770159799 Kaufman 2e the Economics of Labor Markets and Labor Relations (1)

Celpoue
LT E R
The Labor Market
Economics iis concerned wee ae and

the determinatiot ices and ley produc 1 y. aber eco-
nomics focuses on one pail punter) c a Rages thedetermination of
Labor economics is both an interesting and a useful subject. One reason is

that the topics and issues addressed in labor economics are of direct and immedi-
ate relevance to each of us. The focus of labor economics is on the world of
work—who is working, what types of jobs are available, and why people are
being paid what they are. These issues are not of purely academic interest since
the availability, financial remuneration, and psychological gratification of work
are central to most of our lives.
Labor economics is also a popular course because it deals with a subject with
which many students have firsthand experience. Not many of us have been in-
volved with the intricacies of foreign exchange markets, central banking, or
government tax and expenditure programs. Most students do have some previ-
ous experience with finding employment, the advantages and disadvantages of
particular types of jobs, planning careers, and possibly the operation of labor
unions.
Finally, a third factor that makes labor economics an interesting subject is that
it has so many applications to current social issues and public policy debates.
Should the minimum wage be raised? Should it be abolished? Why is the unem-
ployment rate for blacks twice that for whites? Do unions cause inflation? Will
technological change lead to greater unemployment? While labor economics
will not provide definitive answers to all of these questions, it can provide a very
useful framework in which to analyze them more intelligently.
THE LABOR At the present time, slightly over 115 million people in the United States are
MARKET AND THE employed either part- or full-time and approximately 5 million business firms
ECONOMY provide jobs for these people. How do all of these individual workers and firms
get matched up with one another? What determines the rates of pay that go with
these jobs?
1
Chapter 1 The Labor Market
The central analytic device that economists use to answer these questions is
the concept of a labor market. A market exists whenever there is a good or ser-
vice for which there are buyers (demanders) and sellers (suppliers) who engage
in mutual exchange or trade of the item. This exchange or trade is typically con-
summated when there is a common agreement as to the price per unit that the
buyer must pay the seller.
Just as there are readily observable markets for wheat, steel, and automobiles,
so too is there a readily observable market for labor, or more correctly, the ser-
vices of labor (slavery is illegal). Wheat, steel, and other goods are traded in
commodity or product markets. Labor, on the other hand, is a factor of produc-
tion that is owned by individuals and, in effect, rented to business firms for a
period of time (for instance, an hour or a year) to be combined with other factor
inputs such as land and capital to produce a good or service. Labor is traded in
what economists call a factor market, just as the other factors of production are:
land and capital.
The relationship between the labor market and the product market is illus-
trated in more detail by the circular flow diagram in Figure 1.1. It depicts in a
FIGURE 1.1 A Circular Flow Model of the econo

The household sector gives
rise to a demand for goods
(Dg) in the product market
and supply of labor (S,) in
the labor market. The
business sector gives rise
to a supply of goods (Sg)
and a demand for labor
(D,). The interaction of
demand and supply in the
product market determines
the level of prices and
production (P;, Q;); the
interaction of demand and
supply in the labor market
determines the level of
wages and employment
(Wj, L;).
Chapter 1 The Labor Market 3
highly simplified form the manner in which the market forces of demand and
‘supply connect the product market in the top half of Figure 1.1 and the labor
market in the bottom half. In this poses econerny there are two sets of eco-
nomic agents, hous¢ Is (the left-1 ox) and business firms (the right-hand
box). The goal of Srsiness ae is to pee a DORE and tto do so they produce
goods and services to be sold in the product market. This supply of goods and
services is represented in the product market by the supply curve S,. On the
other side of the product market, households desire as high a material standard
of living as possible, leading to a flow of consumer expenditure on goods and
services, represented by the demand curve Dg. The interaction of demand and
supply in the product market gives rise to an equilibrium price (P,) and quantity
(Q)).
The lower half of Figure 1.1 shows how purchases of labor by firms and sales of
labor by individual household members interact in the labor market to deter-
mine the level of wages and employment in the economy. To produce the goods
and services demanded by households, firms must obtain sufficient amounts of
land, capital, and labor. The desire by firms to supply goods and services in the
product market thus gives rise to a demand for workers in the labor market, rep-
resented by the labor demand curve D,. Similarly, for households to purchase
the goods and services they desire requires that individual household members
seek work in order to earn incomes. The demand for goods and services by house-
holds thus gives rise to a supply of labor to firms, represented by the labor supply
curve S,. As in the product market, the interaction of demand and supply in the
labor market determines the wage rate (W,) and level of employment (L, aie
spend. Pe 1987, he ee miaedan rc by penercan ores amounted

to $2,648 billion, 73 percent of total income created.
Microeconomics and macroeconomics focus on the determination of price
and quantity in the product market, either at the level of the individual firm
orBios ihe aggregate economy.Der mcinngeannen
m half igur the le : . Although labor economics deals
ae a erate set A questions and aries the circular flow diagram in Figure 1.1
suggests that labor economics and microeconomic and macroeconomic analysis
are closely linked because of the important feedback effects that events in one
sphere of the economy have on the other. An increase in consumer expenditure
in the product market, for example, will feed back into the labor market in the
form of a greater demand for labor as firms attempt to increase production. Like-
wise, higher wages in the labor market feed back into the product market as they
affect both business costs and consumer income.
Labor economics and general economic analysis are also closely linked at the
conceptual or theoretical level. As Figure 1.1 illustrates, while the product mar-
ket and labor market are quite distinct, economists use the same theoretical ap-
paratus to analyze each market, namely, the theory of demand and supply in
competitive markets. Just as in microeconomics, where changes in price bring
demand and supply in the product market into a state of balance or equilibrium,
so too in labor economics attention is focused on the role of wages as the “bal-
ancing wheel” in the labor market.
UNIQUE FEATURES There are several ways in which the labor market differs in its structure or char-
OF THE LABOR acteristics relative to product markets. These differences, in turn, have a large
MARKET impact on both the theory and the actual operation of the labor market.
Labor Is Embodied Probably the most important feature of the labor market that distinguishes it
in the Seller from all other markets is that the item being exchanged is embodied in a human
being:! The ownership and possession of a commodity such as wheat is com-
pletely transferable between buyer and seller and, thus, neither party has any
interest in the personal characteristics of the other (for instance, age, sex, color,
or personality); the only interest of each party is to secure the most advantageous
price possible. Likewise, the wheat, since it is an inert commodity, does not
have any preference as to where and to whom it is sold.
The same conditions are not true in the labor market. Labor
as a service is
inseparable from the person providing it and, thus, the worker supplying labor
and the firm buying it must have a direct, personal relationship with each other.
This, and the fact that human beings have definite preferences with respect to
the conditions they work under, causes the exchange in the labor market to be
determined not only by the price of labor but also by a host of noneconomic
factors that are largely absent in commodity markets. These noneconomic fac-
tors are partly physical in nature, such as the risk of injury on the job or the
pleasantness of the work environment, and partly social in nature, such as the
prestige of the job, the race or gender of workmates, and the attitude of the man-
agement. The importance of these factors for labor economics is that the deci-
sion of workers concerning for whom to work and the decision of firms regarding
whom to hire are based on a complex package of considerations, including not
only the wage but all the nonpecuniary advantages and disadvantages associated
with the job or worker.
The Long-Term A second significant feature of the labor market is the long-term nature of the
Nature of the . With many goods and services, such as automobiles,
Employment clothing, or airline travel, buyers and sellers may switch from one to another on
Relationship a daily basis in pursuit of the most attractive price. Such shopping around occurs
in the labor market, but in a more attenuated form due to the costs to both
workers and firms of frequent job-changing. Employers, for example, find it
to their advantage to cultivate stable workforces, since they make substantial
investments in workers in the form of costs of hiring and, more importantly,
training and experience. Likewise, individual workers find it to their advantage
to remain with one employer for a considerable time, particularly as they get
beyond their teen years and early twenties. One reason for this is economic in
‘Eli Ginzberg, The Human Economy (New York: McGraw-Hill, 1976): 3-13.
Chapter 1. The Labor Market 5
nature, since wages and fringe benefits normally increase with tenure on the job;
another more psychological reason is related to the value people place on secu-
rity and familiar surroundings. How long do most people stay on one job? One
recent study found the typical worker will remain on a job for about eight years.
The period varies by age; the average duration of a job for a teenager is only
three months, while 40 percent of workers over the age of 40 hold jobs for 20
years or more.’
The most important implication of the long-term employment relationship
between CORES and firms is that it n
$ : ya h > abili yes to clear the market. The clas-
sic type of auction market where prices rise and fallcee exists for commodities,
such as wheat, or financial issues, such as stocks and bonds. In the wheat mar-
ket, an excess supply of wheat quickly leads to a drop in its price as sellers under-
bid each other to attract a buyer; buyers, in turn, have little reason not to switch
from one seller to another since all bushels of wheat are exactly the same. In the
labor market, however, an excess supply of labor typically does not lead to a fall
i es. While workers who are unemployed and want a job might
offer to work for a firm at lower wages, most firms would find it unprofitable to
hire them because the costs of hiring and training, as well as the disruptive influ-
ence on morale, would far outweigh the savings in lower wages. Thus, while in
he la (satticutaely
in the ee fea ple iibdrhiniartinaena SS
frequent turnover. Because of this sluggishness of wages, an imbalance between
the demand and supply of labor may persist for a considerable length of time
before wages rise or fall enough to bring about the necessary adjustments in the
labor market to restore equilibrium. However, in labor markets where the em-
ployment relationship is short-term and turnover costs are negligible, as for day
laborers or migrant farm workers, wages exhibit a flexibility similar to com-
modity markets.’
Heterogeneity of A third feature of the labor market that distinguishes it from most product mar-
Workers and Jobs kets is the extreme diversity in the characteristics of the good being traded. For
agricultural commodities, raw materials, or semifinished products like steel, each
unit is identical and the decision to buy or sell is made strictly on the basis of
price. Other products, particularly consumer goods such as automobiles, cereals,
and so on, are differentiated from each other by their physical appearance,
quality, and brand. The decision to buy and sell these products is influenced not
only by price, but also by these nonprice factors.
2Robert E. Hall, “The Importance of Lifetime Jobs in the U.S. Economy,” American Economic
Review 72, no. 4 (September 1982): 716-724. The average job duration for teenagers is provided in
Kim B. Clark and Lawrence H. Summers, “Labor Market Dynamics and Unemployment: A Recon-
sideration,” Brookings Papers on Economic Activity 1 (Washington: Brookings Institution, 1979):
13-60.
3See Lloyd Fisher, The Harvest Labor Market in California (Cambridge, Mass.: Harvard University
Press, 1953).
The same consideration holds true in the labor market except the degree of
differentiation in the characteristics of jobs and workers is frequently much
greater. Individual workers differ by age, race, sex, education, experience, skills,
and complex personality factors, such as motivation and congeniality. While a
firm may have several types of machinery to choose from in making its product,
every single worker it interviews is in some way unique. In choosing between
employers, workers face the same diversity in characteristics of jobs. Employers
and jobs, for example, differ in the type and difficulty of the work, commuting
distance, fringe benefits, and quality of employee relations, as well as wages.
This extreme diversity in the characteristics of workers and jobs has two con-
sequences for the operation of the labor market. The first is to make exchange in
the labor market a function not only of wage rates but also of the many nonwage
crmmmmemniaticoataintilaisiid alii ieanceiaaanimemnmnsen ws" Workers
do not choose between jobs solely on the basis of pay, but also take into account
a wide range of nonpecuniary factors; likewise employers hire workers partly
with an eye to the wage each will work for, but also with regard to many non-
economic factors. The second way in which the differentiation of jobs and work-
ers affects the labor market is by complicating the acquisition and evaluation:of
a
place.‘ In the wheat market, for example, each bushel of wheat is alike and the
buyer need only acquire information about the price that various sellers are de-
manding. In the labor market, however, both buyers and sellers must invest
much more time and effort in evaluating the many nonpecuniary and hard-to-
measure characteristics that differentiate each worker and job. The result is that
exchange in the labor market is both more costly to undertake and less likely to
result in the most efficient match of buyer and seller compared to markets where
the product is more nearly standardized.
The Multiplicity A fourth characteristic that distinguishes the labor market is the multiplicity of
of Markets individual submarkets. A labor market is the area over which demand and sup-
ply determine the going wage rate for labor. While the concept of one national
labor market, as shown in Figure 1.1, is a useful conceptual device, in reality
there is a multiplicity of markets separated by geographic location, occupation,
skill,and so on. In analyzing the determination of wages, it is obvious that a
distinction should be made between the market for plumbers, teachers, and |
bankers, for example, since the supply and demand for each occupation is likely
to be quite distinct. Likewise, geographical location gives rise to distinct labor
markets; the demand and supply for legal secretaries in Los Angeles is likely to
be distinct from that in Boston. For some-occupations, such as college professor,
lawyer, or baseball player, the labor market may be national or even inter-
national in scope. Although many product markets are also less than national in
scope, few are as fragmented as the labor market.
*See George Stigler, “Information in the Labor Market,” Journal of Political Economy 70, Part 2
(October 1962): 94-105.
The boundaries between many of the individual labor markets are relatively
porous so workers can flow from one market to another in response to changes in
wages; a teacher can become a flight attendant, or a plumber in Michigan can
move to California. This movement between markets becomes progressively
more difficult the greater the disparity in skills or geographic distance. For ex-
ample, a truck driver may be able to compete for an auto assembly job, but would
not be in competition in the market for orthopedic surgeons.
LABOR MARKET There is a useful distinction between labor market outcomes and the labor mar-
OUTCOMES ket process. Labor market outcomes are the observed events, behavior, or de-
velopments in the labor market we are trying:to understand and explain; the
outcomes represent the end-product or final result of the operation of the labor
market. In order to understand the cause of the outcomes, however, it is neces-
sary to understand the process that gave rise to them, that is, the mechanics of
how the labor market works.
A great number of labor market outcomes are examined in this book. Of
these, six in particular are the central focus of research in labor economics. °
The Changing Level Labor is a factor input into the production of goods and services and, thus, the
and Composition of quantity and quality of labor that individuals are willing to supply for market
Labor Supply work is an important determinant of the economy’s level of production and rate
of growth. In calculating the amount of labor input available to the economy, it
is important to realize that there are several different dimensions of labor ne
5)
During the post-World War II period, some dramatic changes in the size and
composition of labor supply have taken place. Several aspects of this are illus-
trated in Figure 1.2. Graph (a) shows the trend over time in the average number
of hours worked per year by full-time production workers for selected years be-
tween 1947 and 1984. In 1947, average hours worked per year were 2,236; by
1984 they had fallen to 1,920 per year, a 14.1 percent decline. Thus, one im-
portant trend in labor supply has been the long-term decline in annual hours
of work.
Graph (b) illustrates the trend over time in two other dimensions of labor
supply. The first, the labor force participation rate, represents the fraction of the
noninstitutionalized population (persons not in a prison or mental hospital) 16
years of age or over who were either employed or looking for work. The graph
shows that in 1947, 87 percent of men but only 32 percent of women were partici-
pating in the labor force. Between 1947 and 1987, these percentages changed
5These and other important outcomes in the labor market are examined in more detail in Richard
Freeman, “The Evolution of the American Labor Market,” in Martin Feldstein, ed., The American
Economy in Transition (Chicago: University of Chicago Press, 1980):349—-414.
8 Chapter 1 The Labor Market
FIGURE 1.2 Trends in Three Dimensions of Labor Supply
(b) Labor Force Participation Rate (LFPR)of = =«_—

Men and Women, Percentage ofLabor
(a) Annual Hours of Work Force with College Degree - s
Hours per Percent
Year
2300 100 =
Male LFPR
2200 80
2100 60
Female LFPR__
2000 40
1900 20 College educated _
1947 1958 1968 1978 1984 1947 1958 1968 1978 1987
sources: Herbert Northrup and Theresa Greiss, “The Decline in Average Annual Hours Worked in the United States, 1947-1979," Journal of Labor
Research (Spring 1983): 97; Bureau of Labor Statistics, Handbook of Labor Statistics, 1985, Tables 2,61; and Employment and Earnings (January 1988),
Table A—2. The annual hours of work for 1984 were extrapolated from hours of work in 1978, based on data in Shirley J. Smith, “Estimating Annual Hours
of Labor Force Activity,’ Monthly Labor Review (February 1983): Table 9.
rather noticeably; the labor force participation rate for men declined to 76 per-
cent, while that for women increased to 56 percent. A second important trend
in labor supply, then, is that a somewhat smaller proportion of men are in the
labor force today, counterbalanced by a much larger proportion of women.
Also shown in the bottom part of graph (b) is the trend in one measure of the
quality of labor supply—the percentage of the labor force having a college de-
gree. Between 1952 and 1986 the proportion of the people in the labor force
who were college graduates more than doubled, from 8 percent to 21 percent. A
third important trend in labor supply, therefore, is the rapid increase in the edu-
cational attainment of the workforce. Taken together, these three trends in
labor supply have led to a fundamental shift over the last three decades in who is
working, the hours devoted to work, and the skills that individual workers bring
to the job.
The Changing Level Just as the number and characteristics of people wanting to work have changed
and Composition of significantly over the last 40 years, so have the number and the characteristics of
Labor Demand jobs that employers have to offer. One important feature of employers’ demand
for labor is the growth over time i the et ee of ig made available. As
the labor force has grown year after year, has the number of new jobs created
Chapter 1 =The Labor Market 9
been able to keep pace? In general, the answer has been yes, witnessed by the
large-scale expansion of employment in the economy from 60 million to over
114 million between 1950 and 1987.
A second equally significant feature of labor demand is the types of jobs that
are available in the economy. Demand for the goods and services produced in
the product market of the economy shifts over time. In reaction to this, the
demand for labor in some occupations and industries expands while it contracts
in others, leading to large changes in the skills, education, and geographic loca-
tions required of workers. One particular aspect of the changing composition of
labor demand is illustrated by the data in Table 1.1, which show the distribution
of employment in the economy among 10 occupational groups for the years 1950
and 1987. Ea live digiaicncecne Me
fe emen sly, expandit rom one-third of tota loyment to
. Among these white-collar occupations, two have grown
exceptionally quickly i ns (such as engi-
neers and teachers) and clerical occupations (such as secretaries). As the propor-
tion of white-collar jobs grew, the 4
from 39.1 percent of total employment in 1950 to only 27.6 percent in 1987.
The decline has been particularly rapid among jobs in the “operatives” occupa-
tion (including assemblers, machine operators, and truck drivers), reflecting in
part the relative decline of manufacturing as an important source of jobs in the
economy. Also shown in Table 1.1 are two other major occupational groups,
service workers and farm workers. Not unexpectedly, the proportion of service-
related jobs outside of private household work (such as waitresses, dental hygien-
ists, and guards) has grown over time while farm-related jobs have plunged from
12.4 percent of total employment in 1950 to only 3.1 percent in 1987.
TABLE 1.1 Percentage Share of

