
Aldrich Report provided a convenient means for measuring the movement of
wages over time. The development of sophisticated survey sampling
techniques was, of course, far in the future. 1900 to the Great
Depression



The Bureau’s work in wage statistics during the first three
decades of the 20th century achieved a coherence that distinguishes it
from its 19th century origins. This cohesion was attained despite the
great economic and social changes that occurred during the period,
including U.S. participation in World War 1. Between 1900 and 1930,
gross national product in constant dollars rose at an average annual
rate of 3.1 percent. The sharpest downturn in economic activity was the
comparatively short postwar recession beginning toward the end of 1920.
The civilian labor force grew by about 20 million, with nonfarm workers
accounting for approximately 77 percent of total employment by 1930.
The automobile began to provide greater mobility for both people and
industry. The growth in trade union membership that had begun in 1897
continued with only minor pauses to 1920, when membership reached 5
million. A sharp decline occurred during the postwar recession, but
membership stabilized at about 3.6 million by 1924. Finally, a
substantial body of protective social legislation was enacted at the
State level during this 30-year period, largely with reference to work
performed by women and children.



During the winter of 1900-01, the Bureau began a major study of
occupational wages by industry, with the data carried back to 1890. The
study was undertaken with the view that “. . . the constant demand
for current data could be met only by a very painstaking and complete
investigation which would result in thoroughly representative figures
for a period of years and which would serve as the basis for the regular
annual collection and presentation of data concerning wages. . . .”
Due to staff limitations, the study required several years for
completion. Data for 1902 and 1903 were included in the final results
of the investigation, which appeared in 1905 as the Commissioner’s
Nineteenth Annual Report, a volume of almost a thousand pages.



The study was confined to “the leading manufacturing and
mechanical industries” and to the “distinctive occupations
which are considered representative of each industry,” and covered
payroll periods most nearly representing “normal conditions”
of operations for each establishment during each year. In all, 67
industries, 519 occupations, and 3,475 establishments were surveyed.
Industry coverage was largely confined to manufacturing and the building
trades.



Each year from 1890 to 1903, the Commissioner’s Annual Report
presented average rates of pay by occupation, industry, and region, and
for selected occupations, by city and State. Complete wage
distributions also were given for the occupations included in the city
and State tabulations. An outstanding achievement was the presentation
in a text table of an overall index of hourly rates, computed from
averages of industry relatives weighted by aggregate wages paid by each
industry as shown by the census of 1900. As Paul H. Douglas pointed
out in his great study of real wages in the United States from 1890 to
1926, the Bureau’s index computations were for some years the
preferred statistical basis for generalizing on the trend of wages.
(Actually, the correspondence between the Bureau’s wage-weighted
index and Douglas’ own wage rate index for manufacturing, which was
prepared from the same data but differently constructed, is strikingly
close. For example, both indexes show wages in 1903 to be 16 percent
higher than the average for 1890-99.)



The general format of the 1890-1903 study was followed in annual
surveys during the next 4 years. The results of these studies were
reported in the Bureau’s bimonthly bulletin. The surveys again
were confined largely to manufacturing, but their scope was limited to
industries in which wages paid amounted to $10 million or more as shown
by the 1900 census. The 1904 survey, for example, covered 350 selected
occupations, 3,732 establishments, and 42 industries. The 1904 wage
index for all industries was linked to the general index for 1890-1903
on the basis of changes in those establishment studied in both 1903 and
1904, and this chaining procedure was followed for subsequent years to
1907.



After 1907, there was a 4-year interruption in the Bureau’s
wage survey program. This was due primarily to the pressure of other
work, notably a large investigation into the conditions of women and
child wage earners and a study of wages, hours, and working conditions
in the iron and steel industry, both of which were published as Senate
documents. The program was resumed in 1912 with two series of surveys.
The first consisted of studies based on payroll records of rates of pay
(or of earned rates for incentive workers) in selected occupations in 12
industries: cotton, wool, and silk textile; lumber, millwork, and
furniture; boots and shoes; hosiery and knit goods; cigars; clothing;
iron and steel; and building and repairing of steam railroad cars.
Except in the case of cigars and clothing, data were carried back to
1907. With the next few years, five of these industries (silke,
millwork, furniture, cigars, and car building) were dropped while
slaughtering and meatpacking was added. The industries that remained in
the program were surveyed approximately every 2 years until 1933.



The 1912 survey of cotton-goods manufacturing and finishing
illustrates the essential nature of this group of studies. Bureau
agents obtained data for the “principal occupations” directly
from mill payroll records. An innovation was the publication of job
descriptions in the survey report. The coverage of establishments was
considerably broader than in earlier studies of cotton-goods
manufacturing, and the data, as previously noted, were carried back to
1907. Distributions of workers by hourly rates of pay at 1-cent
intervals were shown by occupation for the industry as a whole and by
State. The Bureau’s continued interest in the trend of wages was
manifest by the linkage of annual wage changes during 1907-12 to the
existing 1890-1907 index for cotton textiles.