Changes in the Total Employment
Composition of Labor Occupation 1950 19874
Demand, 1950-1987 White Collar 37.5% 55.8%
Professional and technical > 75 16.6
Managers and administrators 10.8 12.5
Sales workers 6.4 8.8
~ Clerical workers 12.8 17.9
Blue Collar 39.1 27.6
Craft and kindred : 12.9 12.1
Operatives 20.3 iile2
Nonfarm laborers 5.9 43
Service Workers 11.0 13.4
Private household 3.2 8
Service workers, except private household 78 12.6
Farm Workers 12.4 3.1
fo a a Ee A ee a ae a a ee
2A new occupational classification system was introduced in connection with the 1980 census, making the pub-
lished data for 1950 and 1987 noncomparable. The 1987 data were adjusted by the author to correspond roughly to
the old system.
sources: Bureau of the Census, Statistical Abstract of the United States (1970), Table 334; Bureau of Labor Statistics,
Employment and Earnings (January 1988), Table 22.
i
TABLE 1.2 Percent Change
The Structure of in Average
Average Weekly Hourly Earnings,
Earnings Occupation Earnings, 1986° Industry 1976-1986°
Chemical engineer $721 Hospitals 124%
Economist 704 Tires 115
Pharmacist 607 Photographic equipment 104
Accountant 478 Banking 94
Electrician 473 Knitting mills 83
Registered nurse 460 Basic steel 80
Police officer 431 Highway construction 73
Truck driver 371 Leather footwear 7)
Bank teller 231 Meatpacking 62
Cashier 181 Newspapers 59
Full-time wage and salary workers
> Production workers
sources: Earl F Mellor, “Weekly Earnings in 1986: A Look at More than 200 Occupations,” Monthly Labor Review 110,
no.6 (June 1987): 41-46; Bureau of Labor Statistics, Employment and Earnings (March 1977 and March 1987),
Table C.2.
The Structure A third labor market outcome of considerable importance concerns the structure
of Earnings i i among individual workers, occupations, in-
dustries, and geographic areas. The concept of earnings may relate to average
hourly earnings, weekly earnings, annual earnings, or even lifetime earnings;
whichever concept is focused on, a central task of economic analysis is to ex-
plain what gives rise to the observed differential in earnings among workers at a
point in time and why these differentials change over time.
Some of the issues involved here are illustrated in Table 1.2. The left-hand
portion contains data on the average weekly earnings in 1986 of full-time work-
ers in 10 different occupations. There is a considerable range or dispersion in
earnings among these occupations; the weekly earnings of chemical engineers,
for example, were three times as large as those of bank tellers while truck drivers
earned twice as much as cashiers. One of the tasks of labor economics is to iden-
tify the factors responsible for this particular spread in earnings. To what extent
are these earnings differentials related to the educational qualifications needed
for each occupation? What about the role of unionization? Does the fact that
the occupation employs predominantly men or women make a difference?
A second important aspect of the structure of earnings concerns the change in
earnings differentials over time—why the earnings of some workers grow: rela-
“ivel.sapilhy whileshs.eamings foncherlasalaag! This issue is illustrated in
the righthand portion of Table 1.2, which contains data on the percentage
change in average hourly earnings between 1976 and 1986 for workers in 10 dif-
ferent industries. Much as with the structure of earnings at a point in time, the
change in earnings exhibits a wide dispersion among different groups of workers.
Workers in hospitals, for example, gained an increase in wage rates (124 per-
cent) that was over twice as large as that received by workers in the newspaper
industry (59 percent). A central focus of economic research in recent years has
been to identify the factors responsible for this pattern of earnings growth.
Changes in the Few labor market outcomes gain more public attention and concern than does
Level and UNS aen A > does nc
Composition of
Unemployment person who hasa paying job is Baned as “employed”; a person who does not
3/1)
Unemployment entails costs both to the individual worker because of the conse-
quent loss of earnings and to society because of the loss of the goods and services
that could have been produced with that labor input.
Two aspects of unemployment are of particular importance. The first is the
overall level of unemployment and its change over time. This is illustrated in
Figure 1.3 by the line marked “Total,” which shows the annual unemployment
rate for workers of all demographic groups for the years between 1950 and 1987.
It is apparent that the movement in the unemployment rate has two dimensions:
. During EE aerecession th as 1974

through 1975, and 1980 through 1982), for example, the unemployment rate
temporarily increases as firms lay off workers as a result of the decline in their
sales; with the end of the recession, unemployment falls again as economic ac-
tivity recovers. What Figure 1.3 also shows, however, and what is a serious con-
cern to policymakers, is that once the ups and downs of the business cycle are
netted-out there has taken place over the last 30 years aclear increase inthe
average level of unemployment in the economy.,Between 1950 and 1960, for
FIGURE 1.3 Level and Composition of Unemployment, 1950-1987
Unemployment
rate (%)
27 |
24
Zi
>
18
Teenagers
s :
12
g | Blacks
6
Total
3
1950 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 1987
and Handbook of Labor Statistics, 1985, Table 26.

SOURCES: Bureau of Labor, Employment and Earnings 35, (January 1988): Table A-1, A—3;
example, the average level of unemployment was 4.5 percent; between 1977
and 1987, it rose sharply to 7.2 percent.
Figure 1.3 also points out a second area of public concern regarding unem-
ployment, which is its unequal distribution among different groups:in the labor
force. In addition to the unemployment rate for all workers, Figure 1.3 also
shows the unemployment rate for two particular demographic groups, black
workers and teenaged workers. Both groups experience much higher levels of
unemployment: the unemployment rate for black workers is generally double the
nt is even
greater, Thus, while all groups in the
labor force are hurt by unemployment, some groups bear a much larger burden
than others.
The Continued The fifth labor market outcome concerns the growth rates of money wages and
Growth in Money real wages. The data in Figure 1.4 chart the annual percentage growth in money
Wage Rates and wage rates and real wage rates received by workers between 1961 and 1987.
the Diminished Money wage rates are the number of dollars and cents received per hour of work;
Growth in Real real wages are measured by dividing the money wage by an index of prices in the
Wages economy to determine the actual purchasing power of the money wage. Thus, if
the money wage should increase by 10 percent and the level of prices by 6 per-
cent, the real wage has gone up 4 percent.
Several features stand out with respect to the trend in money wages and real
wages. The first is the marked acceleration that took place in the 1960s and
1970s in the annual growth rate of money wages. During the early 1960s, work-
FIGURE 1.4
Percentage Change
in Money Wages
and Real Wages,
1961-1987
Money Wages
Real Wages
1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 19851987
sources: Bureau of Labor Statistics, Handbook of Labor Statistics, 1985, Table 92; Monthly Labor Review (April
1988): Table 44.
ers’ paychecks were increasing by an average of 3 to 4 percent a year; by the late

1970s, this rate had more than doubled to the 9 to> 10 percent range. It is com-
mon knowledge that the i i ta-
matically during this period, reaching double-digit figures in 1979 through 1981.
Were these two events related? In particular, did the faster growth in wages
cause the rise in prices or, alternatively, did rising prices pull up wages? This
question is frequently a source of puzzlement, for people know that wages and
prices are somehow related, but many times they are not sure just how. One of
the major tasks of this book is to explain the precise role of wages in the infla-
tionary process, and, in particular, why in the 1970s wages and prices seemed to
ratchet upwards, while in the mid-1980s the two have remained so stable.
A second feature of interest in Figure 1.4 concerns the trend over time in real
wages. In the early 1960s, real wages were growing by about 2.5 percent a year.
While this seems relatively unimpressive, through the magic of compound inter-
est a 2.5 percent growth rate translates into a doubling of per capita real income
every 28 years or so. Observe, however, what has happened to real wage growth
since the early 1970s. Not only has the growth of real wages slowed down, in
several years real wages have actually declined as prices rose faster than money
wa oe
eained aS loti TRS a
What was responsible for the decline in real wage growth? Most people would
immediately say “inflation.” Unfortunately, this answer mistakes the symptoms
for the cause; both the increase in inflation and the decline in real wages were
linked, to a significant degree, to a third development—the sharp decline after
1973 in the growth of productivity (output per employee hour). A central lesson
of this book is that productivity growth is the ultimate source of a country’s rising
standard
of living. Because this fact is so important, we devote considerable at-
tention to explaining both the link between productivity and real wages, and
why productivity growth has stagnated in the United States.
Poverty and Income The sixth labor market outcome of importance is the in-
Inequality come inequality. In 1986, the median family income in the United States was
$29,460. (This income figure includes both “earned” and “unearned” income,
the former being composed of earnings from work, such as wages and salaries,
and the latter being composed of nonwork-related sources of income such as
stock dividends or Social Security payments.) Some families, however, received
considerably more income than this while others received much less.
This is illustrated in Figure 1.5 which shows the distribution of families by
income level. In 1986, 6.8 percent of families had annual incomes of $75,000
or more. Whether $75,000 is enough to be considered “rich” is an open ques-
tion, but there is no doubt that this small minority of families was quite well-off
relative to their counterparts at the bottom end of the income distribution.
Among the have-nots in American society were the 13.6 percent of families
who fell below the official poverty line as calculated by the federal government.
FIGURE 1.5 The Pattern of Poverty and Income Inequality, 1986
Percent of 12
families
10
Poverty line (family

of four)=$11,203
j= 5 10- 15- 20-, 25- @0- 35 40- 45- 50 SR 0. 68. 70. 7h.
5 10 15 20 25 30. 35. 40.45 50. 55 60.65. (0 775 3
Annual _.
income |
source: Bureau of the Census, Money Income and Poverty Status of Families and Persons in the United States, 1986, Series P-60, No. 15 (Washington:
G.P.0., 1987).
In 1986, the poverty line for a family of four was $11,203. (The poverty line
varies with family size, being somewhat lower for smaller families, and higher for
larger ones. )
These data point out several interesting features concerning the income distri-
bution. Perhaps most important, why do some people earn barely enough to eke
by while others have incomes 5, 10, or 15 times as large? In particular, to what
degree is economic success tied to such factors as education, innate ability, fam-
ily background and class, inheritance, and simple luck. A second interesting fea-
ture of the income distribution is that it isskewed to the right. majority
A of
families are in the income range of $15,000 to $40,000, but a relatively small
minority of families are spread out in the right-hand tail of the distribution with
quite large incomes. If ability is normally distributed (a bell-shaped distribution
where the left- and right-hand tails are symmetric around the mean) among in-
dividuals, as is often thought to be the case, why doesn’t the income distribution °
have a similar shape? A final issue of considerable importance is the impact that
government has on the degree of economic inequality in the United States. To
what extent do the income-tax system or transfer programs such as Social Secu-
rity reduce inequality? Likewise, how successful have government welfare pro-
grams been in reducing poverty? Do these programs actually bring about an
increase in the number of people in poverty, as some critics charge?
THE LABOR The six developments outlined above represent the most important labor market
MARKET PROCESS outcomes that are covered in this book. The question that immediately follows
is what caused these labor market outcomes? Why have annual hours of work
declined? Why has the level of unemployment in the economy grown over the
last 30 years? To answer these questions, it is necessary to study the labor market
process—how the labor market works, the nature of the cause and effect rela-
tionships in the labor market, and the major actors or institutions in the market.
The labor market process involves three broad forces that together determine
each of the outcomes. These three forces are, respectively, market forces, in-
stitutional forces, and sociological forces.
Market Forces The first process to consider in the determination of the price and allocation of
labor is the market forces of supply and demand in the labor market. While a
number of important noneconomic factors influence the pricing and distribution
of labor, most economists would agree that the interaction of the market forces
of supply and demand in the labor market isthe single most important determi-
nant of labor market outcomes.
On the demand side of each labor market are all the business firms, plus various
nonprofit organizations such as government, which all actively compete for work-
ers of a particular skill or trade. The demand for labor on the part of business
firms is a derived demand—derived from the demand for the good or service
. The business firm has as its primary goal the maximization
of profit, and its demand for labor arises only to the extent that labor as a factor
input is necessary to the production of the firm’s product. This goal of profit
maximization motivates the business firm to economize on labor as much as pos-
sible and to seek out those workers who are likely to be the most productive and
efficient and who will work for the least remuneration. Given the wide range of
products and services produced in the American economy, and the great diver-
sity in skills and training that such production requires, the demand for labor
among business firms is quite heterogeneous.
The supply side ofeach labor market is comprised of all the individuals who
are working or looking for work. Just as the demand for labor is derived from the
demand for the business firm’s product, the supply of labor is derived from the
a eT Gee Much as with the business firm, the indi-

vidual worker is motivated by self-interest and the desire to maximize his or her
well-being to seek out the type and location of work that yield the highest re-
turn. Given the diverse backgrounds, education, skills, and preferences of the
American labor force, the supply of labor is likely to be as variegated as is the
demand for labor.
Because it happens more or less routinely, the efficiency with which a market
system allocates labor in the economy is often taken for granted. It is, however,
a monumental task. How do all the different jobs and all the diverse individuals
in any one of our larger cities get matched up? How is it that there are enough
truck drivers to deliver food to grocery stores, enough tellers to wait on custom-
ers at banks, and enough telephone repairers to keep the phone system working?
A market system solves this problem through two related mechanisms—
cet SE ERASE This process is illustrated in Figures
1.6 and 1.7, which show, respectively, the determination of the equilibrium
FIGURE 1.6 The Determination of Wage Rates by Demand and Supply

Equilibrium in the labor
market occurs at the wage
W, (point E) where labor
demand and supply are
equal. At the wage W,,
labor demand (point V) is
greater than labor supply
(point X). In this situation,
competition among firms
for scarce labor will cause
wages to rise to We,
leading firms to cut back
on their desired level of
employment from L; to Le,
while inducing L_ — L,
additional people to supply
labor in this market. At
the wage W,, supply
is greater than demand,
and competition will force
the wage to fall to We. Lo Le Ly Labor (L)
ges
ge.
Figure 1.6 shows a representative demand curve D, on the part of employers
and supply curve S, on the part of workers in some particular labor market, say
the market for truck drivers in New York City. The vertical axis of Figure 1.6
measures the money wage rate W, while the horizontal axis measures the num-
ber of workers L. The labor demand curve D, depicts the relationship between
the quantity of labor demanded by firms in this market and the wage rate they
have to pay if other things are held constant, such as the state of the economy.
The demand curve slopes downward to the right, illustrating the imverse»rela-
tionship that exists between wages and firms’ demand for labor—the higher the
i Thus, at
the wage W,, firms in this market would want to hire L, truck drivers (point V);
should the wage rise to W,, however, the demand for truck drivers would decline
to L, (point Z).
The labor supply curve S, depicts the relationship between the wage rate and '
the number of people willing to work in this market. The»supply curve slopes’
,, the quantity of
labor supplied wou point X); a rise in the wage to W, would increase the
number of people wanting to be truck drivers in the New York City market to L,
(point Y).
s. To see how this works, assume that the wage rate for truck drivers in New
York City is only W,. At this relatively low wage, delivery companies in New York
City would find it profitable to hire L, workers (point V), but only L, (point X)
Chapter 1 ~The Labor Market 7
people would offer to work at this low wage. In this situation demand is greater
than supply, and the competition among delivery companies to hire drivers
should result in wages being bid up, This rise in wages would continue until both
enough new workers were attracted into this occupation and enough firms de-
creased their desired employment level until a balance or market equilibrium is
reached where demand just equals supply. In Figure 1.6 this is at point E, with a
wage of W, and a level of employment of L;.
This process also works in reverse. What if the wage rate is W, and supply
(point Y) is greater than demand (point Z) in Figure 1.6? The excess supply of
people wanting to be truck drivers should put downward pressure on wages as
delivery firms find they have many qualified applicants willing to work. In such a
situation, wage rates in the truck drivers’ market will decline, if not through
actual cuts in wage rates by employers, then at least in relative terms as wages
in other occupations continue to grow while wages for truck drivers remain
stagnant.° This decline in wages will eventually bring the market back to equi-
librium as firms are motivated to increase hiring and some workers leave this
market for higher-paying jobs elsewhere. This process of changes in wages and
labor mobility will continue until demand and supply are again in balance.
Another way to see the essential role that wages and labor mobility play in the
labor market process is to examine how the market adjusts to a shift in the labor
demand or supply curve. This is illustrated in graphs (a) and (b) of Figure 1.7.
The demand curve D, and supply curve S,, reproduced from Figure 1.6, each
show how the quantity oflabor demanded orsupplied varies with changes inthe
wage rate, other things held constant. In real world labor markets, however, vari-
ous outside events, such as the rise and fall of the economy over the business
cycle, continually shift the labor demand curve to the right or left, while the
labor supply curve undergoes similar shifts due to changes over time in the size of
the population, the number of graduates of educational institutions, and other
such factors.
How does the labor market adjust to shifts in demand and supply? Graph (a)
of Figure 1.7 illustrates the case of a demand increase. Given the labor demand
curve D, and the supply curve S,, the equilibrium wage in this market (again
assumed to be truck drivers in New York City), is W, (point X). Assuming the
economy of the New York City area now experiences a boom, at any given wage,
such as W,, trucking companies will want more drivers to handle the increased
business. This event is illustrated by shifting the labor demand curve to the right
from D, to D,. At the wage W, the demand for truck drivers has risen from L,
(point X) to L; (point Y). It is evident that the labor market is now out of equi-
librium, for, at the wage W,, supply (point X) is less than demand (point Y).
What will happen?
The sequence of events will be exactly the same as for the situation of excess
demand depicted in Figure 1.6. The wage rate will rise from the original equi-
6If changes in relative wages bring about equilibrium in the market, then the vertical axis of
Figure 1.6 needs to be changed from W (the wage rate of truck drivers) to W/Wa, the wage of truck
drivers relative to wages of all other occupations (Wa).
FIGURE 1.7. The Market panne toseit in oLreny and Ey

Graph (a) shows the
adjustment of the labor
market to an increase in
the demand for labor. The
initial equilibrium wage/
employment level is W,/L,
(point X). The increase in
labor demand shifts the
demand curve to the right
from D, to D2, creating an
excess demand for labor of
L3 — L,. The result is that & Oe
the wage rises until a new (peti eek eat a On OSSera caety
equilibrium is established
at the wage/employment =
level W,/L, (point Z). Graph eee
(b) illustrates the market
adjustment to an increase
in labor supply. The wage/
employment level is initially
W,/L, (point X). The
increase in labor supply
shifts the supply curve to
Sp, resulting in an excess
supply of labor of L3 — Ly.
SE
SY,
SE
The result is a decline in Soe
ie
ae
a oe
ee
a
a
—
the wage until a new

equilibrium is reached at
the wage/employment level
of W/L, (point Z).
ow
ee
oe
—
eee
ant
Gee
ome
ce
on
— ee
ee
ee
ee
=\
librium of W, (point X) to a new equilibrium of W, (point Z) where demand and

supply are again equal. A necessary second step is that this tseinthewagealso
induces individuals in different occupations (such as shipping clerks) or different
geographic locations (such as Philadelphia) to quit their jobs and move to take
the now higher-paying jobs in the New York City truck drivers market. As a
result, equilibrium employment rises from L, to L, (but not all the way to L;).
Graph (b) illustrates the market adjustment to a rightward shift of the labor
supply curve, say because of a liberalized immigration law. At the original equi-
librium wage of W,, the number of people willing to work as truck drivers in
New York City is L,; (point X) on the labor supply curve S,. Assume now that
because of the change in the law, a large number of persons from foreign coun-
tries immigrate to the New York City area. At the wage W, trucking companies
now have L; (point Y) people who want a driver's job, an outcome that is repre-
sented by the rightward shift of the labor supply curve from S, to S,. At the wage
W, a situation of excess supply has emerged that, as demonstrated earlier, should
cause a decline in the wage until a new equilibrium is reached at W, (point Z).
The examples in Figures 1.6 and 1.7 illustrate how changes in wages and the
balance in the labor market between demand and supply. In this Tr) it is
important to point out that the process of labor mobility not only helps achieve
an efficient allocation of labor, but also acts as an important
check on the em-
ployment practices and working conditions offered by employers. Because work-
ers are free to quit one employer and find another, a trucking company that
requires excessive work hours, uses unsafe equipment, or treats employees in an
arbitrary or authoritarian manner will find that it is unable to attract and keep a
work force. The competition among employers for labor and the existence of
labor mobility, therefore, automatically police the social conditions of labor.
This discussion of the labor market process is not a new observation; it was
perhaps best analyzed over 200 years ago by Adam Smith. In describing the labor
market, Smith said:
The whole of the advantages and disadvantages of the different employments of labour and
stock must, in the same neighborhood, be-either perfectly equal or continually tending to equal-
ity. If in the same neighborhood, there was any employment evidently more or less advan-
tageous than the rest, so many people would crowd into it in the one case, and so many would
desert it in the other, that its advantages would soon return to the level of other employments.'
This is as clear a statement as possible of how the forces of supply and demand
determine wages and how, in reaction to these wage rates, people flow from one
employment to another seeking their highest advantage. Smith’s second bril-
“Adam Smith, The Wealth of Nations (Chicago: University of Chicago Press, 1976): 111.
Empirical Evidence
The Market for Teachers
Do market forces really work as described above, or is this just theory? For some
evidence on this issue, consider the market for teachers. Graph (a) of Figure 1.8
illustrates data on the number of job openings for elementary and secondary
school teachers and the number of new college graduates qualified to teach for
the years 1960 through 1984. The number of job openings is a measure of the
strength of labor demand; the number of new college graduates qualified to teach
is a measure of labor supply. Graph (b) shows data on the relative earnings ratio
of teachers, measured as the annual earnings of teachers as a percent of the
median income of men employed year-round full-time in the economy.
Graph (a) shows that the market for teachers went through a distinct boom
and bust cycle over this 25-year period. Up to about 1967, the demand for
teachers grew faster than the supply as the children born in the “baby boom” of
the 1950s passed through the school system. In a situation of excess demand
such as this, the model of demand and supply predicts that two things should
happen to bring the market back to equilibrium: first, the relative earnings of
teachers should rise and, second, the rise in earnings should lead to an expan-
sion of the supply of teachers as people who had previously been teachers (for
example, women who had quit teaching to raise children) are lured back into
the field, and as more people decide to obtain teaching degrees. These predic-
tions are borne out by the data. Relative earnings of teachers rose substantially
during the 1960s as communities upgraded teacher salaries in an attempt to find
and keep scarce teaching personnel. Graph (a) shows that the supply of new
teachers also expanded rapidly as more people majored in education in college
and obtained teaching certificates.*®
In 1967 and 1968, demand and supply in the teachers’ market were roughly in
balance. Beginning in 1969, however, demand for school teachers fell off sharply
as the baby boom children graduated and entered the job market. Compounding
the situation in the mid-1970s was the emergence of an actual “baby bust” as
young adults postponed marriage and the birth of children. Graph (a) shows that
this decline in demand created a glut in the teachers’ market as the number of
people trained to be teachers far exceeded the number of new job openings. This
imbalance was particularly acute in 1970 through 1973 as students, lured by the
*In representing this sequence of events with a demand/supply diagram, it is important to distin-
guish clearly between movements along a labor supply curve and shifts in the curve. At any one
point in time, there is an upward sloping supply curve for teachers, illustrating that, for the given
stock of people trained to be teachers, a higher salary will induce more of them to offer to work (for
example, the housewives cited above might reenter the job market). Over time, however, this
higher salary will also lead more people to obtain teaching certificates, which will then shift the
supply curve for teachers to the right, showing that at any given salary more people will now offer to
work. Thus, excess demand for teachers in the 1960s caused a rise in salaries, which, in turn,
brought about two types of supply adjustments: a movement up a given labor supply curve and, some-
what later, a rightward shift of the supply curve. As discussed in Chapter 7, the rightward shift in the
supply curve will continue until teachers’ salaries fall back to a level just competitive with other
occupations.
FIGURE 1.8
Demand and Supply : see “ie eve
in the Market for (in faves tea
Teachers
350 Newly qualified ,.

320 teachers oA
290
260
230
200
170
140 Openings for
teachers
110
80
1960 1964 1968 1972 1976 1980 1984
(b) Relative Earnings Ratio

Percent
110
105
100
Relative
earnings ratio
90
1960 1964 1968 1972 1976 1980 1984
sources: The Condition of Education, 1986 ed., Table 1.18, 1.19; National Center for Education Statistics, Bureau of
the Census, Money Income of Households, Families, and Persons in the United States, 1984, Series P-60, no. 151,
Table 28.
ZZ Chapter 1 The Labor Market
prospects of a buoyant job market several years earlier, graduated in record num-
bers with teaching certificates, only to find that the market had gone flat.
Given the excess supply of teachers in the 1970s, the model of demand and
supply predicts that two things should have happened to restore equilibrium.
First, relative wages for teachers should have fallen, and second, this should
have caused a decline in the number of people choosing to be teachers. Both of
these events clearly took place in the teachers’ market. From 1970 to 1980, the
earnings ratio of teachers fell precipitously, from 102 percent to 91 percent.
Likewise, in the space of 8 years (from 1973 to 1980) the number of college
graduates with majors in education declined nearly in half (from 19.5 percent
to 11.6 percent).
Graph (a) shows that the buyer's market for teachers probably ended in 1981.
Since then, demand for teachers has turned up, primarily because of an increase
in the school-age population, while the growth in the supply of newly qualified
teachers remained relatively unchanged from 1980 to 1984. The result? A grow-
ing shortage of teachers and a noticeable increase in teachers’ salaries. Can an
increase in the number of education majors be far behind?
Institutional Forces
. . . ll
institutional forces. ecu Porees represent the influence of various or-
Institutions affect labor outcomes in two different ways. First, they fragment
or “balkanize” the labor market into a number of segmented, loosely connected
submarkets. As described in a classic article by Clark Kerr, institutions introduce
structure, artificial boundaries, and rigidities into the labor market.’ In the clas-
sical labor market of Adam Smith, wages change automatically in response to
changes in demand and supply, and workers in turn flow from the occupation or
area of excess supply to those experiencing shortages. The boundaries of this
“natural” market are defined only by the limits placed on the sphere of competi-
tion by the flow of information, the transferability of skills, and the willingness
to travel. ‘nesleutiantaahow2\eieainisodige Lule.aie caulatisn’ that furrhes
define the dimensions of the labor market. These rules and regulations may be
either formal (written) or informal (customary). In either case, they effectively
delineate who can compete for particular jobs, who is most preferred, and which
business firms can compete in the bidding. These rules are established in the
form of corporate personnel policies, union contracts, and government legisla-
tion, among others.
Perhaps the best example to illustrate how institutions can segment the labor
market is the concept of the internal labor marke
policy of pre ing people
*Clark Kerr, “The Balkanization of Labor Markets,” Labor Markets and Wage Determination
(Berkeley: University of California Press, 1977): 21-37.
from the outside. Such a seniority or in-house bidding system creates an internal
labor market in that the effective sphere of competition is limited to individuals
already employed in the firm. Thus, when workers who have been with the com-
pany for some time retire or leave to work for another employer, vacancies are
created on the firm’s internal “job ladder.” Rather than hiring workers from the
outside to fill these jobs, the firm limits the competition to workers already em-
ployed with the company, since they are the ones most likely to have the re-
quired skills and knowledge. As these workers inove up the job ladder, vacancies
then appear at the bottom—the *‘port of entry.” It is at the port of entry where
inscompetition with each other to win the

jobs at the port of entry; once having done so, they can then move up the com-
pany job ladder.
This balkanization of the labor market is a consequence not only of corporate
personnel policies but also of union and government policies. Union contracts
often contain detailed seniority provisions that specify which individual or group
of individuals have first bidding rights at a new job vacancy. Similarly, the opera-
tion of union hiring halls in such trades as construction and longshoring restricts
the sphere of competition for work largely to union members. Finally, govern-
ment legislation in the form of protective labor laws that limit the ability of
teenagers or women to compete for certain types of work also establishes barriers
to entry that segment the labor market.
By delineating and narrowing the effective area over which the forces of de-
mand and supply can interact, smasnusmmelcaiicnlceilesGiatiallcildicig?
determinants
of wages and employment. This does not mean, however, that the
operation of the labor market is necessarily any less efficient. Internal labor mar-
kets, seniority provisions, and other such rules often have a positive impact on
the operation of the labor market by giving firms a method to train workers in
the idiosyncratic tasks of the job and as a means to promote on-the-job experi-
ence and continuity of employment. '°
The second way in which institutions affect labor market outcomes is through
Uisuinsepensientieliechanauaceels s In the classical market of many buyers
and sellers, no one employer or worker can independently affect the going mar-
ket wage; individual firms and workers are what economists call “price takers.”
The same is not true, however, in real-world labor markets populated by large
corporations, powerful unions, and firms regulated by the government.
The best example of how institutions can independently affect wage rates is
the labor unions In a competitive market, an individual worker has little or no
power to raise the wage above the going market rate, since the firm can easily let
one worker go and hire someone else. But what if the workers band together to
form a union? The power of a union to raise wages comes from its threat of a
strike. While the firm may be able to replace individual workers, if all workers
See O. E. Williamson, M. L. Wachter, and J. E. Harris, “Understanding the Employment Re-

lation: The Analysis of Idiosyncratic Exchange,” Bell Journal of Economics (Spring 1975): 250-278.
walk off the job together the company faces a potentially long shutdown and loss
of profits. The strike threat, then, gives trade unions an institutional influence
on rates of pay (and other forms of compensation) that originates outside of the
market forces of supply and demand. Market forces do impinge on the process of
union wage determination, however, since if a union pushes up wages too far it
risks causing substantial layoffs among its members. Empirical evidence suggests
that unions are able to raise the wages of their members above the going market
rate quite substantially, often on the order of 20 to 30 percent."
Large business firms and government also have an institutional influence on
wages. In competitive theory, any firm with labor costs higher than its com-
petitors’ will suffer the ultimate penalty of bankruptcy. In today’s economy, how-
ever, some firms operate in oligopolistic or regulated markets where the forces of
competition are attenuated, allowing some discretion to such companies in set-
ting their wage policies. Government also has a wide influence on wages that
transcend market forces. Examples are minimum wage floors, prevailing wage
standards in government construction, and legislated wage increases for govern-
ment employees.
Sociological Forces The third determinant of labor market outcomes is sociological forces, which
represent thei S
and the allocationof labor. Important — rane a
culture,
discrimination,
and custom.
Analogously with institutional forces, sociological forces influence the labor
market process in two distinct ways—first, through their influence on who can
and second, through their influence on the
determination
of wage rates in the market."
With respect to the first, sociological considerations such as family back-
ground, parents’ occupations, and socioeconomic class may have an important
effect on an individual’s range of choice and mobility in the labor market. This is
clearly stated, for example, in the following passage written over 100 years ago
by J. E. Cairnes in describing his Sy
No doubt various ranks and classes fade into each other by imperceptible gradations, and
individuals from all classes are constantly passing up or down; but while this is so, it is never-
theless true that the average workman, from whatever rank he be taken, finds his power of
competition limited for practical purposes to acertain range of occupations, so that, however
high the rates of remuneration in those which lie beyond may rise, he is excluded from sharing
them. We are thus compelled to recognize the existence of noncompeting industrial groups as a
feature of our social economy. '
"Richard B. Freeman and James Medoff, What Do Unions Do? (New York: Basic Books, 1984),
Table 3-1.
"This distinction is made by Henry Phelps Brown, who argues that sociological factors affect
labor market outcomes “before the market” and “within the market.” See The Inequality of Pay
(Berkeley: University of California Press, 1977), particularly Chapters 1 and 10. On the role of
sociological factors, also see Ivar Berg, ed., Sociological Perspectives on Labor Markets (New York:
Academic Press, 1981).
"J. E. Cairnes, Political Economy (New York: Harper, 1974): 67-68.
Chapter 1 The Labor Market, 25
The labor market described by Cairnes is balkanized much like the one dis-
cussed by Kerr, but for quite different reasons. Here the stratification between
labor market groups stems from one’s class and background as well as from spe-
cific labor market ski
To better appreciate the extent to which family background and socioeconomic
class noncompeting
create groups in the labor force, consider a current day ex-
ample. Presently fierce competition exists for admission to medical school be-
cause of the high income and status that it leads to. Do family background and
social class have any bearing on the individuals who are ultimately accepted? To
be more specific, is the fact that an applicant is the son of a construction worker
or daughter of a bank executive likely to have any independent influence on who
eventually becomes a doctor and who does not? Some economists would argue
that family background has relatively little impact, pointing to numerous ex-
amples of people who have risen from very humble beginnings to positions of
great wealth (and vice versa). Others, however, would argue that family back-
ground is quite important, admitting that while there is some movement be-
tween broad socioeconomic groups, individuals
as a rule do not move far from
their
occupational
parents’ and social position.
This question is obviously quite complex and subject to considerable debate. '*
It is also quite important because one’s position on this issue is likely to signif-
cantly influence one’s view of the social merits of relying on market forces as the
distributor of economic rewards in society. The greater the opportunity of indi-
viduals (and their sons and daughters) to move up and down the occupational
ladder as a consequence of their own efforts and decisions, the greater the argu-
ment for relying on the market. If, however, one’s station in life is largely deter-
mined by the accidents of birth and inheritance, then the market merely passes
on the status quo from one generation to another."
The second way in which sociological factors affect labor market outcomes is
through their influence on the supply of labor and the determination of wage
rates and earnings in the labor market. In some cases sociological factors influ-
ence wages independently of the market forces of supply and demand; in other
cases these sociological factors work through market forces as they influence the
supply of labor to the economy or to specific occupations and industries. With
respect to the latter, an individual’s decision about the amount of work versus
leisure or choice of occupations, for example, involves not only a consideration
of market opportunities such as wage rates and employment prospects, but also
the individual’s psychological feelings about which choice is valued more. Econ-
omists have always recognized this psychological aspect of choice and include it
in economic theory under the rubric of “preferences” or “tastes.” Thus, whether
a person chooses to become a teacher or a lawyer, for example, depends not only
on the wage rate that can be earned in each occupation, but also on the individ-
ual’s preference or taste for each line of work.
'4See, for example, Christopher Jencks, Who Gets Ahead? (New York: Basic Books, 1979).
5 As one piece of evidence on this, a survey of the 400 richest Americans (Forbes, October 1,
1984), fourid that 146 had inherited their wealth while 254 had largely earned it themselves.
In order to focus on the role of economic factors such as wage rates, income,
and prices as determinants of labor market outcomes, and because the determi-
nants of tastes and preferences have generally been thought to lie outside the
domain of economics, most economists assign a passive role to these variables,
assuming that they are either given or, alternatively, that they vary among indi-
While such an assumption is both useful
and empirically correct in certain labor market contexts, in other contexts the
tastes and preferences of individuals not only vary systematically over time and
by age, race, culture, and work group, but also influence their choices regarding
education, occupational attainment, and labor force participation, for example.
An illustration of this influence is the dramatic increase over the last two de-
cades in the proportion of working women in the economy. Research finds that
part of the rise in female labor force participation can be explained by economic
factors such as the -
mmm
eh The influx of millions of women
married women working outside the home.'’®
into the job market has caused, in turn, a large expansion in labor supply to
those occupations that traditionally employ a large number of women (such as
secretary and school teacher), depressing
wage rates for those jobs below what
they would have otherwise been.
cess. Culture inculcates in people a set of values or beliefs common to the larger
social group. These cultural values may have an important effect on the earnings
and economic success of different ethnic groups in the labor market as they
shape the attitudes of individuals concerning the work ethic, entrepreneurship
and risk taking, occupational choice, the importance of education, and the
desirability of getting ahead in a materialistic society. Are cultural differences
important in explaining wage and income differences among people? In several
influential books, economist Thomas Sowell argues that they are.'” In terms of
average family income, the two ethnic groups that have experienced the greatest”
eae ger oeregr O The great
majority of people of both groups came to America penniless and often suffered
intense discrimination here. Despite these roadblocks, both groups worked their
way up the economic ladder, passing by other ethnic groups such as the Irish and
the Italians. Why have other ethnic groups not succeeded economically to the
same extent? One reason is that some groups came to America later than others
and are several generations behind in the economic race.'* Sowell argues, how-
ney that some groups are also penalized
by cultural attitudes that conflict with
'°See Frank L. Mott, The Employment Revolution (Cambridge, Mass.: MIT Press, 1982).
See Thomas Sowell, Race and Economics (New York: David McKay Co., 1975); and Ethnic
America: A History (New York: Basic Books, 1981).
'*Dating the beginning point for when black Americans started the economic race is more diffi-
cult than for most other groups. Was it the early 1800s, when most blacks were brought to America;
1865, when slavery was abolished; or the 1950s when segregation was outlawed?
are perhaps the best example of this conflict. Despite being the first group to
settle in America, they still have the lowest average family income of all ethnic
groups.
Sociological factors also influence wages in ways that are independent of sup-
ply and demand conditions in the labor market. One example isdiscrimination.
Discrimination occurs whenever one person is treated preferentially over an-
other even though both individuals are equal except for some characteristic such
as sex, race, religion, or nationality. While each of us has individual likes and
dislikes, discrimination typically involves a common taste or preference on the
part of one social group against another, making discrimination a sociological
phenomenon. Doés discrimination play a role in the wage determination pro-
cess? The answer is surely yes. One issue this book examines in considerable de-
tail, for example, is the causes of the earnings gap between men and women. In
1986, women who were year-round full-time workers earned only 64 percent as
much as their male counterparts. The existence of an earnings differential be-
tween two groups does not, of course, prove that discrimination is the cause.
Research shows that half or more of the male/female earnings gap isdue tolegiti-
Taaleatnapielaisiateel factory such as length of work experience, college major,
and number of career interruptions for family reasons.'’ There is, however, in-
controvertible evidence that f-
fe with respect to pay and
promotions.”° To the extent that discrimination occurs, therefore, wages and
other market outcomes are determined by factors that are separate from the mar-
ket forces of supply and demand. This is not to say that market forces are irrele-
vant, however, for the employers that desire to discriminate are constrained in
the exercise of their prejudice by the need to hire the most qualified, productive
employees possible for the firm to make a profit.
Custom is a second example of an independent sociological influence on the
labor market. Custom is a practice or belief that has gained acceptance and le-
gitimacy simply because it has been followed for a long time. Some economists
argue that custom has a strong influence on wages, particularly relative wage dif-
Such pay ences,for instance those
ferences between groups of workers.*+ differ
between fire fighters and police officers or between locomotive engineers and
conductors, are regarded by the workers involved to be as much a matter of so-
cial standing and equity as economics, and often become a major criteria in the
wage determination process.
'°See Mary Corcoran and Greg J. Duncan, “Work History, Labor Force Attachment, and Earn-
ings Differences between the Races and Sexes,” Journal of Human Resources 14, no. 1 (Winter 1979):
3=20)
20See Paul Osterman, “Sex Discrimination in Professional Employment: A Case Study,” Industrial
and Labor Relations Review 32, no. 4 (July 1979): 451—464; and Maryellen Kelly, “Discrimination in
Seniority Systems: A Case Study,” Industrial and Labor Relations Review 36, no. 1 (October 1982):
40-55.
21Michael J. Piore, “Fragments of a Sociological Theory of Wages,” in Michael J. Piore, ed.,
Unemployment and Inflation: Institutionalist and Structuralist Views (White Plains, N.Y.: M.E. Sharpe,
1979), 134-143; and David Marsden, The End of Economic Man? Custom and Competition in Labor
Markets (New York: St. Martin’s Press, 1986).
THE EVOLUTION OF Most economists would readily agree that each of the market, institutional, and
LABOR MARKET sociological forces has some role in shaping labor market outcomes. There is,
THEORY nevertheless, considerable debate and disagreement over the relative impor-
tance of each force and how these forces work to bring about specific labor mar-
ket outcomes. Out of this debate have grown two distinct schools of thought or
paradigms in labor economics that have vied back and forth over the years for
dominance in the field.”? To provide some historical perspective on the develop-
ment of labor economics, as well as a heightened appreciation for why labor
economists disagree on fundamental issues of theory and policy, each school of
thought is briefly examined.”
The Neoclassical The neoclassical school is by all accounts the dominant paradigm in labor
School economics today. The word neoclassical means derived from or an extension of
the classical
economics of Adam Smith, David Ricardo, and other early 19th-
century economists. The founder of the neoclassical school in economics was
the brilliant English economist Alfred Marshall, who taught at Cambridge Uni-
versity from 1885 to 1908. Marshall was not himself a labor economist, but his
writings on microeconomic theory and its applications to labor-related issues set
the foundation upon which later economists built the neoclassical theory found
in today’s textbooks and academic journals.** After Marshall, neoclassical labor
economics received its next big boost in the late 1920s and 1930s from two econo-
mists, Nobel laureate John R. Hicks of Oxford University and Paul Douglas of
the University of Chicago. Hicks developed a number of crucial theoretical con-
cepts relating to labor demand and labor supply; Douglas’s major contribution
was to show how these theoretical concepts (such as the supply curve of labor)
could be estimated from available statistical data.”
Beginning in the 1930s and extending into the 1950s, the neoclassical school
fell out of favor in the United States, largely because the events of the Depres-
sion seemed to discredit it. From about the mid-1950s, however, the neo-
classical school staged a remarkable comeback, primarily as a result of the writ-
ing and research of a highly influential group of economists associated with the
University of Chicago. These economists included Nobel laureates Milton
Friedman, George Stigler, and Theodore Schultz, as well as H. Gregg Lewis,
Gary Becker, Jacob Mincer, Melvin Reder, and Albert Rees.
* A third school of thought includes Marxist or “radical” economists. While this paradigm has
had little influence on labor economics in the United States, it is more influential in European coun-
tries and Japan. For a Marxist view of the labor market, see David Gordon, Theories of Poverty and
Underemployment (Lexington, Mass.: Lexington Books, 1972); and Harry Braverman, Labor and
Monopoly Capital: The Degradation of Work in the Twentieth Century (New York: Monthly Review
Press, 1974).
A detailed history of thought in labor economics is provided in Paul McNulty, The Origin and
Dev velopment of Labor Economics (Cambridge, Mass.: MIT Press, 1980).
“Every economics student owes Marshall a debt of gratitude—he rigorously developed and popu
larized the familiar diagram of demand and supply (sometimes known as the ‘Marshallian cross”).
“See John R. Hicks, The Theory of Wages (New York: Macmillan, 1932); and Paul H. Douglas,
The Theory of Wages (New York: Macmillan, 1934), Chapters 11 and 12.
The neoclassical school of labor economics is distinctive in terms of both its