The second group of studies begun in 1912 were concerned with union
wage scales and standard hours of work in the building industry,
newspaper printing, book and job printing, marble and stone work, metal
trades, baking, and millwork. In all, 49 crafts were surveyed in 39
cities. The initial report states that the data were “. . . in
every case furnished by officials of the local unions to special agents
of the Bureau of Labor Statistics, and wage scales, written agreements,
and trade union records were used wherever possible.”



As with the payroll studies, the union wage data were extended back
to 1907. Wage scales were shown by trade and city for each year from
1907 to 1912, and indexes of the movement of scales back to 1890 were
computed for many of the crafts in the industries covered. These
studies, with some changes in industry coverage, were to continue on an
annual basis for almost 80 years.



These payroll and union wage studies had the great virtue of
providing a large measure of consistency in the Bureau’s
occupational wage survey program over a period of roughly two decades.
They also provided a measure of continuity with the major survey that
had produced the 1890-1903 report on occupational wages and with the
annual surveys of 1904-07. They developed data on the structure of
wages by skill and sex for manual jobs in a variety of relatively
low-wage and high-wage industries. Data provided by locality and State
yielded a measure of insight into interarea wage differences.



As previously mentioned, these surveys were the source for the
manufacturing and building components of the seminal Douglas study of
money and real wages from 1890 to 1926. Shortly after World War I, the
Bureau itself ventured to put together an annual general index of wages
in response to inquiries that “. . . have generally related to
recent years but . . . frequently ask for an index that shall compare
Civil War Changes with those during and following the late World
War.” The index was prepared (with some hesitation) from “all
sources available,” and was published initially for the period from
1840 to 1920. It was later extended in several stages to 1934.



Two additional observations should be made concerning the
Bureau’s work in wage statistics during this period. The first is
that World War I had minimal impact on the generation of wage data.
Although several Government agencies were established to deal with
wartime labor problems, there was not effort, as there would be during
World War II, to impose comprehensive wage controls. Wage adjustment
efforts for particular industries (such as shipbuilding) tended to focus
on those made necessary by rising living costs, which led to the
establishment of the Bureau’s Consumer Price Index. Toward the end
of the war, the War Industries Board asked the Bureau to undertake wage
surveys in a number of industries for use in the solution of labor
problems and to provide a record of industrial conditions at the height
of the war effort. These surveys, with occupational wage data for 28
industries, were not completed until after the war, and the results were
published in 1920.



The second observation is that, while generally continuing the
group of payroll studies begun in 1912, the Bureau extended its wage
survey activity to additional industries during the 1920’s. A
number of these were manufacturing industries essentially new to the
20th century, including motor vehicles, rubber tires, synthetic
textiles, and airplane and aircraft engines. Several nonmanufacturing
industries also were added to the program, most notably bituminous coal mining and air transport.



This major phase of the Bureau’s work in wage statistics came
to an end about 1932. Although the surveys of the period were confined
to manual jobs and largely to selected industries in the manufacturing
sector, they provided a reasonably consistent body of data on both the
structure and trend of wages for industrial workers. They undoubtedly
also played a role in private wage determination through collective
bargaining and employer personnel administration. In their absence, we
should know much less than we do about economic conditions during the
first three decades of this century. Depression and war, 1930-45



The great Depression began toward the end of 1929. Unemployment,
which was estimated at 3.3 percent of the civilian labor force during
the years 1923-29, rose to an estimated 25 percent in 1933. Recovery was
only partial during the remainder of the decade; even in 1940, the
unemployment rate was estimated at 14.6 percent. However, the steady
expansion of war production and of the Armed Forces after mid-1940
brought the rate to 1.2 percent by 1944.



The singular decade of the 1930’s witnessed the beginning of
an unprecedented and continuing involvement of the Federal Government
with the economy. Aside from social insurance programs, this
development expressed itself with respect directly to labor in two major
forms. The first was machinery through the Davis–Bacon Act of 1931 and
the Walsh–Healey Act of 1936 for the establishment of wage standards
for workers employed by contractors or subcontractors on public
construction or in the provision of manterials and supplies to the
Federal government, and, by the Fair Labor Standards Act of 1938, for
minimum wages for most workers engaged in or producing goods for
interstate commerce. The second, embodied in the National Labor
Relations Act, Protected the right of workers to join unions and imposed
upon employers the duty to bargain collectively over wages and other
terms of employment. Partially as a result of this latter act, U.S.
union membership, which had declined to 2.7 million by 1933, reached
14.3 million by 1945, and collective bargaining was extended to many
strategic sectors of the economy. During the same year that the act was
passed, however, a deep split occurred in the trade union movement, a
breach that was not to be closed for 20 years.



By the time the United States became involved in World War II as a
combatant on December 7, 1941, the economic impact of the conflict had
already begun to be felt and measures to deal with its consequences had
emerged. One group of measures designed to contain the inflationary consequences of resource diversion to war production was comprehensive
control of changes in wage rates and prices. This had a decisive effect
on the Bureau’s wartime wage statistics programs, and postwar
consequences as well.