theoretical framework and its methodological approach to research. With re-
spect to theory, the neoclassical school gives primary
attention to the operation
To highlight the role of market forces, factors such as the nature of legal and
business institutions, sociological considerations of culture and class, the distri-
bution of property and wealth, and the pattern of tastes and preferences are gen-
erally abstracted from or treated as a given in the theory. This is not because
neoclassical economists necessarily view these factors as unimportant, but be-
cause these factors lie outside of what is perceived as the labor economist’s area
of expertise.
The neoclassical theory of the labor market has two important parts that
heavily influence the conclusions and predictions derived from it. The first con-
cerns the nature of human behavior. Neoclassical theory makes several simple
but very important assumptions about human behavior, assumptions that col-
lectively give rise to what is called the model of economic man.” One key as-
sumption concerns human motivation. The model of economic man holds that
or optimal outcome, given the constraints they face. A second key assumption is
that human beings have the cognitive ability to exercise rational choice. This
implies that the human brain is powerful enough to calculate the value of alter-
native outcomes so that the optimal outcome can be chosen. Finally, a third
assumption is that human beings are individualists in that behavior and prefer-
ences are largely independent of what others outside of the family think or do.
The second important part of neoclassical theory concerns the nature and
operation of markets. While neoclassical economists recognize that the labor
market has certain unique features, they do not perceive these differences to be
sO great as to preclude analyzing the labor market with the same theoretical
model used to study other product and factor markets.’’ Furthermore, it is usu-
ally assumed that the labor market is highly competitive, having a large number
of buyers and sellers and allowing relatively easy entry into and exit from the
market. The importance of these assumptions is that they ensure that demand
hat impersonal market forces are the
major determinant of wages and the distribution of labor, and that market out-
allocation
comes result in an efficient of resources.
In addition to its theoretical framework, the methodology of neoclassical eco-
nomics is also noteworthy. The word methodology means the methods and pro-
cesses by which economists seek to explain real-world behavior. The first feature
of methodology that distinguishes neoclassical economics is the reliance on
26The model of human behavior in neoclassical theory is described in detail in Gary S. Becker,
The Economic Approach to Human Behavior (Chicago: University of Chicago Press, 1976): 3-14. A
critique of the model of economic man is given in Harvey Leibenstein, Beyond Economic Man (Cam-
bridge, Mass.: Harvard University Press, 1980).
27See Milton Friedman, “The Supply of Factors of Production,” in Price Theory: A Provisional Text
(Chicago: Aldine Publishing, 1962): 199-225.
to derive theoretical predictions and testable hypotheses.

Deductive reasoning is logic that progresses from the general to the specific; its
opposite is inductive reasoning, which progresses from the specific to the general.
Either method can be employed to predict what will happen to a firm’s desired
level of employment if it is forced to pay a higher wage. One method is to study a
number of individual firms in which the wage has gone up and, based on the
change in employment in each, formulate a general law or prediction. This is an
example of inductive reasoning. A second method is to start with several general
assumptions about the goal of business firms and the determinants of revenues
and costs and from this deduce the likely employment effect of a wage increase.
This is a case of deductive reasoning. Neoclassical economists make heavy use of
the deductive method, believing it is both more general and more fruitful than
an inductive, case-study approach. The benefit of the deductive method is that
given only a few rather general assumptions, it is possible to derive a very ele-
gant and sophisticated model of labor market behavior. The drawback is that the
model and its predictions may be seriously in error if the assumptions do not
accord with real-world labor markets.
The second noteworthy aspect of methodology is the reliance
heavy on mar-
ginal decision rules in neoclassical theory. Given that the goal of individuals is to
maximize some objective such as profits or utility, the “golden rule” is always the
same: whatever the activity, continue to do more of it as long as the marginal
increase in benefit exceeds the marginal increase in cost; when the two become
equal, the optimal amount has been reached.
The Institutional The main rival to the neoclassical theory in labor economics is the institutional
School school. Institutional economics emerged in the United States at the turn of the
century, mainly as a reaction against the neoclassical economics of Marshall and
his American disciples. It is generally considered that institutional economics had
three founders: Thorstein Veblen, John R. Commons, and Wesley C. Mitchell.”
The history of institutionalism in labor economics is a complicated one and
involves three separate phases. The first phase dates from the 1920s to the early
1940s and is associated with Commons and his students at the University of Wis-
consin. This branch of institutionalism largely eschewed formal models or quan-
titative analysis of the labor market, preferring instead to conduct painstaking
and investigative studies of current-day
labor problems. The second phase is associated with the “neoinstitutionalists” of
the 1940s and 1950s.”” The neoinstitutionalists dominated labor economics over
*®See Joseph Dorfman, Institutional Economics: Veblen, Commons, and Mitchell Reconsidered (Berke-
ley: University of California Press, 1963).
See Glen G. Cain, “The Challenge of Segmented Labor Market Theories to Orthodox Theory:
A Survey,” Journal of Economic Literature 14, no. 4 (December 1976): 1215-1257. Not all econo-
mists agree that this second phase was actually a continuation of the institutional school. For al-
ternative points of view on this subject, see Clark Kerr, “The Neoclassical Revisionists in Labor
Economics (1940—1960)—R.I.P.,” and Bruce E. Kaufman, “The Postwar View of Labor Markets
and Wage Determination,” in Bruce E. Kaufman, ed., How Labor Markets Work: Reflections on
Theory and Practice by John Dunlop, Clark Kerr, Richard Lester, and Lloyd Reynolds (Lexington, Mass. :
Lexington Books, 1988): 1-46, 145-203.
Chapter 1 The Labor Market a1
this 20-year period and included many names that remain famous in the field
today, such as John Dunlop, Lloyd Reynolds, Clark Kerr, Arthur Ross, and
Richard Lester. The neoinstitutionalists were different from Commons and the
“Wisconsin school” in that they placed more emph on the
asis
study and theory
of how labor markets actually work. While this moved them closer to the posi-
tion of the neoclassical economists, the case studies of labor markets conducted
by the neoinstitutionalists convinced them that the neoclassical theory was
quite unrealistic as a description of how wages and employment are actually
determined. With the revival of neoclassical theory by the Chicago school in
the 1960s, the influence of the neoinstitutionalists waned in labor economics,
although their présence remained considerably stronger in the related field of
industrial relations. In the late 1960s, a third phase of institutionalism emerged
which continues to exist today. This group is known as “segmented
labor mar-
titiebatcialiesadtluseomsconamistsesuchassablichael Piore, Peter
Doeringer, and Barry Bluestone.*° Like the neoinstitutionalists before them, the
SLM economists emphasize how the policies of unions and corporations, as well
as the effect of social forces such as discrimination and class, segment the laboie
ee ASGHECMESEo-
classical theory.
Although the institutional school of labor economics includes a number of
diverse people and strands of thought, there are certain common features in
terms of both theory and methodology that unite institutionally oriented econo-
mists. One issue of theory concerns the appropriate model of human behavior.
I ic man. Many institu-
tional economists, for example, deny that people maximize in pursuit of their
economic objectives, believing instead in an alternative theory of motivation
known as satisficing.” Satisficing means that an individual pursues a goal only
. Rather than
strive for the optimal outcome, people who satisfice settle for the outcome that
they regard as good enough, even if it is not the best they could do. Insti-
tutionalists also believe that people’s choices are less rational and consistent
than the model of economic man presumes. The major problem, as they see it,
is that the human brain is too limited to assimilate all the data and to make all
the complex calculations that are required to arrive at optimal and consistent
choices. Nobel laureate Herbert Simon calls this feature of human decision
making bounded rationality.» Finally, institutionalists also place great stress on
See, for example, Peter B. Doeringer and MichaelJ.Piore, Internal Labor Markets and Manpower
Analysis (Lexington, Mass.: Heath Lexington Books, 1971); and Peter Doeringer, “Internal Labor
Markets and Noncompeting Groups,” American Economic Review 76 (May 1986): 48-52.
The assumption that people satisfice rather than maximize can be found in any of the neo-
institutional labor texts of the 1950s. See, for example, Joseph Shister, Economics of the Labor
Market, 2d ed. (Chicago: J. B. Lippincott Co., 1956): 362-375.
32See, for example, Herbert Simon, “Rationality as Process and as Product of Thought,” Ameri-
can Economic Review 68, no. 2 (June 1978): 1—16. Perhaps the best-known study in labor economics
that makes use of the concept of bounded rationality (although the term was not known then) is
Richard Lester, “Shortcomings of Marginal Analysis for Wage-Employment Problems,” American
Economic Review 36, no. 1 (March 1946): 63-82.
two aspects of preferences. The first is that people’s preferences or tastes are in-
terdependent; 3
. The second aspect of preferences emphasized
by institutionalists is that preferences are not exogenous or a “given,” but are
endogenous, meaning that they are shaped by and change with the economic
Institutionalists also hold a different view concerning the structure and opera-
tion of labor markets. While neoclassical economists emphasize the competitive
nature of labor markets and the primacy of market forces in determining wages
and employment, institutional economists tend to emphasize just the opposite.”
As they see it, institutional forces, such as internal labor markets and unions,
and sociological factors, such as class and discrimination, segment or stratify the
’
labor market into noncompeting groups, preventing the free flow oflabor from _
one sector to another. Institutionalists also do not believe that market forces
work as efficiently or strongly as neoclassical economists maintain. Partly institu-
tionalists place far more emphasis on the unique features of the labor market and
how these features impede or diminish the role of supply and demand. Another
factor is that institutionalists give more weight to the importance of market im-_
perfections such as rigid wages, persistent unemployment, barriers to labor mo-
bility, and poor information. Finally, institutionalists give a larger role to the
independent influence of unions, corporations, discrimination, and custom in
the process of wage determination than do neoclassical economists.
Institutional labor economists also approach research methodology in a differ-
ent way than their neoclassical counterparts. Institutionalists
prefer acase-study,
how the labor market “really” works: Institutionalists also claim that the mar-
ginal decision rules of neoclassical theory often unrealistically describe how de-
cisions are actually made. In their view, decisions are more often made on the
basis of average values rather than marginal values. One reason for this, accord-
ing to institutionalists, is that organizations such as unions typically make policy
decisions on the basis of a majority vote of the membership instead of some type
of marginal calculation of benefit and cost.** A second reason is that the stan-
dard accounting data available to firms provides ready information on average
cost, average output per worker, and so on, but very little information on the
marginal value of such variables.
*See, for example, James Annable, Jr., The Price of Industrial Labor (Lexington, Mass.: Lex-
ington Books, 1984). The best examples of the neoinstitutional view of the labor market are Richard
Lester, Labor and Industrial Relations: A General Analysis (New York: MacMillan, 1951): 37—73; and
Lloyd Reynolds, The Structure of Labor Markets (New York: Harper and Brothers): 207-256.
‘As shown in Chapter 11, under a system of majority vote the wage policy of a union will reflect
the preferences of the median or “average” worker. In this case, the union's optimal wage is deter-
mined through a political process rather than an economic process involving a calculation of the
marginal gain to the union from higher wages versus the marginal cost from reduced employment, as
is presumed in neoclassical models.
THE IMPORTANCE Given the different theories and schools of thought in labor economics, how
OF THEORY AND does an individual as a policymaker, economist, or just as an interested student
HYPOTHESIS go about deciding which explanation of labor market behavior best fits the facts?
TESTING As an example, consider the decline that has taken place in the 20th century in
average hours of work per week. What were the factors responsible for this
trend? Was it the result of government legislation requiring overtime pay after 40
hours of work per week, or does it reflect the bargaining success of unions in
lowering hours of work? Was it reflective of changing shift and production pat-
terns of employers or was it a consequence of the growing demand by workers for
more hours of leisure time?
Here are four different explanations for the same event. Which one is correct?
Are they all correct? To decide this question involves the use of both theory and
empirical hypothesis testing.»
The first requirement in explaining a particular labor market outcome is to
have a theory or model that shows the logical cause and effect relationships be-
tween some original Event A and the resulting outcome in the labor market,
Event B. Without such a theory we have no way of knowing why Event A caused
Event B nor could we predict what would be the effect if Event C were to take
place instead of Event A. Such a theory can be verbal, graphical, or mathemati-
cal; the major requirements are that it be logically consistent and generate pre-
dictions or conclusions that can be at least potentially refuted from examination
of actual labor market events or experience. An additional requirement of most
theory is that it be an abstraction from real life so that the main cause and effect
relationships are not obscured by the complexity of the labor market.
Having constructed one or several theories to explain a particular event, how
does a person then choose among them? The next step involves the role of
empirical research and hypothesis testing. To be useful, a theory has to yield a
prediction or hypothesis in the form of “if A, then B.” The role of empirical
research is to test this hypothesis by examining actual empirical data gathered in
the labor market and seeing whether the predictions of the theory reasonably fit
the facts. If the predictions do not fit the facts, the theory can be rejected as
incorrect; if they do coincide, then we can infer support for the theory, but not
its final proof.
This two-step process of theory and empirical research is crucial to under-
standing how the labor market works and why events occur as they do. In the
terminology of economics, this process involves positive economics—the study
of how and why things are the way they are. Positive economics involves ques-
tions of fact—why is the unemployment rate for blacks twice that of whites, or
why did productivity growth decline in the 1970s? The answers to these ques-
tions are varied and not always in agreement, but through scientific investiga-
tion and examination of empirical data it should be possible to reject competing
hypotheses and accept the one that best fits reality. In practice this process is not
35For a more detailed discussion of methodological issues in economics see Mark Blaug, The Meth-
odology of Economics (Cambridge, England: Cambridge University Press, 1980).
IN « THE « NEWS
Different Views on the Merits
of Raising the Minimum Wage
Senator Kennedy (D—Mass.) and Congressman one group of workers was left behind: those paid
Hawkins (D—Calif.) have introduced bills in the the $3.35 per hour minimum wage. The failure to
current session of Congress to raise the federal raise the minimum wage has contributed greatly to
minimum wage from its 1987 level of $3.35 per the high incidence of poverty among workers paid
hour to $3.85 in 1988, to $4.25 in 1989, and then the minimum. Nearly 3 million minimum wage
to $4.65 in 1990. Is this legislation a good idea? earners, or 35 percent of all those at the mini-
Given below are two very different perspectives on mum, were in households below the poverty level
this issue. The first passage, excerpted from a re- last year.
port issued by the AFL—CIO, argues in favor of Opponents of raising the minimum wage claim
raising the minimum wage; the second, published that a higher minimum would increase unemploy-
by the U.S. Chamber of Commerce, argues against ment. Such an argument ignores both present and
the legislation. Note how the different policy con- past experience. Today, unemployment is at its
clusions reached by each side reflect fundamen- highest level in any post-war recovery period, and
tally different assumptions about how labor mar- the minimum wage has not changed in 6 years.
kets work, and how each side attempts to buttress Unemployment is related to many factors, but not
its position by showing that real world events the minimum wage.
match the predictions of its model of the labor Also contradicted by real world events is the
market. charge made by opponents of the minimum wage
that teens lose jobs when the wage floor is raised,
The AFL—CIO. During the last six years,
and, conversely, that a decline in the minimum
while food, housing, and clothing prices rose and
would put young people to work. If this were true,
while most workers received wage adjustments,
as clear-cut as it sounds because of the complexity of human behavior and the