The early 1930’s had seen the end of the relatively small but
systematic program of payroll-based industry wage surveys that had been
conducted during the previous two decades. The annual survey of
“common labor entrance rates” was discontinued in the early
1940’s in large part because of the increasing ambiguity of the
concept of “unskilled” or “common” labor. Dropped
also, in 1934, was the occasional publication of the broad annual wage
index that had first been published in 1920, having been pieced together
from such data as were available. Its discontinuance appears to have
been related to the institution in 1932 of the Bureau’s average
employment and payroll reporting system.



Between 1932 and the beginning of the defense buildup in 1940, the
Bureau’s wage survey activity was largely, although not entirely,
geared to the informational needs of the new Federal agencies concerned
with labor standards. Thus, a number of studies were undertaken for use
in the administration of the short-lived National Industrial Recovery
Act of 1933 which provided by industry for “codes of fair
competition” containing minimum wage and maximum hour provisions.
At the direct order of the President, an especially noteworthy study was
made in the cotton textile industry covering pay periods in 1933 and
1934, following a general strike in that industry in 1934. Several
surveys were undertaken in cooperation with the Works Progress
Administration. Work was done also in connection with prevailing
minimum wage determinations under the Walsh-Healey (Public Contracts)
Act of 1936, which covered work performed by Federal government
contractors.



The pace of survey activity accelerated after the passage of the
Fair Labor Standards Act (1938), which initially provided for Federal
minimum wage determination (above a statutory level) on an industry
basis. During 1938 and 1939, about 45 industry wage surveys were
conducted for use in minimum wage proceedings. Most of these studies
developed data on the distribution of workers by pay rates or
straight-time hourly earnings, without occupational detail. They
typically related to relatively low-wage consumer-goods industries. In
a few cases, they provided the basis for appraisal of the wage and
employment effects of minimum wage orders before these effects were
masked by the upsurge of economic activity associated with the war.



Concern with minimum wage determinations faded as the defense
program got underway in mid-1940. With expansion of the defense effort,
unemployment declined, shortages of skilled workers began to appear, and
the incidence of strikes rose sharply during the first half of 1941. In
March of that year, a tripartite National Defense Mediation Board was
appointed to assist in the settlement of labor disputes; this agency was
superseded in January 1942 by the National War Labor Board, which was
given the additional function of stabilizing rates of pay as part of a
comprehensive economic stabilization effort.



As a consequence of these developments, the Bureau’s wage
survey activity shifted initially from consumer-goods industries to
heavy industries essential to war production. Occupational wage studies
were undertaken in such industries as shipbuilding, aircraft, rubber,
nonelectrical machinery, and the mining, smelting, and refining of
nonferrous metals. Beginning in 1941, special studies were also
requested for use in the settlement of specific labor disputes in
industries vital to the war effort.



With the beginning of comprehensive wage stabilization under the
War Labor Board, the governmental need for wage statistics increased
manyfold. This led to a substantial expansion of Bureau staff and the
establishment of regional offices to parallel those set up by the Board.
For the most part, two types of wage information were provided for Board
use. The first involved data on wage rates or straight-time earnings by
occupation, industry, and labor market for Board decision in thousands
of claims for increases on interplant wage inequity grounds, together
with some studies in connection with specific labor disputes. The
second was a general wage rate index to measure the effectiveness of the
wage stabilization program.



The surveys for use in Board decisions in inequity cases involved
special procedures to expedite data collection and presentation. These
included the development of patterns of “key” jobs by
industry, the preparation of uniform job descriptions, and the
standardized presentation of survey results. Althogether, job patterns,
each typically including from 10 to 20 occupations, were prepared for
more than 120 manufacturing and nonmanufacturing industries. It was
reported in 1945 that, under this crash program, data on pay rates in
key occupations had been collected from more than 100,000
establishments, and that some 8,000 reports on an industry-locality
basis had been transmitted to the Board. This was an extraordinary
achievement over a period of little more than 2 years, and represented a
vital contribution to the wartime wage stabilization program.



The second Bureau contribution to the wage stabilization effort was
the construction of an occupationally based index of urban wage rates.
Becuase the War Labor Board sought to stabilize basic rates of pay
rather than earnings, such an index was needed to provide some measure
of its effectiveness. A properly constructed index can eliminate many
of the factors that affect earnings rather than rates, including premium
pay for overtime and late-shift work, interindustry employment shifts,
and changes in occupational skill ratios.



The urban wage index was based on data for key occupations in about
6,000 establishments, and covered selected manufacturing and
nonmanufacturing industries in some 69 labor markets, It was issued
semiannually beginning in April 1943, but was estimated back to Januar
1941 for the manufacturing sector. Discontinued largely for budgetary
reasons in 1947, this index would have been a valuable tool for the
analysis of postwar wage movements. Its final publication gave
estimates of wage-rate change for manufacturing between January 1941 and
September 1947, and for selected nonmanufacturing industries between
April 1943 and April 1947. Postwar program adjustment, 1945-47



For more than 3 years, almost all of the Bureau’s expanded
resources for wage survey purposes had been devoted to data needs for
wage stabilization and labor dispute settlement. By the beginning of
1945, the pressure of these needs had eased substantially. The war was
coming to an end, and problems of industry reconversion to peacetime
production were being discussed. It was now necessary to devise an
interim program to meet anticipated requirements for wage statistics
during the immediate postwar period.