limitations and shortcomings in the available data.
It is a short step from questions of fact—positive economics—to questions of
how things ought to be—normative economics. Normative economics involves
value judgments; for example, women should earn as much as men, or the mini-
mum wage should be abolished. While facts alone cannot answer these or other
normative debates encountered in labor economics, theory and empirical evi-
dence are usually (but not always) a crucial input into shaping decisions on such
issues.
THE IMPORTANCE Since the New Deal of the 1930s, government regulation and intervention in
OF THEORY FOR the labor market have expanded tremendously. Before that time, the minimum
PUBLIC POLICY wage, Social Security, overtime pay requirements, health and safety regulations,
and affirmative action and equal opportunity laws were unknown. Particularly
:
rr
SS ee eee
then teen employment should have risen consider- to households with an employed adult and income
ably during the 1980s because of the sharp drop in above the poverty line. To those young, inex-
the minimum wage compared both to inflation perienced workers, minimum wage jobs are an ex-
and to average hourly earnings for all workers. In tension of schooling since those jobs offer them
reality, teen employment dropped 10 percent or the training and the work experience to move
three-fourths of a million from 1981 to 1986. on to higher paying jobs. The minimum wage
Many wages are not set in free and perfectly deprives the youth of America, not only of the
competitive markets. The lowest paid workers in Opportunity to get a job, but also of the needed ex-
society often suffer from a lack of bargaining power. perience and on-the-job training, while it does
They are easy targets for exploitation by business, nothing to alleviate poverty. Ever since the Fair
especially when there is a large pool of unem- Labor Standards Act (FLSA) of 1938 instituted a
ployed individuals seeking work, as has been par- federal minimum wage, numerous studies have
ticularly true in the United States during the years shown that increases in the minimum wage do, in
since the last rise in the minimum wage. Some fact, reduce employment below the level that
nonmarket institution or arrangement is often would prevail in the absence of such increases. In
needed, therefore, to prevent such exploitation. 1981, the Minimum Wage Study Commission
found that a 10 percent increase in the minimum
The U.S. Chamber of Commerce. The main
wage reduced teenage employment between 0.5
thrust of the argument for raising the minimum
and 1.5 percent.
wage is that doing so will raise earnings of low-
Usually, businesses will deal with the increase
income workers and reduce poverty. This reason-
in labor costs resulting from the higher mini-
ing, however, completely ignores the fact that
mum wage by either (1) passing the burden on
only one-fifth of all minimum wage workers are
to consumers in the form of higher prices and/or
classified as living below the poverty line. Most
minimum wage workers are teenagers who belong continued
with the election of President Reagan.in 1980, however, a growing debate has
taken place as to the proper role of government in the economy and, in particu-
lar, whether or not the economic and social welfare of the country would be
better promoted by dismantling or expanding some of the legislative and regula-
tory programs enacted over the previous 50 years.
The two sides to this debate start out with different normative assumptions
about what ought to be the goals and direction of society. The important point,
however, is that both sides attempt to buttress their public policy programs on
an appeal to the “facts,” that is, on a positive analysis of how the labor market
works and the effect of government intervention on it. The fact that each side
reaches a diametrically opposite conclusion concerning the desirability of a
minimum wage law or equal opportunity requirements, for example, stems from
their different assumptions as to the labor market process and in particular to the
relative roles of economic, institutional, and sociological forces in the labor
market (see the accompanying “In the News” section).
IN « THE « NEWS
curtailed services; (2) absorbing the added costs other economic distortions which reverberate
by reducing profit margins; or (3) passing these through the economy, thus reducing overall eco-
costs along to the workers themselves by reducing nomic efficiency.
the number of hours worked, reducing the number
of employees, freezing new hiring, or requiring
Sources: Rudy Oswald and Rick Krashevki, “Increase the Mini-
that workers earn their increased wages through mum Wage Now,” AFL-CIO Reviews the Isues, Report 13
added work effort. Thus, the first impact of a (Washington: AFL—CIO, July 1987); and Graciela Testa-Ortiz,
“Raising the Minimum Wage and Derailing the Great Ameri-
higher minimum wage is lower levels of employ-
can Jobs Machine,” Policy Working Papers, No. 3 (Washington:
ment and higher inflation. But this is just the be- U.S. Chamber of Commerce, April 1987). Both passages were
ginning. Increases in the minimum wage cause condensed from the original to promote readability.
‘Ee ee ee ee SS EE SS SS
As Ray Marshall, President Carter's secretary of labor, argued, every proposed

or actual piece of government labor legislation is based or justified on either an
explicit or implicit theory of the labor market.*° As a rule, economists and
policymakers who believe in the primacy of market forces and the competitive
model of demand and supply favor a minimal amount of government inter-
vention in the labor market.*” Others who give greater weight to the defects of
the market system, as well as to the role of institutional and sociological consid-
erations, favor a greater intervention by government in the labor market.** In
either case, the normative public policy proposals of what ought to be done are
based directly on a positive theory of how the labor market works.
Seen in this light, it is clear that the theory of the labor market process is
much more than just an interesting economic debate; the outcome of this debate
has profound implications for the direction of economic and social policy in the
country. The resolution of this debate (if there ever is one) will come from the
weight of empirical evidence for and against each point of view as citizens and
policymakers observe the labor market outcomes generated by the respective
programs of one side or the other. Thus, the interaction between theory and
empirical evidence guides not only economists in their search for “the” answer
to how the labor market works, it also provides the basic tool for policymakers
and the average voter to judge the relative merits of particular labor market pro-
grams and the broader issues of government and society.
*°F, Ray Marshall, “Implications of Labor Market Theory for Employment Policy,” in Gordon
Swanson and Jan Michaelson, eds., Manpower Research and Labor Economics (Beverly Hills, Calif.:
Sage, 1979): 18.
See, for example, Gary Becker, “Let’s Put Deregulation to Work in Labor,” Business Week
(July 14, 1986); and Dan C. Heldman, James T. Bennett, and Manuel H. Johnson, Deregulating
Labor Relations (Dallas: Fisher Institute, 1981).
“For an example of this point of view, see Sar Levitan, Peter Carlson, and Isaac Shapiro, Protect-
ing American Workers: An Assessment of Government Programs (Washington: Bureau of National
Affairs, 1986).
Chapter 1 ~The Labor Market
37
SUMMARY The science of economics is concerned with the allocation of resources in the
economy and the determination of prices and levels of production. For a capi-
talist economy, the primary theoretical construct used to understand these issues
is the market model of supply and demand. In labor economics, one particular
market, the labor market, is studied. Labor is a service that households supply to
business firms in order to earn an income and that business firms demand in
order to produce their product. It is this interaction of the demand and supply of
labor that determines wage rates, the level of employment, and the distribution
of income in the economy. Market forces of supply and demand are not the only
determinants of these market outcomes, however, institutional and sociological
forces also play roles. Institutional forces affect labor market outcomes through
the influence of organizations such as labor unions, large corporations, and gov-
ernment; sociological forces exert their influence through such factors as cul-
ture, class, discrimination, and custom. There is no one model or paradigm in
labor economics that weaves these market, institutional, and sociological forces
into a universally accepted theory of how the labor market works. Rather, there
are different schools of thought that give varying degrees of emphasis to the rela-
tive importance of these forces in the labor market and the degree to which the
market forces of supply and demand efficiently work to set wage rates and allo-
cate labor. To discriminate between these alternative theories, therefore, re-
quires a process of hypothesis testing and empirical validation of competing
theoretical predictions.
GLOSSARY
Derived demand— The demand for labor by the Internal labor market — The group of jobs in a
firm that is derived from the demand for its firm that are filled through internal promotion
product. rather than hiring from the external market.
Economic man—A theory of human behavior Labor demand curve—A curve that illustrates
that assumes persons maximize, exercise ra- the inverse relationship between the wage rate
tional choice, and are individualistic. and the quantity of labor demanded by business
External labor market— The labor market exter- firms, holding other factors constant.
nal to the firm where workers compete for entry Labor market — The area within which demand
level jobs. and supply determine the price for a particular
Factor market — A market where a factor of pro- type of labor.
duction such as land, labor, or capital is bought Labor market outcomes
— The observed events,
and sold. developments, or trends in the labor market.
Institutional forces — The influence that organi- Labor market process— The mechanics of the
zations such as unions, corporations, and gov- labor market—how it works, the nature of the
ernment have on the pricing and distribution of cause and effect relationships, and the major
labor. actors and institutions.
Institutional school — A school of thought in la- Labor supply curve— A curve that illustrates the
bor economics that emphasizes imperfections in positive relationship between the wage rate and
the market mechanism and the importance of the quantity of labor supplied by workers, hold-
institutional and sociological forces. ing other factors constant.
Market equilibrium — A state of balance in the among which there is little competition in the
labor market where demand and supply are labor market due to differences in family back-
equal. ground, class, or skills.
Market forces— The influence that the market Normative economics— The study of economics
forces of demand and supply have on the pricing based on value judgments of what ought to be.
and distribution of labor. Positive economics— The study of economics
Neoclassical school— A school of thought in la- based on objective facts or logic.
bor economics that bases its theory on the twin Sociological forces —The influence that social
assumptions of competitive markets and the groups and norms have on the pricing and dis-
model of economic man. tribution of labor.
Noncompeting groups— Socioeconomic groups
REVIEW QUESTIONS
1. Where do the other factor markets for capital ciently allocate labor in the economy? What role
and land (natural resources) belong in the circular does labor market information have in this pro-
flow model depicted in Figure 1.1? Can you draw cess? Do you think people move from one labor
them in? Are there any connections or feedbacks market to another in search of their highest ad-
between one factor market (such as labor) and an- vantage, as economic theory assumes? Why or
other (such as capital)? How would you represent why not?
them in Figure 1.1? 5. What is the difference between an external
2. If human beings could be bought and sold like and internal labor market? How are they con-
commodities such as wheat, how would it change nected? Can a firm decide on its own what wages
the operation of the labor market? Would the sup- to pay in the internal labor market, or do the
ply of labor become more or less sensitive to the market forces of supply and demand have some
nonpecuniary aspects of employment (such as influence?
working conditions)? Would market forces still be 6. In 1986, the average weekly earnings of truck
able automatically to police the social conditions drivers was $371 while average weekly earnings of
of labor? bank tellers was only $231. How might market,
3. Are there any differences between the labor institutional, and sociological forces, respectively,
market for day laborers or migrant workers and the be responsible for the higher earnings of truck
market for teachers or pilots? What are these dif- drivers?
ferences and how do they affect the pricing and 7. Describe the major differences between the
allocation of labor? neoclassical and institutional schools of labor’
4. Why is the role of labor mobility crucial if economics.
market forces of demand and supply are to effi-
ADDITIONAL READINGS
Ashenfelter, Orley, and Richard Layard, eds. Friedman, Milton, and Rose Friedman. Free to
Handbook of Labor Economics. Amsterdam: Choose. New York: Harcourt Brace Jovanovich,
North-Holland, 1986. Contains a number of ar- 1980. A classic and very readable discussion of the
ticles written by prominent economists that survey virtues of a free market system. Chapters 1 and 8
important areas of research in labor economics. are particularly pertinent for this course.
Chapter 1 =The Labor Market 39
Kalleberg, Arne, and Ivar Berg. Work and Indus- kets.” Industrial and Labor Relations Review 9,
try: Structures, Markets, and Processes. New no. 2 (January 1956): 183-199; and Lampman,
York: Plenum Press, 1987. Written by two soci- Robert J. “On Choice in Labor Markets: Com-
ologists, the book examines the evolution of institu- ment,” Industrial and Labor Relations Review 9,
tional structures in the labor market and their im- no. 4 (July 1956): 629-636. These two articles
pact on various outcomes such as earnings and provide a very interesting debate over the extent to
career mobility. which real-world labor markets actually operate in
Kerr, Clark. “Labor Markets: Their Character and accordance with the assumptions of the competitive
Consequences.” American Economic Review 40, market model.
no. 2 (May 1950): 278-291. Distinguishes among Thurow, Lester. Dangerous Currents. New York:
five different models of the labor market and the im- Random House, 1983, Chapters | and 7. Argues
plications of each for the determination of wages and that the auction model of the labor market is un-
employment. realistic; provides a general critique of neoclassical
Rottenberg, Simon. “On Choice in Labor Mar- theory.
SS ~=
im Sonn)
= cy . y
PTT 4 ath,
7 ‘i } itt sores ib oe
= treads iol iin

:
“eo = We
' eanyt ‘ speciemea

Cah
Ag be lah R
Hours of Work
is chapter and the next deal with the subject of labor supply. Labor supply
has several dimensions encompassing both the i ity of the
labor input made available for market work. This chapter begins the analysis of
labor supply by focusing on onesOheneduapiinaiimens nea nung sacks)he
first part of the chapter develops the basic neoclassical theory of labor supply,
known as the labor/leisure model. This is followed by an examination of em-
pirical evidence on the determinants of work hours. The theory and empirical
evidence are then used to analyze two public policy debates. These are the im-
pact on labor supply of the Reagan tax cut and welfare programs such as Aid to
Families with Dependent Children. The next section analyzes two more institu-
tionally oriented aspects of labor supply, the impact of employers’ preferences
, and whether blue-collar and white-
collar workers have different preferences for work versus leisure. The final part of
the chapter examines the causes of the long-term decline in average work hours
per week in the economy.
THE PATTERN OF Before the theory of labor supply is introduced, it is important to point out the
HOURS OF WORK real-world behavior that the theory is meant to explain. This is done in Figure
2.1. Graph (a) shows the distribution of people employed in July 1987 by their
hours of work per week. This is an eXample ofcross-sectional data— it shows a
. Graph (b), on the other hand, provides an illustration of time-series

data, since itshows the change over time in average hours worked per week for
all workers as a group.
Three features of the cross-sectional distribution of hours are noteworthy. The
first is the wide diversity among individuals in the hours worked per week. At
one extreme are the 3.7 percent of the employed who worked 14 hours or less
per week; at the other are the 7.7 percent who worked 60 hours or more. The
second feature is the eek. Al-
though a great many Americans work either relatively few or many hours, the
single greatest number in 1987 (42 percent) worked exactly 40. Finally, also
shown is the proportion of people who were s. To be
: 41
42 Chapter 2. Hours of Work
FIGURE 2.1
Distribution of (a) Cross-Sectional Pattern
Employed Persons by Percent
Part-time Full-time
Hours of Work, 1987;
Average Weekly
Hours of Work,
1900-1987
1-4 5-14 15-29 30-34 35-39 40 41-48 49-59 60+ Hours

per week _ a
(b) Time-Series Pattern

Hours
per week
55
50
45
40
35
1900 1910 1920 1930 1940 1950 1960 1970 1980 1987 Year
source: Bureau of Labor Statistics, Employment and Earnings 34, no. 8 (August 1987): 32; and G. Moore and
J. Hedges, “Trends in Labor and Leisure,” Monthly Labor Review 94, no. 2 (February 1971): 5.
classified as a full-time worker, a person must work 35 or more hours a week,

depicted in graph (a) by the broken vertical line. Despite the growing propor-
tion of part-timers in the work force, they continue to account for fewer than ,
one-fourth of all the persons employed in the economy.
Graph (b) shows the change in average hours worked per week in the econ-
omy from 1900 to 1987. At the turn of the century the average workweek was 53
hours; for many workers a 9 or 10 hour day was standard, with time off only on
Sunday. Over the course of the 20th century, however, the workweek declined
significantly until by 1987 the < urs for full-time
workers and . A second interesting feature of the secu-
lar decline in hours of work is the marked difference in the trend in hours of
work before and after World War II. Most of the decline in hours of work oc-
curred between 1900 and 1945; since then there has been some further decline,
but it has not been nearly as sharp as before.
Chapter 2 Hours of Work © 43
Z
In the terminology of Chapter 1, the cross-sectional and time-series pattern of

hours of work are the labor market outcomes we wish to explain. To do so, we
need a theory that can identify the major influences on people’s choice concern-
ing work versus leisure.
THE THEORY There is a wide diversity in the hours per week that people choose to work. Why
OF THE do some individuals work only part-time while others moonlight on two or three
LABOR/LEISURE separate jobs? Presumably both the part-time worker and the moonlighter have
CHOICE made a conscious choice to find jobs that provide the work schedule they want.
This section analyzes how this choice is made.
There are 168 hours in a week, and each individual decides how to allocate
those hours among various uses. A certain number of hours per week are not
really discretionary, however, since there are fixed biological needs for sleeping,
eating, and so on. Assuming, for simplicity, that these fixed needs require 68
hours per week (or a little less than 10 hours per day), 100 hours remain about
which choices can be made. This chapter assumes that there are tw ible
uses forthisdiscretionary time—work and leisure. Wns AGT a
paying job; leisure is the remainder of time used for all other activities. Leisure is
used here in a very broad sense to cover nonmarket activities as diverse as
watching a movie, going to school, or working in the home. This somewhat
artificial definition of leisure keeps the analysis at a relatively simple level; in the
next chapter, the theory is made more realistic by distinguishing among three
possible uses of time: market work, nonmarket work, and leisure.
Since the amount of time not working is by definition the amount of time
spent at leisure, it is possible to treat the demand for leisure and the supply of
a naSar EISEN Since economics has a well-developed the-
ory of demand, the choice between labor and leisure is analyzed in terms of the
demand for leisure; thesupply oflabor canthen befound bysubtracting hours of
According to the theory of consumer demand, the demand for a specific good
or service depends on a number of variables: in particular, the price of the good
or service, the amount of income that potential buyers have, and tastes or pref-
erences for the good or service. The theory predicts that the higher the price of a
good is, holding all other things constant, the lower the quantity demanded of it
will be. Similarly, this theory predicts that if income should increase, the de-
mand for the good will increase, assuming that it is what economists call a nor-
change simply because of a change innae preferences or tastes concerning it.

This theory of demand can be used to analyze an individual’s decision about
how much leisure time he or she desires. While it may seem strange to talk about
a “demand for leisure,” leisure time provides utility or satisfaction much like any
other good or service, and it also has a definite price attached to its use. Given
these parallels, the following sections analyze
Preferences and about

Indifference Curves
’
. They are inherently sub-
jective in nature and potentially influenced by a host of factors related to one’s
ethnicity, socioeconomic class, and occupation, as well as random personality
factors. Although preferences are subject to considerable variation among indi-
viduals, research shows that at a point in time people possess a clear notion of
how they rank the desirability of competing goods or services and how much
they would trade of one to get more of another.’
The theory developed here assumes that individuals have to rank and choose
between two goods: ¢iiaiaeanchincomencannedsdaatemnD Leisure is desired for
its own intrinsic qualities; income is desired for the goods and services it can buy.
Figure 2.2 graphically represents the process of ranking and choosing among al-
ternative combinations of leisure time and income.
In Figure 2.2 the horizontal axis measures hours of leisure per week ranging
from zero to the hypothetical maximum of 100 hours. The vertical axis measures
money income (Y) and ranges from zero to $1,000 per week. Point A is an arbi-
trary combination of leisure and income. At point A the individual has 40 hours
of leisure and $400 of income per week, yielding some particular level of utility.
aseneasnlfaenaa
aE
e=ninp
meee
istnet
ESET ret ene
i rari
i
The next step in the analysis is to derive what is known as an indifference
curve. The derivation of an indifference curve is based on an individual’s answer
to the following question: “What other combinations of leisure and income,be -
‘ticaphandineiniehaulileneimuiuuomnglagiiaabslaa’ (Or, put another
way, “Are there other combinations of leisure and income besides those at point
A that you would just be indifferent to?”) The answer typically will trace out a
curve much like the one labeled I, in Figure 2.2.’ On an indifference curve,
Te ac several important properties. First, as a matter of

logic, they have to have a negative slope—that is, they slope down and to the
right. The reason is that if we are to find some new combination of income and
leisure that has just the same utility level as point A, clearly the person must
receive more of one but at the same time lose some of the other. Such a new
combination might be point D on indifference curve I,, with an income of
$500, but leisure of only 30 hours.
The second important property of indifference curves is that they aré
€onvex
to
origin.
the At point A, how much income would this person have to be given
‘See Arnold A. Weinstein, “Transitivity of Preference: A Comparison among Age Groups,” Jour-
nal of Political Economy 76, no. 2 (March/April 1968): 307-311.
*See L. F. Dunn, “An Empirical Indifference Function for Income and Leisure,” Review of Eco-
nomics and Statistics 60, no. 4 (November 1978): 533—540.
Chapter 2. Hours of Work 45
Same Vera o& Okt :

FIGURE 2.2 A Set of JLT Curves Conve
Points A, B, and C
represent successively incon
higher levels of utility since per week (Y)
the person has more of
both leisure and income.
$1,000
For each level of utility, an
indifference curve may be
constructed that shows all
the other combinations of
income and leisure that the
person regards as equally $700
satisfying. Points A, D, and
E on indifference curve /,,
for example, all yield the
same level of utility. Each $500
indifference curve is
convex, giving rise to a $400
diminishing marginal rate
of substitution (MRS)
between income and
leisure. The MRS at point
A is given by the slope of
the dashed line.
0 20 30 40 100 Hours of
: leisure
per week
to compensate for the loss of 10 hours of leisure time? Point D gives the answer,
which is $100. To keep utility constant, how much income would have to be
given to induce this person to give up another 10 hours of leisure (that is, to have
only 20 hours of leisure)? Would he or she be willing to trade away that 10 hours
of leisure for another $100? The answer is most likely no—with fewer and fewer
able. Thus, to persuade the person pictured in Figure 2.2 to reduce leisure from
30 hours to 20 would require not an additional $100 but, it is assumed, an addi-
tional $200. This combination of income ($700) and leisure (20 hours) is shown
as point E on indifference curve [,.
This
©). The MRS t
measures the rate at which a person is willing
; ? >» MR‘ +4
Se ae acta > at point A in Figure 2.2, for example,
is given by the slope of the broken tangent line. (The slope of a line is given by
the rise over the run, measured here as Aincome/Aleisure.) As leisure hours de-
crease from points A to D to E, the’slope becomes steeper and steeper, illustrat-
ing the diminishing rate atwhich the person iswilling to trade leisure for
‘income.
FIGURE 2.3 Inconsistent Preferences and Differences in Preferences for Income versus Leisure
One property of indifference curves is that they never intersect one another. If they did, as shown in graph (a), preferences
would no longer be consistent. Thus, points V and X are of equal utility as are points Z and X. Logic implies that points V and Z
should also be of equal utility, but clearly they are not since Z provides more of both income and leisure. A second property
of indifference curves is that their shape and degree of convexity will vary from person to person depending on individual
preferences regarding income versus leisure. The indifference curve /, represents a “workaholic” person who willingly trades off
an hour of leisure for only a small increase in income. The curve /g represents a “laid-back” person who will give up an hour of
leisure only in exchange for a relatively large increase in income.
(a) An Individual with (b) Two Individuals with

Inconsistent Preferences Different Preferences
Ig
Ia “Laid-back”
person
x “Workaholic”
person
Hours of Hours of
leisure leisure -
A third property ofindifference curves isthat there isawhole set ofseparate _

curves such as 1,, 1,, and I, for each level of utility. Having derived one indif-
ference curve such as I,, for example, it is then an easy matter to increase both
income and leisure, move to a new point such as B, and at this higher level of
utility derive an entirely new indifference curve, such as I,. Every point on I,
represents a higher level of utility than I,, and every point on I, corresponds to a
higher level of satisfaction than on I).
es. This is illustrated in

graph (a) of Figure 2.3. From the definition of an indifference curve, a person
would be indifferent between point V or point X on indifference curve I,; like-
wise he or she would be indifferent between points Z and X on indifference
curve |,. Logic suggests that if points V and X and points Z and X, respectively,
yield the same ied of utility, then points Z and V should also be equal. But in
Figure 2.3, pointsS# and Z are not equally desirable, since Z gives more of both
The two indifference curves in grap of Figure 2.3 illustrate this. The indif-
ference curve of person A (I,) is relatively flat; that for person B (I,) is rela-
tively steep. The flat curve I, indicates that person A would willingly give up
another hour of leisure in exchange for a relatively small increase in income; the
steep curve I,, on the other hand, says that person B must be offered a much
larger increase in income to induce him or her to work another hour.
There are at least three possible sources of these differences in tastes. One may
simply be innate personality differences. Person A may be a “workaholic” by na-
ture, requiring little monetary inducement to work extra hours; person B, on the
other hand, may possess a “laid-back” personality and, thus, place a high value
on every hour of leisure. A second reason is related to the type of work people
do. mp On D fe fas he nNditteren e € = ¥ Ope
A UTRRNA TR RE nee «fi B, for example, may have a rela-

tively disagreeable job, such as assembly-line worker or janitor, while person A
. . . . rs
may have an agreeable job, such as airline pilot or college professor. Since each
hour of work for person B involves a greater amount of disutility, he or she would
require a larger increase in income before willingly giving up another hour of
leisure, making I, steeper than I,. Finally, a third factor that might make person
B’s indifference curve steeper is if he or she has a relatively more valuableuse for.
leisure
time than person A, such as attending school or caring for children.
Wages, Income, The demand for a good or service is shaped not only by preferences, but also by
-and the Budget amma ear saincemne. Consider first the price of leisure.
Constraint An hour of leisure has no necessary explicit cost—it is possible to sit under a
tree and daydream without spending a cent. There is a definite cost to that hour
of leisure, however, in the sense of an opportunity cost. Every hour spent in lei-
sure is an hour that could have been devoted to market work; the opportunity
cost of anhour of leisure, therefore, is equal to the wage rate per hour of work.
The higher the wage is, thehigher istheprice ofleisure.
The wage earned per hour multiplied by the number of hours worked per week
yields total weekly earnings from work. relationship
This between the wage rate,
i ; it shows all
the various combinations of income and hours of work (and thus of leisure) that
are available to an individual, given the wage that he or she can earn in the
market.
Figure 2.4 shows a graph of the budget constraint. It again has income on the
vertical axis and leisure (up to the assumed maximum of 100 hours per week) on
the horizontal axis. What is different here is that the horizontal axis represents
Neiiliquunclelsiauneenshisomesstev Ae ). Hours of leisure are read from left to
right; hours of work are read in the opposite direction. When leisure is zero,
hours of work are 100, and when leisure is 100, hours of work are zero.
Each person, based on education, experience, occupation, and so on, can
earn some particular wage rate in the market. Here it is assumed to be $5.00 per
hour. The budget constraint can be derived by plotting in Figure 2.4 all the vari-
ous combinations of hours of work and levels of income available to this person.
48 Chapter 2 Hours of Work
FIGURE 2.4 A Set of Budget Constraints

hour and zero
The budget constraint AB shows all the attainable combinations of income and leisure given a wage rate of $5 per
nonlabor income
nonlabor income. An increase in the wage rate to $7 rotates the budget constraint to AC. The addition of $200 of
is
per week would then cause a parallel shift to the right in the budget constraint to ADE. The slope of the budget constraint AB
equal to —5 (the negative of the wage rate), showing that each additional hour of leisure results in a reduction of income of $5.
Income (Y)
$900 KE
$700
$500 &B
Slope = —5
$200
0 100 Hours of leisure —>

100 0 Hours of work (H) <-
If zero hours are worked (100 hours of leisure), and assuming for the moment no
other sources of income, total weekly income is also zero. This is shown as point
A. Likewise, if all 100 hours are worked at $5.00 per hour, weekly income will
be $500, shown as point B. It is possible to plot every other combination of
_income and hours worked; doing so yields a straight lineor budget constraint
shown
as line AB.’
There are several important properties of the budget constraint. First, its slope
is negative, reflecting the fact:that income falls'as hours’of leisure tise. Second,
the slope of the budget constraint equals the negative of the wage rate. In this
Car ee
‘It is assumed for the sake of simplicity that each hour of work pays a constant $5.00 per hour.
There are several other possibilities, however. One is the presence of an overtime law that requires
time-and-one-half to be paid after 40 hours. The effect of such a law on the budget constraint is left
as an exercise for the student (see Review Question 1 at the end of the chapter). A second possibility
is that the wage varies with the number of hours a person is willing to work. One study (Robert
Moffitt, “The Estimation of a Joint Wage-Hours Labor Supply Model,” Journal of Labor Economics 2,
no. 4 (October 1984): 550-566] found the budget constraint actually had an inverse S-shape—the
wage rate rose by about 5 cents (in 1972 dollars) for each 5 additional hours worked until 35 hours
per week, after which it levelled off and then began to decline. Also see Shelly Lundberg, “Tied
Wage-Hours Offers and the Endogeneity of Wages,” Review of Economics and Statistics 67 (August
1985): 405-411.
Chapter 2 Hours of Work 49
A third property of the budget constraint imanecntinondlithinlitOalntis

changéin the wage rate. For example, what would happen to the budget con-
straint in Figure 2.4 if the wage increased from $5.00 to $7.00 per hour? Point A
would remain the same since no matter how high the wage became income
would still be zero if hours of work were zero. If all 100 hours were devoted to
work, however, total weekly income would become $700, or point C. Thus, a
in the wage and flatter for a fall in the wage.

a
. Assuming a wage of $7.00 and zero nonlabor in-
come, how does the budget constraint AC change if the person receives $200
per week of nonlabor income in the form of, say, rental income from property? If
hours of work are zero, total weekly income will be $200, or point D in Figure
2.4, representing zero dollars of earned (labor) income and $200 of rental in-
come. If all 100 hours are devoted to work, total income will be $700 of earned
income plus $200/0f nonlabor income for’a total of $900 per week (point E).
This yields a constraint
new budget ADE in Figure 2.4. The addition of non-
labor income results in a parallel shift of the budget constraint, from AC to
ADE, not a change in its slope since, by assumption, the wage remained con-
stant at $7.00 per hour.
The Equilibrium An individual’s decision concerning hours of work is the result of the interaction
-Hours of Work of preferences, wages, and income. To demonstrate this, Figure 2.5 brings to-
gether in one diagram both the series of indifference curves I,, I,, and I, from
Figure 2.2 and the budget constraint AB from Figure 2.4.
FIGURE 2.5 Equilibrium Hours of Work

The equilibrium hours of work is 50 per week (point X), where the indifference curve /, is tangent to the budget constraint AB.
Point Z would yield a higher level of utility, but it is unobtainable since it lies to the right of the budget constraint. Point V is
obtainable, but would not maximize utility since it is on a lower indifference curve.
Income (Y)
0 50 100 Hours of leisure >

100 50 0 Hours of work (H) +
—in other
words, to reach the ary The level of utility that
is obtainable is constrained by the wage rate and amount of nonlabor income.
not an
Given this, it is clear, for example, that point Z on I, is attainab lecom-
bination of income and leisure since it lies to the right of the budget constraint.
Conversely, point V on I, is attainable, but it does not maximize utility. By mov-
ing down the budget constraint, it is possible to reach successively higher indif-
ference curves until point X is reached, where the indifference curve I, and the
t. Point X is therefore the point of maxi-
mum utility, yielding an equilibrium combination of 50 hours of work, 50 hours
of leisure, and $250 of income.
The equilibrium at point X has the unique property that it is the only point
curve: Since the slope of the indifference curve is given by the marginal rate of
substitution, and the slope of the budget constraint is the wage (neglecting the
minus sign), the equilibrium hours ofwork is given by the condition
MRS = W. (2.1)
This equality has a special economic interpretation that further demonstrates
why point X is the optimal level of hours of work.
The slope of the budget constraint measures the wage rate, the dollars that are
obtainable from an additional hour of work. The slope of the indifference curve
measures the number of dollars that the individual psychologically feels each
hour of leigyre is worth. To maximize utility, the appropriate decision rule is to
keep on working additional hours as long as the wage earned exceeds the psycho-
logical valuation of that hour of leisure; when the two become equal, utility will
be
highest.
at its For hours of work less than point X, the slope of the budget
constraint is greater than that of the indifference curve and, thus, utility could
be increased by working additional hours. Utility would continue to increase un-
til hours of work reached point X, where the two slopes are equal. For hours of
work to the left of point X the wage would be less than the person’s psychological
valuation of time, and he or she would not voluntarily work additional hours.
Hours of Work The tangency between a budget constraint and indifference curve yields the
and Changes in . The next ques-
Nonlabor Income tion to consider is how the number of hours supplied to the market changes
when either nonlabor income or the wage rate change. First the effect of a
change in nonlabor income is considered.
Before the addition of any nonlabor income, the budget constraint in Figure
2.6 is AB, and the equilibrium hours of work are H, per week at point X. Should
this individual now receive Y, amount of nonlabor income, the budget con-
straint becomes ACD. Because the budget constraint has shifted to the right, it
is possible to reach a higher indifference curve and level of utility. The optimal
combination of leisure and income is given at point Z, where the new budget
constraint is just tangent to the higher indifference curve 1,.
FIGURE 2.6 The Effect on Hours of Work of an Increase in Nonlabor Income

The addition of $Y, nonlabor income shifts the budget constraint from AB to ACD. Since leisure is assumed
to be a normal good,
equilibrium hours of work fall from H, (point X) to H (point Z).
Hours of leisure —>

ywcomt 2d) Hours of work (H) =—
With the addition of nonlabor income, what happens to hours of work? Like
most goods and services,
This is re-
flected in Figure 2.6 by the movement from point X to Z. At point Z hours of
leisure are now greater, while hours of work have fallen from H, per week to H,
per week.
This relationship between changes in income and changes in hours of work
leads to one of the most important concepts in the theory of labor supply. This
concept is known as the income effect, defined as the change in hours of work
(AH) resulting from a change in income (AY), keeping the wage rate constant
(W). If leisure is a normal good, the sign of the income effect must be negative:
2 eg!
Income effect = Ay lw <4) (222)
The income effect attempts to capture only the pure, isolated influence of
changes
of income (the income may be from any source), netting out the influ-
ence of possible changes in the price of leisure.* Graphically, the income effect is
given by the change in hours of work resulting from a parallel shift of the budget
constraint as a result of changes in income, such as from AB to ACD in Figure
- +A good example of the income effect in action is the labor supply response of people who have
won a multimillion dollar prize in a state lottery. A study of 1,000 people who won $1 million or
more found that a year afterward, 23 percent had at least temporarily quit working, 17 percent had
permanently retired, and 30 percent of the winners’ spouses had quit working. The study concluded
that while labor supply is responsive to changes in income, many people’s attachment to work is
quite strong. See “What Lottery Winners Prove about the Work Ethic,” Business Week (October 7,
1985) 2
2.6. The income effect, therefore, gives rise to a change in the level of utility as
the person moves from one indifference curve to another. In Figure 2.6, the in-
. To accurately measure the pure effect of
income on hours of work, the slope of the budget constraint cannot change as it
shifts from one position to another. To change the slope causes a change not
only in income, but also in the wage, violating the very definition of the income
effect.
Hours of Work and The next issue to analyze is the effect of a change in the wage rate on hours of
Changes in the work. This is illustrated in Figures 2.7 and 2.8, for two different individuals.
Wage Rate The wage rate is assumed to be W, per hour, giving rise to the budget con-
straint AB and, accordingly, an equilibrium number of hours of work of H, per
week (point V) in both figures. What would happen to hours of work if the wage
rate were to increase from W, to W, per hour? For both individual 1 and individ-
ual 2, :
As before, the new point of maximum utility is where the new budget constraint
AC is just tangent to the highest indifference curve. In Figure 2.7 this occurs at
point X on indifference curve I,; in Figure 2.8 this occurs at point X’ on indif-
ference curve I}.
What happens to hours of work as a reaction to this increase in the wage rate?
In Figure 2.7, the
Conversely, in Figure 2.8 the new
equilibrium at point X’ indicates a decrease in hours worked from H, per week to
H; per week. These diagrams indicate that it is possible for hours of work to go
either up or down as a result of a change in the wage rate. The next question to
consider is why this is so.
The reason for this discrepant behavior is that a change in the wage rate sets
off two conflicting and opposing influences on the demand for leisure. First, a
rise in the wage rate means that if the person works the same number of hours as
before, the total amount of weekly income will be higher. Avrisesinsincome
effect. There is a second force that —_ in i opposite mii sere

Not only does a rise in the wage rate cause an increase in income, it also causes
an increase inthe opportunity cost orprice ofleisure. The theory of demand
quigeeeraennetimsienaispetanininemisemmnentheniemenddenioian |
should decline, causing an individual to work more hours per week. This is
called
substitthe effect.
ution
The essence of the substitution effect is that when the price of leisure goes up,
the individual is motivated to substitute away from the now more expensive lei-
sure toward additional hours of work. The sign of the substitution effect must be
positive, therefore. To measure the substitution effect, it is necessary to let the
relative ) change while everything else is held con-
stant, particularly the person's level of income or, more precisely, the level of
utility associated with that income. Thus, the substitution effect is defined as:
AH
Substitution effect =
aw |y> ° (2.3)
FIGURE 2.7 A Wage Change Leading to Increased Hours of Work

A rise in the wage rotates
the budget constraint from Income (Y)
AB to AC. Equilibrium
hours of work increase
from H, (point V) to H,
(point X). The rise in the
wage to WM sets off both
an income and substitution
effect. The income effect
Causes a reduction in
desired work hours from
H, (point V) to H3 (point Z).
The substitution effect
Causes an increase in
H3 (point Z) to H, (point
X). In this example, the
substitution effect out- A
weighs the income effect,
and hours of work rise by H, H, Hs | Hours of leisure —>
the net amount of H, — H,. <— Hours of work (H)<—
FIGURE 2.8 A Wage Change Leading to Decreased Hours of Work