The most significant Bureau decision with regard to the wage
program was to conduct a large number of nationwide occupational wage
survey on an industry basis, with regional and locality breakdowns
whenever possible. Between 1945 and mid-1947, such studies were made in
no fewer than 70 manufacturing and 11 nonmanufacturing industries. Each
presented data on rates of pay for selected occupations and rate
distributions for all plant workers on a national and regional basis,
and virtually all contained occupational data for specific localities.
Information typically was shown on occupational earnings by size of
establishment, size of community, method of wage payment, and
unionization, and for such plant practices as scheduled hours of work,
shift differentials, paid vacations and sick leave, and insurance and
pension plans. Most of the studies also contained data on salaries for
a few office occupations.



This large group of nationwide studies were issued in a special
series of wage structure reports in mimeographed form. Unfortunately,
few were published in the Bureau’s series of numbered bulletins,
but many summary articles based on the surveys appeared in the Monthly
Labor Review.



During this period, the Bureau continued its annual surveys of
union wage scales in building, printing, and several other industries.
Of great importance in subsequent survey planning were studies of
clerical salaries on a cross-industry basis in a few local labor
markets. A variety of special studies relating to such issues as
guaranteed wage plans in the United States, the economic status of
registered nurses, and the experience of workers in various industrial
reconversion situations also were conducted.



Particular notice should be given to the rise of a board analytical
capability within the Bureau during this period. This, no doubt, was a
function of staff size stimulated by the exceptional volume of
statistical information generated during 1945-47. Several notable
studies were produced, including an analysis of changes in wage
differentials by skill for manual workers between 1907 and 1947 in
manufacturing and the building trades. This pioneer study pointed up
striking changes that had occurred in supply-demand conditions in the
labor market over the four decades. A companion study, confined to
manufacturing, dealt with the course of regional wage differentials
during the same period. Other studies relating to the immediate postwar
period dealt with such topics as shift differentials in manufacturing,
vacation practices in major industry groups, and the prevalence of
insurance and pension plans. Reshaping wage programs, 1947-70



In mid-1947, the Bureau experienced a sharp reduction in its
budget–in effect a postwar readjustment–although its resources
remained substantially above the prewar level. With respect to wage
statistics, this budget reduction, together with other factors, led to a
major reexamination of programs.



A basic question was whether continued concentration on nationwide
industry surveys represented the best use of the available resources.
Experience during the war had indicated the great importance of data at
the level of the local labor market, where most wage decision occur. It
was clear, moreover, that the white-collar segment of the labor force
had shown impressive growth and would continue to expand. In pay
determination for most types of white-collar workers, the importance of
the local labor market was, if anything, greater than for manual
workers. Most white-collar jobs are found in a great variety of
industries, suggesting the need for some type of cross-industry survey
for this type of employment.



Other influences also were at work. The rapid spread of trade
unionism in the 1930’s and during the war, and the
“rounds” of wage increases during the years immediately
following the war, produced great interest in the dynamics of wage
adjustment. Another influence was the spread of supplementary benefits,
public and private, representing income to workers and cost to
employers. Aside from the existence of a broad demand for data for use
in private wage decisions and for public policy purposes, great academic
research interest also had developed.



A general factor that contributed importantly to program
development during this period was the emergence during the late
1940’s of efficient modes of probability sampling in government
statistical surveys. With relation to the Bureau’s studies of
occupational wages, pathbreaking work was done in adapting probability
sampling to universes of establishments. The use of probability
sampling, rather than purposive sampling, permitted a more efficient use
of survey resources, the more reliable calculation of population values,
and the estimation of sampling errors.



The essential problem in 1947 was to devise, with limited
resources, a program that would throw as much light as possible on the
structure of occupational wages and salaries, on supplementary forms of
compensation, and on the dynamics of wage development. Considerable
thought and experimentation suggested that two occupational wage survey
systems were needed, together with several other types of recurring reports. The principal ingredients of the revised program are
summarized below.



Cross-industry labor market surveys. In 1948, pilot studies on the
collection of salary data for office clerical occupations on a
cross-industry basis were undertaken in a few large labor markets. These
were successful technically, and the response from users of wage data in
the markets covered was encouraging. In 1949, experimental survey were
made in a number of smaller cities. The occupational coverage was
broadened in these studies to include selected skilled and relatively
unskilled manual jobs that were not unique to particular industries.
The skilled manual occupations were drawn primarity from those involved
in plant maintenance, and the unskilled from material movement,
warehousing, and custodial work. Our of this experimental effort grew
the concept of the community wage survey.



As it happened, this exploratory work was coming to a close at the
outbreak of the Korean War in June 1950, which called forth another wage
stabilization effort. The National Wage Stabilization Board, formed in
September 1950 to administer wage controls, concluded that community
wage studies were well suited for use in its work. It provided
resources for the conduct of such studies in a large number of labor
markets, with occupational coverage extended to jobs peculiar to major
industries in each area surveyed. The results of these and other Bureau
wage studies and reports were extensively used in Board decision, and
provided the basis for a series of analyses of interarea differences in
the level of wages, occupational differentials, fringe Benefits, extent
of unionization, and the formalization of wage structures.



Wage Stabilization Board budgetary support for the labor market
survey program ended after the Korean emergency. The Bureau’s own
resources could provide for only about 20 community surveys annualy,
with coverage limited to cross-industry occupations. If was decided as
a matter of policy to conduct these surveys in the same metropolitan
areas each year. This decision was based largely on the fact that the
use of survey data in wage determinations by employers and unions
depends on its currency. Extensive support developed for the surveys in
the communities in which they were made. Numerous requests for
community surveys in other areas had to be firmly rejected.



In light of the widespread interest in the cross-industry type of
survey, a rational basis was sought for a more elaborate program. The
result was a proposal for annual surveys in a sample of metropolitan
areas (approximately 80 of the 183 then existing), selected to represent
all such areas. The sample was designed to include all metropolitan
areas with employment of 250,000 or more. This would permit estimates of
the level and distribution of wages for a significant group of
white-collar and manual jobs for all metropolitan areas–in effect, for
the urban economy. It would provide a basis for national estimates,
separately for office and plant workers, of scheduled hours of work,
holiday and vacation provisions, the incidence of private insurance and
pension plans, and collective bargaining coverage. Wage comparisons
among areas and broad regions also could be made.



The budgetary requirements for this program were met toward the end
of the 1950’s when an urgent Federal need developed for national
data on white-collar salaries in private industry to implement a
comparability pay policy for the 1.7 million Federal white-collar and
postal employees. An interagency technical committee concluded that the
annual 80-area survey design was an appropriate survey vehicle, with
additional data collected from a subsample of establishments for
selected professional, managerial, and technical jobs. The Congress
approved funds for this program, and for expansion of other aspects of
wage survey activity, for fiscal 1960 (then beginning July 1959). When
Congress passed the Federal Salary Reform Act of 1962, this
cross-industry survey had been tested in 2 years of operation.



With some changes in the design and size of the sample, refinements
in occupational definitions, the inclusion of a number of additional
occupations, and modifications of data collection procedures, this
survey system has continued on an annual basis for more than two
decades.



Industry wage surveys. Occupational wage studies on an industry
basis were not abandoned with the development of the community wage
survey program. Such studies remained highly important for insight into
the structure of wages and benefits for nonsupervisory workers in
establishment groupings differentiated by product, technology, labor
force composition, extent of unionization, and other factors. Partly for
economy in the use of resources, there was a shift in emphasis during
the late 1940’s from the industrywide surveys of the 1945–47
period to surveys in major areas of industry concentration. The intent
behind this industry-locality program was the annual study of wages and
related benefits in some 25 manufacturing and nonmanufacturing
industries, together with less frequent industrywide studies in a few
industries where wages were determined on a national basis. The
long-term program of union scale studies in a few industries was
continued.



During the 1950’s the Bureau also undertook under contract
many surveys, largely of the wage distribution type, for use in
appraisal of the effects of minimum wage actions under the Fair Labor
Standards Act, and as a basis for decisions on minimum wage policy.
These studies continued into the 1960’s as minimum wage coverage
was extended to retail trade, service industries, and other areas of
employment.



The industry wage studies program was enlarged and systematized as
part of the planning process that produced the 80-area community wage
survey proposal. Provision was made for recurring studies in
approximately 50 manufacturing and 20 nonmanufacturing industries,
either nationally or in areas of major concentration. Most of these
industries were scheduled for survey on a 5-year cycle, with others,
predominately in textiles and apparel, on a 3-year cycle. The few
studies based on union scales rather than employer payroll records would
continue on an annual basis. Altogether, about 20 surveys would be
planned for each year.



At the time of their selection, the 70 industries included in the
program accounted for about three-fifths of manufacturing and a third of
nonmanufacturing employment. The 3- or 5-year periodicity for most of
these studies was not ideal, but industry wage structures (that is,
relative rates of pay) tend to change slowly. These surveys developed
data for selected jobs, such as plant maintenance, that cut across
industry lines and alos, for selected processing jobs peculiar to each
industry. Data for the distribution of rates of pay or straight-time
hourly earnings for all production or nonsupervisory workers were
collected, together with information on establishment practices, such as
shift work, and supplementary benefits provisions.



Thus, by the end of the 1950’s two well-articulated
occupational wage survey programs had been developed, one on a local
labor market and the other on an industry basis. The latter program
also provided for a considerable amount of information by labor market
or region. Together, they shed much light on the level and structure of
wages and salaries in the U.S. economy, and provided data for a variety
of governmental uses, private wage and salary decisions, and research.



Supplementary remuneration surveys. Beginnning in the second half
of the 1930’s, a variety of supplements to basic rates of pay began
to assume significance in the U.S. wage structure. These supplements,
some legally required and others established through collective
bargaining or employer personnel policy, provided additional money
income, paid leisure, or income security for workers, and represented a
cost to employers.