In this case the rise in the
wage causes the person to Income (Y)
reduce desired hours of
‘work from H, (point V) to
H (point X’). The income
effect causes a decline in
H, to H3 (point V to Z’);
the substitution effect
causes.an increase in
desired work hours from H3
to H» (point Z’ to X’).
Because the income effect
is quantitatively larger, the
rise in the wage results in
a net decrease in desired
hours of work from H, to
H5. A
H, H’, HH’, Hours of leisure —>
Hours of work (H) —
where Y signifies that all other things such as income that affect a person's level
of utility are being held constant.* Graphically, the substitution effect is shown
as the change in hours of work, resulting from the rotation of the budget con-
straint along a given indifference curve.
5An intuitive example of the substitution effect is the following: Consider a person working 40
hours at $4.00 per hour, yielding a weekly income of $160. Allow the wage to increase to $5.00 per
hour, but also charge this person a $40 lump-sum tax in order to keep income at the original level of
$160. The prediction of the theory is that the person will decide to work more hours now that the
price of leisure is higher.
t a higher wage rate income is also greater, leading a person to “buy”

more leisure and less work. Simultaneously, however, this higher wage rate also
means each hour of leisure costs more, causing a substitution away from leisure
and toward more work. The income effect from a wage increase tugs the individ-
ual toward less work; the substitution effect pulls him or her toward more work.
Whether hours of work actually increase or decrease depends on the relative
strength or size of the substitution effect versus the income effect.
A Graphic Hours of work change due to the wage increase from H, to H; in Figure 2.7 and
Derivation of from H, to H} in Figure 2.8. It is possible to graphically decompose this change
Income and in hours of work into two parts: the part due to the change in income (the in-
Substitution Effects come effect) and the part due to the change in relative prices (the substitution
To isolate the income effect in both diagrams, the following hypothetical ex-
periment can be done. Starting from point V and the original budget constraint
AB, increase each individual’s income just enough to reach the higher level of
utility associated with the indifference curve I, or Ij. Doing so causes the budget
constraint (shown as the broken line in each figure) to shift outward in a parallel
direction until it just touches I, at point Z in Figure 2.7 and I} at point Z’ in
Figure 2.8. S
the
ity as obtained at points X and X’? The answer is that hours of work would de- |
point V to Z) in Figure 2.7, and from H, to H} (point V to
Z') in Figure 2.8. This decrease in hours measures the pure income effect.
Points Z and Z’ are not the actual utility maximizing combinations of income
and leisure; points
X and X’ are. What accounts for the disparity? The answer is
the substitution effect.
The rise in the wage from W, to W, per hour increased both income and the
price of leisure. The movement in each diagram from point V to points Z and
Z’, respectively, represented the income effect, holding the wage at W,. The
substitution effect can be isolated by returning to points Z and Z’ and raising the
price of leisure (the wage) but holding constant the level of utility that the per-
son’s income could purchase. This can be accomplished graphically by the fol-
lowing hypothetical exercise: beginning at points Z and Z’, rotate the broken
line along indifference curves I, and I; until the slope increases from —W, to
—W,, whi . (When the
slope of the rr slope of the
solid line AC.) The increase in hours of work from H; to H, (point Z to X) in
Figure 2.7, and H; to H; (point Z’ to X’) in Figure 2.8, measures the pure sub-
stitution effect from the change in the wage.°
“It is possible to derive the income and substitution effects in reverse order, doing the substitution
effect first and then the income effect. To do so, start at point V and rotate a broken line along I,
until it has the slope of AC, then shift the broken line in a parallel direction to point X (X').
FIGURE 2.9 A Forward Sloped and Negatively Sloped Labor Supply Curve
The supply curve of labor may have either a positive or negative slope. In graph (a), a rise in the wage rate from W, to W, causes
an increase in desired work hours from H, to H>. In this graph the substitution effect outweighs the income effect. In graph (b)
a rise in the wage rate from W, to W, causes a decline in desired work hours from H, to H3. In this case the income effect ;
outweighs the substitution effect, causing the supply curve to have a negative slope.
(a) (b)°
Wage Wage
rate (W) Ss, rate (W)
W, W,
Ww, W,
H, H, Hours of 0 H’, H, Hours of

work (H) work (H)
In Figure 2.7 the higher wage rate led to a net increase in individual 1’s labor
supply from H, to H, hours. In Figure 2.8, however, the wage increase caused
individual 2 to reduce hours of work from H, to H3. For each individual the
ind du 2, however, thei ne effect dominated the substitution effect, and
housolanarkedastinecs The reason for this divergent behavior is attributable

entirely to the different shapes of the indifference curves, reflecting the different
preferences of each individual regarding income versus leisure. Since preferences
vary from one person to another, it is impossible to say a priori which effect will
dominate for a given individual.
The Supply Curve A supply curve of labor represents the

of Labor the hours of labor supplied to the market. The derivation of a labor supply curve
is illustrated in graphs (a) and (b) of Figure 2.9. Each diagram has the wage rate
(W) on the vertical axis and hours of work per week (H) on the horizontal axis
(note that H goes from left to right in this graph). Point V in Figures 2.7 and 2.8
indicates that at a wage of W, per hour, labor supply is H, hours per week. This
combination of the wage rate and hours of work is plotted as point V in graphs
(a) and (b) of Figure 2.9. If the wage should go up to W,, what will happen to
hours of work? Points X and X’ in Figures 2.7 and 2.8 give the answer: for indi-
vidual 1, hours of work increase to H, hours per week, and for individual 2,
hours of work decline to H} per week. In graph (a) of Figure 2.9, plotting point
X and then connecting points V and X with a straight line yields the supply
curve of labor S,. (Other points on S, can be derived in a similar manner.) In
eraph (b) of Figure 2.9, plotting point X’ and then connecting points V and X’
yields the labor supply curve S,. For person 1the supply curve S} slopes upwards
qidteonelsexrighemindicacingatl saci
wage rate. This, in turn, reflects the fact that the substitution effect
higher
a
This is due to the larger size of the income effect relative to the substitution
Althouth not shown here, it is possible that the supply curve of labor could
contain n. One example is what
economists call a b A backward bending
supply curve is positively sloped at low wage rates but negatively sloped at high
wage rates. The presumption behind a backward bending supply curve is that at
low wages the desire for additional income is so great that the substitution effect
will outweigh the income effect; beyond a certain wage rate, however, the per-
son's income is sufficiently hjgh that he or she would react to a still higher wage
by buying more leisure and reducing hours of work.
Empirical Evidence
Labor Supply Curves Estimated from Cross-Sectional Data
Economists have conducted extensive empirical research in an attempt to de-
termine the actual shape of the labor supply curve. Is it positively sloped, nega-
tively sloped, or backward-bending? To answer this question, economists
typically use cross-sectional data on hours of work, wages, income, and other
variables collected in large surveys of individuals to estimate the labor supply
curve and the underlying income and substitution effects.
The process of estimating a labor supply curve from cross-sectional data is
fairly straightforward, although, in practice, there are a number of complex
theoretical and empirical issues involved.’ The basic idea is first to obtain data
on the wage rate, hours of work, age, sex, education, occupation, and other
such variables for a large number of individuals. Then the individuals are sorted
or standardized into homogeneous groups to eliminate the influence of differ-
ences among them in nonlabor income and preferences for income versus lei-
sure. Next the data on hours of work and wage rates for each individual in a
group are plotted. Passing a line through this scatter of points will then trace out
‘For a nontechnical discussion of some of these issues, see Cynthia Lloyd and Beth Niemi, The
Economics of Sex Differentials (New York: Columbia University Press, 1979): 61-66. The more ad-
vanced student might consult Mark Killingsworth, Labor Supply (Cambridge, England: Cambridge
University Press, 1983).
FIGURE 2.10
Estimated Labor Hourly.
Supply Curves in wage
Five Studies
$15
$12 Ae os seceee Kalactok

be & Raines
., = ome coe er al
\°
$9 \ Greenberg
PO a & Kosters
_.. .. Cohen, Rea

& Lerman
$6
Hill
$3
1500 - 2000 2500 3000 Annual hours

of work
source: Robert E. Hall, “Wages, Income and Hours of Work in the U.S. Labor Force,” in Glen G. Cain and Harold W.
Watts, eds., Income Maintenance and Labor Supply (Chicago: Rand McNally, 1973): 102-162. The wage rate figures
on the vertical axis are expressed in 1984 dollars.
the supply curve of labor. This entire process is done today by means of a com-
puter and the statistical technique of linear regression.°
Given the mechanics of estimating a labor supply curve, what are the results
from cross-sectional studies? For adult men, nearly all studies find the labor sup-
ply curve to be negatively sloped or backward-bending. Figure 2.10 shows the
labor supply curve for men estimated in five different cross-sectional studies.”
Three of the studies (Hill; Cohen, Rea, and Lerman; and Greenberg and Kos-
’See the Appendix to Chapter 2 for a more in-depth discussion of linear regression and an ex-
ample of its use in estimating a labor supply curve.
_°The supply curve for the Cohen, Rea, and Lerman study, for example, is for men aged 22 to 54
in 1966 who had $2,000 per year in nonlabor income, were neither ill nor in school, worked one or
more weeks, completed 12 years of school, were married with spouse present, had two children under
18, neither lived in a poverty tract nor in the South, were white, and were not self-employed. If
persons with $3,000 per year of nonlabor income had been analyzed, the labor supply curve would
shift to the left, showing that at any given wage rate fewer hours would be supplied, assuming leisure
is a normal good.
a
IN « THE « NEWS
Why Do the Japanese Work So Hard;
and Are the French Lazy?
IBM Corporation recently calculated the annual wise, if a group of Japanese want to play a game of
hours of work put in by their employees at facili- softball, they will often have to leave home early
ties in ten different countries. At the top of the in the morning and drive 2 to 3 hours to find a ball
list was Japan, where the average IBM employee field. Another discouragement to the consump-
worked 1,964 hours per year. The United States tion of leisure is that Japanese homes are notori-
was second at 1,873 hours, and in tenth place was ously small and uncomfortable (at least by western
France at 1,612. For every 40 hours worked by a standards), making family activities and social en-
French employee, a Japanese employee worked tertainment less attractive.
about 48. Apart from these economic considerations, the
Why do the Japanese work so hard? Part of the Japanese also seem to have, in the words of one
answer reflects economic considerations and part newspaper article, a “lust for labor.” The article
may reflect unique aspects of Japanese culture. explains, for example, that the Japanese govern-
Many leisure time activities in Japan are astonish- ment recently set up an agency called the Leisure
ingly expensive in both time and money due to the Development Center to encourage people to work
scarcity of land and the abundance of people. An less. The staff members of the center pursued their
18-hole round of golf, for example, costs more goal so zealously, however, that they regularly
than $100 at many courses, and country club worked Saturdays and some Sundays and only took
memberships have sold for over $1 million! Like- 40 percent of their allotted vacation time! As one
ters) found the labor supply curve to be negatively sloped throughout; the other
two (Hall; and Kalachek and Raines) found a positively sloped segment at low
wage rates, which then became backward-bending as the wage rate increased.
Because the labor supply curve for men is negatively sloped over all or most of
its range, the implication is that the negative income effect from higher wages
dominates the positive substitution effect. This has been the result found by
nearly all cross-sectional studies. For example, a consensus estimate is that the
labor supply elasticity for men is about —0.1, meaning that an increase in the
wage rate for men of 10 percent will, on average, cause a net reduction in hours
of work of about 1 percent."° This reflects, in turn, the combined influence of a
Labor supply elasticity is defined as the percentage change in hours supplied divided by the per-
centage change in the wage rate. These figures are an average from 19 different studies as reported in
Michael Keeley, Labor Supply and Public Policy (New York: Academic Press, 1981), Table 3: 98-99.
These estimates are from nonexperimental data sources. A second source of labor supply elasticities
is from the negative income tax experiments sponsored by the federal government in the 1970s.
These data yield income effects that are of the same magnitude as found in nonexperimental studies;
the substitution effects, however, are generally smaller. See Robert Moffitt and Kenneth Kahres
“The Effect of Tax and Transfer Programs on Labor Supply: The Evidence from the Income Mainte-
nance Experiments,” in Ronald G. Ehrenberg, ed., Research in Labor Economics, vol. 4 (Greenwich
Conn.: JAI Press, 1981): 103-150.
a
ii ar a
observer noted, “Japanese don’t have a strong ethic and are now a nation of sunbathers and cafe
sense of leisure. They often don’t distinguish habitués.
clearly between work and leisure.” Whether or not the French are lazy is currently
While the Japanese are concerned about work- being hotly debated on French television talk
ing too much, the French are debating whether shows and in the popular press, no doubt to the
they work too little. Not only do the French work enrichment of Mr. Scherrer. Perhaps the best way
significantly fewer hours per week than the Japa- to resolve the debate would be to have the French
nese, they take many more holidays and vaca- and Japanese switch countries. If the French began
tions. (Between May 1 and June 5 of each year, to work Saturdays and Sundays in Japan while the
the French have four extended-weekend holidays, Japanese took frequent 3-day vacations to the
plus many French workers vacation during much Riviera, we could surmise that economic condi-
of August.) Part of the greater demand for leisure tions, rather than differences in tastes, were pri-
on the part of the French no doubt reflects the marily responsible for the cross-country pattern of
lower prices of leisure time activities. Dining at a work hours. Unfortunately, such a test is not pos-
restaurant in Paris, for example, is far cheaper sible, so we are left to conjecture on the relative
than in Tokyo. It is also possible, however, that roles of culture and economics in the work deci-
the French have a greater “taste” for leisure rela- sions of the two peoples.
tive to the Japanese. This thesis has been force-
fully argued by French author Victor Scherrer in
Sources: “Lust for Labor,” The Wall Street Journal (April 21,
his provocatively titled book Lazy France. Accord- 1986): 9D; and “New French Cause Celebré: A Book Calling
ing to Scherrer, the French have lost the work Nation Lazy,” New York Times (June 20, 1987): 1.
negative income effect from higher wages of —0.2 and a positive substitution
effect of +0.1.
Although hours of work for men decrease with higher wages, this decrease is,
quantitatively speaking, not very large, particularly for prime-age males (age 25
to 54). Thus, a 10 percent wage increase (for a given level of prices and all other
factors) would cause a male working.2,000 hours a year (40 hours per week for 50
weeks) to cut back on work only 20 hours a year, on average.
The labor supply curves estimated from cross-sectional data for women have
been substantially different compared to men in two important respects. First,
empirical studies have found that the labor supply curve is forward-sloping;
women, on average, work more hours at higher wage rates, not less as for men.
This increase in hours comes from two sources: the entrance of women to the
labor force to begin work, and longer hours of work by women who already hold
jobs. An average estimate derived from a large group of studies of adult women is
that the labor supply elasticity for females is about 0.9; that is, for a 10 percent
increase in wages, women will increase desired hours of work by about 9 percent,
reflecting a negative income effect of —0.2 and a positive substitution effect of
+1.1." Thus, while the negative income effect for women is about the same size
Keeley, Labor Supply and Public Policy, Table 5, 102. These estimates are based on 11 studies.
as for men, the positive substitution effect is about ten times larger. The second
important difference between men and women, therefore, is that female labor
supply not only increases with higher wages but the response is also quanti-
tatively much larger.
What accounts for this discrepant behavior between men and women? Al-
though a full discussion of this issue must be postponed until the next chapter,
the basic explanation can be briefly stated here. Since over 90 percent of adult
men work full-time, additional hours of work must come primarily from one
source: leisure. A different situation exists for adult women, however. Many
women devote a considerable amount of time to work in the home. These
women, therefore, have three alternative uses of time: market work, home work,
and leisure. A rise in the market wage induces a substitution not only from lei-
sure to market work, but also from home work to market work. It is this substitu-
tion from home work to market work that accounts for the large, positive labor
supply elasticity for women.
Policy Application
The Reagan Tax Cuts and Labor Supply
During the 1960s and 1970s, the twin forces of economic growth and inflation,
together with the progressive nature of the income tax system, pushed the aver-
age wage earner into successively
higher income tax brackets. In 1961, for ex-
ample, only 10 percent of income tax returns had a marginal tax rate (the tax
rate on the last dollar of earnings) exceeding 22 percent; by 1979, this propor-
tion had more than tripled to 35 percent.'* This increase in tax rates was particu-
larly great for high-income families; families with twice the median income
experienced a rise in marginal tax rates from 22 percent in 1960 to 43 percent
in 1980.
An income tax system forms awedge between aworker's gross wage and net or
)
“take-home” wage; for a given gross wage, the larger the marginal tax rate is, the
lower is the additional net income earned from working offe more hour. A num-
ber of economists and policymakers (most notably, President Reagan) believed
that the rise in marginal income tax rates noted above caused a serious decline in
work effort and labor supply precisely because the growing tax wedge stifled indi-
viduals’ incentives to work. For this reason, the centerpiece of the Reagan eco-
nomic program in his first term was a
over
3 years. The major purpose of the cut was to
and
investment. In 1986, income tax rates were lowered further as part of the
“Charles R. Hulten and June A. O'Neill, “Tax Policy,” in John L. Palmer and Isabel V. Sawhill,
eds., The Reagan Experiment (Washington, D.C.: Urban Institute, 1982), 97-128 and Table A—6.
FIGURE 2.11 The Effect of an Income Tax on Hours of Work

With the imposition of a progressive income tax, the budget constraint rotates from AB to AC. The equilibrium hours of work
decline from H, (point V) toA, (point X). The income tax sets off an income effect that causes the person to desire more hours of
work (point V to Z) and a substitution effect that causes the person to desire fewer hours of work (point Z to X). In this example,
the substitution effect outweighs the income effect, and the net result is a decrease in labor supply. The opposite could also
happen, however.
Income (Y)
H3H, =F Hours of leisure — >