It was clear by 1950 that benefits supplementary to basic wages
would continue to account for an increasing share of worker,
compeensation. In 1951, with close industry cooperation, a study was
made of supplementary expenditures in basic steel. In 1953, a
methodological study of problems in the measurement of employer
expenditures on major benefits in manufacturing was undertaken. This was
followed in 1956 by a study of benefit expenditures in the electric and
gas utility industry, conducted as part of an industry wage survey, and
by a 1958 survey of the composition of payroll hours for factory
workers.



Finally, in 1959, a continuous program was launched. The initial
study measured employer expenditures on benefits for production workers
in manufacturing. Data were developed by major industry group, region,
level of wages, size of establishment, collective bargaining coverage,
metropolitan or nonmetropolitan location, and the composition of payroll
hours. This broad-based study was repeated in 1962, following benefit
expenditure surveys in the mining and finance, insurance, and real
estate industries. At the request of the Civil Service Commission, a
special survey was made in 1963 of benefit expenditures for white-collar
workers in metropolitan areas in a broad segment of U.S. industry.
During the following 2 years, numerous studies were conducted in
individual manufacturing and nonmanufacturing industries.



In 1966, an initial survey was made of compensation expenditures in
the entire private nonfarm economy. Data were shown separately for
manufacturing and nonmanufacturing, by establishment size, and, for
nonoffice workers, by union status. This study, in esence, rounded out
more than a decade of experimental work and studies in limited
industrial sectors. Future surveys were to be conducted biennially for
the entire private nonfarm economy, with studies in selected industries
in the intervening years.



Current wage changes. During the years immediately following World
War II, enormous interest developed in changes in wage rates and
employee benefits in major collective bargaining situations. This
reflected the growth of collective bargaining as a mechansim for wage
determination, and the influence that major settlements might have on
the wage bargain–union and nonunion–in other firms and industries.



To facilitate response to inquires, the Bureau began publication of
a monthly report entitled Current Wage Developments in January 1948.
This report, available on a subscription basis, sought to list general
wage changes and changes in benefit provisions in all collective
bargaining settlements covering 1,000 workers or more. Because the
report was based largely on secondary sources, the names of the unions
and employers concerned in the settlements were identified.



This monthly periodical has become a major source of information on
current wage behavior. The year 1954 saw the inclusion for the first
time of quarterly and annual statistical summaries of newly negotiated
wage rate changes. During the mid-1960’s, procedures were devised
for estimating the cost of supplementary benefits, and since 1966, data
have been presented on the total change in compensation in bargaining
settlements affecting 5,000 workers or more. In 1968, statistics were
developed on wage adjustments put into effect from (1) settlements
during the year; (2) deferred changes under agreements negotiated in
earlier years; and (3) provisions for adjustments geared to changes in
the cost of living. The coverage of the report was extended in 1979 to
State and local government collective bargaining settlements involving
5,000 workers or more.



Somewhat related to the monthly report on current wage developments
was the inauguration, also in 1948, of a series of reports on changes in
wages and supplementary benefits in a limited number of key collective
bargaining situations. Through periodic supplements, these wage
chronologies summarized the history of wage and benefit changes
resulting from negotiations between unions and such major employers as
United States Steel, General Motors, and Lockheed Aircraft. The
bargains covered were important in themselves and were thought in many
cases to have significant pattern-setting effects in wage determination.
By the beginning of the 1970’s, about 35 chronologies were being
maintained. Wage rate trends



Several of the survey systems devised after 1947 provided, as a
byproduct, data on the trend of wages for important groups in the
working population.



From the monthly reports on current wage developments, annual
median and mean adjustments and wage indexes were developed for workers
in the universe of major collective bargaining agreements, thus
providing the basis for index computation. On the basis of the
occupational surveys by labor market, annual indexes of wage change were
constructed for office workers, skilled maintenance workers, unskilled
plant workers, and industrial nurses in all metropolitan areas,
nationwide and by region, over the period 1961-74. These indexes were
discontinued in 1974 (with the last report published in 1975) when the
method of computing wage changes in the separate labor markets was
revised and a more comprehensive wage rate index, described in the
following section, was introduced. The annual report on salaries for
selected professional, administrative, technical, and clerical jobs has
provided the basis since 1961 for developing measures of salary trend
for these occupational groups.



In the early 1950’s, the Bureau also began to issue a series
of reports on the trend of salaries for important groups of government
employees. The initial report for white-collar Federal workers covered
the period 1939–50; for city public school teachers, 1925–49; and for
firemen and policemen, 1924–50. These reports were based on
Congressional salary actions for classified employees in the Federal
service; on reports on teacher salaries published by the National
Education Association; and on several data sources for firemen and
policemen.



The reappraisal of the Bureau’s work in wage statistics that
began in 1947 produced over two decades a multidimensional program that
sought to meet, within the budget constraints under which it operated, a
wide variety of governmental and private needs for information. It
provided significant insight into (1) the structure of wage and salary
rates of pay for major groups of workers, manual and white-collar; (2)
the rise of and expenditures for supplementary benefits as part of
compensation for work; and (3) the dynamics and trend of wage
developments. The turbulent years, 1970-84



The 1970–84 period provided a turbulent backdrop for developments
in wage statistics programs. Wide cyclical swings in the economy,
coupled with Federal activities–wage and price controls and guidelines,
minimum wage adjustments, and reevaluation of pay setting for Federal
employees–affected Bureau wage programs. The number and scope of such
programs grew substantially through 1978, but then contracted abruptly and leveled off in the face of Federal budget constraints.



Employment Cost Index. During the early 1970’s, Federal wage
and price controls highlighted a major shortcoming in the Nation’s
economic intelligence system. Information was lacking on changes in
employers’ compensation costs (or labor costs), free from
influences unrelated to cost change, such as emplyment shifts among
occupations and industries with different labor cost levels. Without
such information, it was virtually impossible to gauge the effects of
wage controls in the same way that price controls were assessed on the
basis of the Bureau’s Consumer Price Index. (A similar need had
beeen addressed when the Bureau developed the short-lived,
occupationally based “urban wage rate index” to measure the
War Labor Board’s effectiveness in stabilizing wage rates during
World War II.)



During the 1971–74 controls period, policymakers trying to track
wage rates or compensation costs were faced with a wide array of Bureau
information, all useful for some purposes, including estimates of
average hourly compensation and average hourly and weekly earnings, data
on collectively bargained wage adjustments, and surveys of occupational
pay levels. The various wage measures, unfortunately, gave mixed and
incomplete signals about developments in wage and compesation costs.



It was in this climate that the Bureau began a long-range effort to
develop the Employment Cost Index (ECI), initially called the
“general wage index.” The ECI was designed to be a timely and
comprehensive measure of labor cost change, covering all types of
workers and industries in the economy and all elements of compensation
costs (wages, salaries, and employer costs for employee benefits).



A critical feature of the ECI design was the use of fixed
employment weights by occupation and industry. This feature specifies
measurement of labor cost changes in much the same way that the
Bureau’s Consmuer Price Index measures changes in the prices of a
fixed market basket of goods and services. Like the CPI, the ECI also
yields subindexes (by broad occupational group, industry, union or
nonunion status, and so forth) to provide insights into forces
underlying overall wage and compensation cost trends.



The ECI was developed in stages to meet its design objectives.
Quarterly measures of wage and salary change for workers in the
Nation’s nonfarm economy were first published in 1976. The series
was broadened to cover changes in total compensation costs (employee
benefit costs in addition to wages and salaries) in 1980. The following
year, the ECI was further expanded to cover State and local government
workers. Over the 1976–84 period, the number of ECI subindexes
increased from 21 to 85.



The ECI–designated as a “Principal Federal EConomic
Indicator” in October 1980–now provides measures of quarterly
compensation cost changes for 78 million private-sector workers and 15
million State and local government Employees. It currently excludes
farm, household, and Federal workers, although coverage may be extended
to these groups in the future.



FLSA surveys. PRior to 1970, the Bureau conducted a wide range of
surveys designed to shed light on the impact (3r potential impact) of
changes in the minimum wage and maximum hours provisions of the Fair
Labor Standards Act (FLSA). The surveys, which developed data on
employee wages and weekly hours of work, were narrowly focused on
industries and areas judged to be most heavily affected by changes in
the Federal minimum wage, such as men’s and boy’s shirt
manufacturing, southern sawmills, retail trade, and nonmetropolitan
areas of the South and North Central regions.



The focus broadened in 1970 with the first Bureau study of
distributions of hourly earnings and weekly hours of work for
nonsupervisory employees in the private nonfarm economy. With this
study– designed to estimate the number of worker whose wage rates would
be raised in response to potential changes in the Federal minimum, and
the consequent increases in establishment wage bills–began a period of
accelerated Bureau survey activity related to FLSA and funded by the
Labor Department’s Employment Standards Administration (ESA).
Subsequent years saw several increases in the Federal minimum wage–from
$1.60 to $2 an hour in May 1974 (the first adjustment since 1968),
followed by a series of six adjustments that brought the minimum to
$3.35 an hour on January 1, 1981.



During the mid-1970’s the Bureau conducted surveys of
industries and occupations exempt from FLSA minimum wage and overtime
coverage (including small newspapers; truck-drivers and helpers in local
cartage; and executive, administrative, and professional employees).
Survey results were used by ESA in judging whether an exemption should
be continued, based on the effect it had on wages and hours of work for
affected employees. In the case of executive, administrative, and
professional employees, the survey was used to assist ESA in setting
minimum salaries as one test for FLSA exemption.



The 1977 FLSA Amendments created a Minimum Wage Study Commission to
help “resolve the many controversial issues that have surrounded
the Federal minimum wage and overtime requirements since their origin in
the Fair Labor Standards Act of 1938.” The responsibility for
research on FLSA amendments shifted to the Commission in 1978. The
Commission’s request for Bureau surveys, which continued to be
funded by ESA, resulted in broad-based studies similar to the 1970
survey of nonsupervisory employees in private industry. Also at the
request of the Commission, the Bureau developed a panel study of
establishments with which to gauge the effects of changes in the minimum
wage on employee benefits within individual firms.



The Bureau’s work on FLSA surveys and the panel study ended in
1981, as did the life of the Commission.



Federal pay comparability. As shown earlier, the adoption by the
Congress in 1962 of a comparability pay policy for Federal white-collar
employees led to the conduct by the Bureau of an annual nationwide
occupational survey of salaries in private industry for use in policy
implementation. The 1962 Act was amended by the Federal Pay
Comparability Act of 1970. Under the amended act, the Bureau’s
annual study of Professional, Administrative, Technical, and Clerical
Pay (the PATC survey) continues to provide a statistical basis for
policy considerations. Its results are used by the President’s Pay
Agent (the Secretary of Labor and the Directors of the Office of
Management and Budget and the office of Personnel Management) in making
annual recommendations to the President on pay adjustments needed to
make salaries of Federal white-collar employees comparable to those of
their private-sector counterparts.



The Federal pay determination process, including the PATC survey,
is large, complex, and highly controversial. It now affects they pay of
more than 3 million employees (including the military) and has
substantial impact on the Federal budget; every 1-percent increase in
Federal pay scales costs about $1 billion. The magnitude of the costs
involved and the controversy surrounding the determination process have
triggered no fewer than six procedural reviews and evaluations.



Review of the process began with a General accounting Office audit
in 1972. Subsequent evaluations were conducted by the President’s
Panel on Federal Compensation (the Rockefeller Panel) in 1975, the
Personnel Management Project (the Carter Administration’s task
group on Federal Government reorganization) in 1977, the Grace
Commission in 1982-83, the General Accounting Office again in 1983, and
the Reagan Administration’s Cabinet Council on Management and
Administration in 1984. The reviews generated a variety of
recommendations for improving the pay determination process, including:



* Expansion of the PATC survey to cover smaller establishments and
more private-sector industries;



* Amendment of the 1970 Act to include State and local government
workers in comparability surveys;



* Determination of Federal white-collar pay comparability on an
area, rather than a national, basis for certain types of occupations,
such as technical and clerical jobs; and



* Consideration of employee benefits as well as pay in
comparability determination.



The recommendations from the 1972-77 reviews have already had
direct impact on the Bureau’s PATC survey. To date, improvements
to the study include the establishment of national training programs for
Bureau field representatives, and expansion of coverage to the mining
and construction industries and to smaller establishments in a number of
manufacturing industries.



Pressure to consider employee benefits as well as pay in the
comparability process grew as private-sector benefit costs approached 30
percent of total compensation costs in the late 1970’s. In 1978,
the Bureau began construction of a comprehensive data base on employee
benefits in private industry. Developed from a survey of detailed
employee benefit plan characteristics, the data base has been used
experimentally by the Office of Personnel Management in estimating the
effect of implementing a Total Compensation Comparability concept in
the determination of Federal employee renumeration.



The annual survey of employee benefits was first conducted in 1979,
and has become one of the richest sources of employee benefit data ever
developed. It is nationwide in scope and covers the same industries and
establishment size groups as the PATC survey. Data are collected on
employee participation rates and detailed plan provisions for such
benefits as paid leave, short- and long-term disability benefits, health
and life insurance, and retirement plans.



Shifting program priorities. Expansion and contraction of the
Bureau’s wage statistics programs during 1970-84 followed patterns
of the past: Growth in periods during which the Federal Government had
pressing need for more economic intelligence or for data to administer
Federal law, and cutbacks when dictated by budget constraints.
Difficult priority decisions on cutbacks in wage porgrams included
elimination of the biennial survey of employer expenditures for employee
comppensation; wage chronologies for about 30 major collective
bargaining situations; union wage surveys in construction, printing,
local transit, local trucking, and grocery stores; municipal government
wage surveys in the Nation’s 27 largest cities (initiated during
the early 1970’s); and FLSA surveys. However, the period also
brought development and growth of the quarterly ECI series, construction
of a rich data base on employee benefits, and expansion of the PATC
survey–all contributing to a better understanding of wages and
compensation of the Nation’s working men and women. In retrospect



There are many omissions in this review of 100 years of work by the
Bureau in the compilation of wage statistics. But the main lines of
development have been made clear, and it may not be inappropriate to
recap briefly the significance of this effort.



Without the Bureau’s surveys and studies, with all their
limitations, we would know far less than we do about the money return
for work during the past century in our highly complex and dynamic
economy. The Bureau has provided a reasonably consistent body of
information available from no other source. This reflects an underlying
consistency and continuity of program, despite adaptations necessitated
by fluctuating budgetary levels, special governmental requirements for
survey data, changes in the industrial composition of the working
population, and the increasing complexity of the wage bargain.



In substantial measure, the wage statistics program has been shaped
by Federal Government needs for information for administrative and
policy purposes. But, in line with general Bureau policy, the results
of surveys and studies consistently have been made available to the
public. They have found extensive use over the years in wage
determination through collective bargaining and employer personnel
action, and in university and other private research.