Hours of work (H) <—
Reagan tax reform package. Altogether, the two pieces of legislation reduced the
top marginal tax rate on income from 70 percent to 28 percent.
Although it seems reasonable that higher tax rates, by reducing the net earn-
ings from an hour of work, would cause people to work less, the answer is by no
means so simple.A number of economists have argued that the Reagan tax cuts
may, in fact, cause people to work less, not more.’? What effect does an income
tax have on labor supply? The labor/leisure model is used to answer this question
for the case of an individual worker, and then the implications are considered for
labor supply in the aggregate economy.
Graphic Analysis Figure 2.11 depicts the labor supply decision of a represen-
tative individual before and after the imposition of a progressive income tax on
earnings. By definition of a progressive income tax, the marginal tax rate t in-
creases with the level of earnings. To begin, assume there is no income tax (t =
0), the person’s wage is W,, and he or she has zero nonlabor income. These as-
sumptions give rise to the budget constraint AB in Figure 2.11; with zero hours
of work, earned income is zero (point A). With each additional hour of work,
income rises by W, dollars up to a maximum of Y; amount (point B). Given the
budget constraint AB, and assuming the person has the indifference curve [,,
B James Tobin, “Supply-Side Economics: What Is It? Will It Work?” Economic Outlook USA 8,
no. 3 (Summer 1981): 51-53.
the tangency between the two at point V yields the equilibrium number of hours
of work of H).
Next consider the impact on hours of work if a progressive income tax is im-
posed on labor eamminey Alchoue the gos wage paid the worker remains Wi,
thé a ances Fi ete REE Sey Pac es
. The effect of the neon tax, then, is to pie rise to the
new, curved budget constraint AC. For zero hours of work, total income is still
zero (point A); with each additional hour of work, however, the net increase in
earnings is no longer W, but the lesser amount (1 — t) W,, causing the budget
constraint to rotate to the left from AB to AC. The slope of the budget con-
straint AC also becomes progressively flatter as additional hours are worked,
since the marginal tax rate t rises with additional earnings.'* The maximum
amount of earnings available to the individual even if all 100 hours are worked is
reduced to Y. dollars (point C); the area between AB and AC thus shows the
total wedge between gross and net earnings caused by the income tax.
Given the new budget constraint, utility is maximized at point X, where the
indifference curve I, is just tangent to AC. The key issue is whether hours of
work at this new equilibrium will be more or less than they were before the im-
position of the tax. The answer depends on the size of the income and substitu-
tion effects. The income tax results in a substitution effect that motivates the
leads to a decrease in the demand for all normal goods, including leisure. Which
effect will dominate? The answer depends on the shape of the indifference
curves, or more generally, the preferences of each individual regarding income
versus leisure. Some people who have a dogged determination to maintain or
increase their standard of living will work longer in reaction to the tax; others,
however, will respond more strongly to the disincentive effects of the tax and
will cut back their hours of work. In the example shown in Figure 2.11, the
income tax leads to a net reduction in hours of work from H, to H, (point V to
X), implying that the substitution effect outweighed the income effect.
The income effect can be identified by beginning at point V and decreasing
income just enough to reach the new indifference curve I,, holding the price of
leisure constant. This is represented by the broken line tangent to I, at point Z.
Assuming leisure is a normal good, the reduction in disposable income due to
the tax should cause an increase in hours of work, shown by the increase in
weekly hours from H, to H; (point V to Z). The substitution effect can be deter-
mined by rotating the broken line along I, until it has the same slope as the
‘The budget constraint AC is drawn as a smooth, continuous line for ease of exposition. It actu-
ally should be drawn with a series of linear segments corresponding to each discrete tax bracket.
budget constraint AC. The result of doing so is a reduction in hours of work of

H, — H; (point Z to X). The income effect, therefore, leads to an increase in
hours from H, to H;, and the substitution effect leads to a decrease in hours
from H, to H). The observed result is the net decrease in hours from H, to H).
The outcome shown in Figure 2.11 is only one of several possibilities. A dif-
ferent set of indifference curves can be drawn in Figure 2.11 that results in a net
increase in hours of work. In this case, the income effect would dominate the
substitution effect.
Income Taxes and Aggregate Labor Supply The basic conclusion of the
labor/leisure model is that in reaction to a tax cut a person may either work more
or work less, depending on the relative sizes of the income and substitution
effeets. What implication does this have for the labor supply impact of the
Reagan tax cut? The traditional approach to answering this question is ex-
plained first.
As discussed earlier, the estimated labor supply elasticity for men is=Orl1, im-
plying that their labor supply curve has a slight negative slope. For men, then,
the Reagan tax cut would increase the net take-home wage, causing a movement
up the supply curve and a small reduction in hours of work. If men were the only
group in the labor force to receive the tax cut, the critics of the Reagan tax cut
would be right; it would lead to less labor supply, not mores It%was also shown
earlier, however, that women’s estimated labor supply elasticity is quite different,
being both positive’and much larger (+0.9). Since the labor supply curve of
women is forward-sloped and more responsive to wages, the rise in the after-tax
wage would lead to a significant increase in hours of work for women, particularly
on the part of married women who would be drawn from home work to market
work by the rise in potential earnings. For women, then, the supporters of the
Reagan tax cut are right; it would lead to more labor supply in the economy.
What would be the net impact? Based on elasticity estimates such as those
cited above, Michael Evans concluded that a 1-percent cut in the marginal tax
rate would raise labor supply by an average of 0.2 percent.'’’ Given that in 1980
the median-income family was in a 24 percent marginal tax bracket, the Reagan
3-year, 25-percent cut in tax rates would lower the marginal tax rate by a total of
6 percentage points, munncammemmmrprrenes oe
ond study estimated that the Reagan tax cut would result in a
increase
in labor supply. '*®
The traditional approach to estimating the impact on aggregate supply of an
income tax cut has recently been challenged, however, by James Gwartney and
Richard Stroup.'? They claim that economists such as Evans have committed
5 Michael K. Evans, The Truth about Supply-Side Economics (New York: Basic Books, 1983),
Chapter 9.
Robert Haveman, “How Much Have Reagan Tax and Spending Policies Increased Work
Effort?” in Charles R. Hulten and Isabel Sawhill, eds., The Legacy of Reaganomics: Prospects for Long-
Term Growth (Washington: The Urban Institute, 1984).
‘James Gwartney and Richard Stroup, “Labor Supply and Tax Rates: A Correction of the
Record,” American Economic Review 93, no. 3 (June 1983): 446-451.
the fallacy of composition, that is, they have assumed that what holds true for an
individual also holds true for the economy as a whole. Gwartney and Stroup
argue that with respect to taxes and labor supply, this is not a valid assumption.
For an individual worker, a cut in income tax rates causes both a substitution
effect and an income effect since it simultaneously increases the price of leisure
and the amount of goods and services the person can buy, given his or her cur-
rent level of work hours. The net impact of the tax cut on labor supply is inde-
terminant, therefore.
If income tax rates are reduced for all individuals in the economy, however,
the result is quite different, say Gwartney and Stroup. Aneaggregatecut-in-in-
come.tax rates produces a substittition effect because*the opportunity cost-of lei-
sure.is greater for each person in.the economy. To generate an income effect, the
tax cut would also have to increase people’s level of real income (i.e., the
amount of goods and services they can buy). The crux of Gwartney and Stroup’s
argument is that this can not happen at the economy-wide level. They reason
that the total amount of goods and services (GNP) available to people is deter-
mined by the state of technology and the amount of the factor inputs land, capi-
tal, and labor utilized in production. Holding these determinants of the level of
production constant, iteis-clear-that.a.tax.cut,.by. itself;.cannot-merease-every-
one’s'level of real income since it will have produced no corresponding increase
in GNP.”
All a tax cut does in the aggregate is change*the*relativeshare of GNP that
goes
households
to versus: government. The paychecks and purchasing power of
workers increase with a tax cut, but the amount of goods and services that can be
provided by the government (e.g., defense, interstate highways) necessarily di-
minishes. The net impact on total real income in the country is zero, therefore,
since the extra goods and services households gain from a higher after-tax in-
come is just offset by the loss of government-produced goods and services. The
net-impact-of -an-aggregate.tax..cut.on. labor.supply,..therefore,.. must. unam-..
Peasiseneposli accosdindnid- iar
SRE
netaRSStioinenccause
AON ORCS psonls-aie
effect
demand-greater-hours
leads them to: of leisure. This conclusion implies, in
turn, that estimates such as Evans’ probably understate the increase in labor sup-
ply that resulted from the Reagan tax cuts."
Gwartney and Stroup’s position has not gone unchallenged. In articles subse-
quent to theirs, several sets of economists have argued that their conclusions rest
“If the substitution effect is sufficiently large, it is possible that a cut in income tax rates, by
engendering more labor supply, production, and income, may actually increase the total dollars of tax
revenue collected by the government. This would occur if the economy is to the right of the peak of
the Laffer curve (the hypothesized U-shaped relationship between the size of tax collections and the
size of the marginal tax rate). Although some supply-side economists believe this to be the case,
empirical evidence suggests that aggregate labor supply is not sufficiently responsive to tax rate
changes for a tax cut to “pay for itself.” See Don Fullerton, “On the Possibility of an Inverse Rela-
tionship between Tax Rates and Government Revenues,” Journal of Public Economics 15 (October
1982): 3-22.
on special or highly restrictive assumptions.'? When these are taken into ac-
count, it turns out that it is still possible for a tax cut such as those in the Reagan
years to lead to a reduction in aggregate labor supply. The bottom line appears to
be, therefore, that at the economywide level, a*tax cut will probably lead toa net
i Whether the actual impact of the Reagan tax cuts on
hours of work has been more or less than the estimates derived by Evans and
other economists is still unknown, however.
Policy Application
Income Maintenance Programs and Labor Supply
The discussion in the preceding section concerning the labor supply effects of
the Reagan tax cut points out one important way in which the concept of in-
come and substitution effects, and the labor/leisure model in general, can be of
considerable relevance to analyzing questions of public policy. This section fo-
cuses on a second application of the labor/leisure model to public policy, the
question of how income transfer programs affect hours of work and labor supply.
Types of Income The fa g g C government expenditures in the last 30 years

Transfer has been various income transfer or “entitlement” programs. Table 2.1 shows
Programs the major income transfer programs in the United States and the public expen-
diture on each program for 1965 and 1985. Expenditures on transfer programs
have expanded at a rapid rate, increasing from 4.6 percent of GNP in 1965 to
9.0
percent in 1985.
Income transfer programs fall into two basic groups: social-insurance.pro-
grams
maintenance
and income programs (or “welfare” programs).”° These pro-
grams have two basic objectives— i
l 2
, ATEN
David Betson and David Greenberg [‘“‘Labor Supply and Tax Rates: Comment,” American Eco-
nomic Review 76 (June 1986): 551-556] show that a tax cut may still lead to a reduction in labor
supply if some groups in the population cannot adjust their hours of work. For example, assume
government transfer payments go to disabled persons who cannot work, while the taxes to pay the
transfer payments come from able-bodied workers. A tax cut decreases the income of the former
group and increases the income of the latter group by offsetting amounts, but the change in hours of
work will not net-out—the decreased labor supply of the able-bodied engendered by the income
effect will not be offset by any increase in labor supply on the part of the disabled. It is possible,
therefore, for the tax cut to lead to a reduction in hours of work to the extent that a negative income
effect remains. For other qualifications to the Gwartney and Stroup analysis, see Cecil Bohanon and
T. Norman Van Cott, “Labor Supply and Tax Rates: A Comment,” American Economic Review 76
(March 1986): 277-279.
. 2 Arthur Williams, Jr., John G. Turnbull, and Earl F. Cheit, Economic and Social Security, 5th ed.
(New York: Wiley, 1982).
a
lr
TABLE 2.1 Public Expenditures
Expenditures for (Billions of Dollars)
Social Insurance and Progrant
ee ee
Income Maintenance Social Insurance
Programs Social Security (OASDI) $18.1 $186.1
Unemployment Insurance 2.4 14.8
Workers’ Compensation 1.2 14.3
Veteran's Disability Compensation 3.0 14.1
Railroad Retirement (loll 6.3
Black Lung — 1.0
Medicare — 47.5
Income Maintenance
Aid to Families with Dependent Children (AFDC) ee 14.9
Supplemental Security Income (SSI)* Wy 11.1
General Assistance 8} 2.4
Medicaid® 5 37.5
Food Stamps iS 10.8
Total Expenditures $31.3 $360.8
Expenditures as a Percent of GNP 4.6% 9.0%
8 Aid to the Blind, and to the Permanently and Totally Disabled, and Old Age Assistance in 1965.
> Medical Aid to the Aged in 1965.
source: U.S. Department of Health and Human Services, Socia/ Security Bulletin, Annual Statistical Abstract, 1986
(Washington, D.C.: Government Printing Office, 1987).
The first objective is largely served by social-insurance programs. Under these

programs the receipt of benefits is conditioned on the loss of labor market earn-
ings from some identifiable problém, such as old age, illness, disability, or unem-
ployment. To receive benefits, indivi sho
need; eligibility and level of benefits are tied to the amount of past contributions
into the program (as in Social Security). The second objective is served by pub-
lic assistance or welfare programs. Welfare benefits are not conditioned on con-
tributions into the program, but rather are determined by a demonstration of
“need”under
the eligibility guidelines. One characteristic of most welfare pro-
grams in the United States is that they are categorical in nature, that is, to be
eligible the individual not only has to fall under the asset and income cut-
offs, but also must possess certain characteristics (such as single mother with
children, blind, or disabled) that determine which program the individual
qualifies for.
With the rapid growth in income transfer programs has come a growing debate
as to the effects of these programs on such things as labor supply, savings, and
economic growth. One particular income maintenance program, Aid to Fami-
lies with Dependent Children (AFDC), serves to illustrate some of the issues
involved.
AFDC and Hours One of the most controversial of the income maintenance programs is Aid to
of Work Families with Dependent Children (AFDC). AFDC originated with the Social
Security Act of 1935 and was set up to provide federal grants to states to help
families with children who are in
. The AFDC program today is financed jonthe by
the federal government and the states, with each state determining the specific
eligibility criteria and benefit levels. In 1985, the program paid out benefits of
$15 billion to 3.7 million families (10.7 million persons). The monthly payment
per family averaged $348. The typical AFDC family is a single mother with two
children; 41 percent of the recipients are white; 13 percent of the mothers work
part-time or full-time.*! One reason AFDC has been controversial is that by con-
ditioning payments on the absence of a male in the household, it-allegedly-pro-
motes marital instability and abandonment of the family by fathers:”? Since pay-
mentssincrease-with=family:size, there is also concern that AFDC provides an
incen for additional
tive children, particularly among unwed mothers. In recent
years, one-half of AFDC expenditures went to women who were teenagers when
their first child was born.”
The AFDC program has several key features. One is the income guarantee
(G), the amount of money per month paid to the family if the family has zero
income from work. In 1982 the income guarantee for a family of four ranged
from a low of $120 per month in Mississippi to $634 per month in Alaska.
The second feature of AFDC and other welfare programs is the existence of an
implicit tax rate (t) on earnings. The idea is that if the family has zero income,
then the guarantee provides them with an income floor of $G. If the mother
should go to work, however, and bring in extra income, then family need is pre-
sumably less and the AFDC payment is reduced. Under the provisions of the
law, prior to 1981, for every $1 of earned income, the AFDC payment from the
state was reduced by 67¢. This reduction is in essence an implicit tax on earn-
ings; in this case t = .67, and the family’s net income rises by 33¢ for every dollar
earned
from work:
The third feature of the AFDC program is the break-even point (BE), the
l i re
i /t. For the state of Missis-
sippi, fe example, it would be $180= $120/.67. With zero earned income, the
family in Mississippi receives $120 per month; with every additional dollar
earned from work the guarantee is reduced by 67¢, until the guarantee G has
been reduced to zero at an earnings level of $180. Beyond the break-even point,
the family no longer qualifies for the program and is affected by neither the guar-
antee G nor the implicit tax rate t. From the formula for the break-even point it
is evident that the lower the tax rate or the higher the income guarantee, the +
21U.S. Department of Health, Education, and Welfare, Social Security Administration, Aid to
Families with Dependent Children: A Chartbook (Washington, D.C.: Government Printing Office,
1979).
2Twenty-three states permit AFDC payments to husband-wife families if the husband is unem-
ployed and the family’s income and asset holdings meet the eligibility requirements. A detailed de-
scription of the AFDC program, as well as all other income transfer programs, is provided in “Social
Security Programs in the U.S., 1987,” Social Security Bulletin 50, no. 4 (April 1987): 5—66.
23D, Lee Bawden and John L. Palmer, “Social Policy,” in John L. Palmer and Isabel Sawhill, eds.,
The Reagan Record (Washington, D.C.: The Urban Institute, 1984): 200. Also see Chapter 14 ior
more discussion of this and other issues concerning the social and economic impacts of welfare
programs.
FIGURE 2.12 Hours of Work with an AFDC Program

Prior to the introduction of an AFDC program, the budget constraint is AB, and the person chooses to work 45 hours per week
(point V). With an AFDC program, the budget constraint becomes ACDB. As a result, equilibrium hours of work fall from 45 to 15
(point V to X). The income effect from the AFDC program causes a reduction in desired work hours from 45 to 30 hours (point V to
Z); the substitution effect causes a further reduction in desired work hours from 30 to 15 (point Z to X).
Income (Y)
45 30 15 Hours of leisure
—>
higher will be the break-even point and the more families that would then
qualify for the program.
Effect of AFDC on Labor Supply Figure 2.12 illustrates the effect of such a
welfare program as AFDC on the labor supply decision of individual adults. It
assumes that the mother, if she works, can earn $4.50 per hour. Her budget con-
straint, in the absence of an AFDC program, is AB. If she works zero hours,
total income is zero; if she works all 100 hours that are possible per week, her
income is $450. Given the budget constraint AB, and the individual’s prefer-
ences for income versus leisure as represented by the indifference curve I,, she
will choose to work 45 hours per week (point V) and will earn $202.50.
With an AFDC program, the budget constraint becomes ACDB. At zero
hours of work, total family income is $200 (point C), comprised of zero labor
income and $G of AFDC payments, assumed in this example to be $200. If the
yi iaiilideeeaisieg erieTREHSEbSdrSOm
guarantee (Gis reduced by $3.00! The budget constraint between point C and
the break-even point (point D) will have a slope of only —1.5. Beyond the
break-even point of $300 of income, the family is no longer on the AFDC pro-
gram, and the budget constraint thus coincides with the original one from point
D to point B.
Given the new budget constraint ACDB, the individual maximizes utility at
point X where indifference curve I, is just tangent to the budget constraint. At
point X, hours of work have been reduced to 15 per week. The prediction of the
theory, then, is that the introduction of an income maintenance program such
lation: Based on previous empirical research, one study concluded that AFDC
caused recipients to reduce labor supply by an average of 600 hours of work
per year."
Income and Substitution Effects The income and substitution effects set off
by the program help explain why an income maintenance program such as
n Figure 2.12, the observed change in hours of
work due to the AFDC program is from point V on indifference curve I, to point
x on indifference curve
c I,. Tosdetermine theincome effect,incomefortheindi
sal is incr ough to reach the higher in ver'l,,
but the
waemate skepeconsan Te is represented as a parallel shift of the original
dge he broken e ngen at point Z. At pointZ
poe of work Pave ne Sealed from 45 to5 30 hours per ee which represents
the income effect stemming from the addition of nonlabor income inthe form*of
the income guarantee payment of $G. a
The AFDC program also gives rise to a substitution effect. Because of the im-
plicit tax rate t, the net hourly wage is reduced from $4.50 to $1.50 per hour
under the program, reducing,the,price,of leisure:and leading the individual:to
substitute leisure for work. This substitution effect is represented by keeping the
individual on indifference curve I, but rotating the budget constraint from a
. The resulting change in hours of work from
point Z (30 hours) to point X (15 hours) represents the substitution effect.
Under an income maintenance program, both the income and substitution
effects lead to a . The income effect on hours of work
arises from the addition of extra nonlabor income in the form of the income
guarantee G. Assuming leisure is a normal good, additional income motivates
the individual to consume more leisure and work less. This negative income
effect on hours of work is reinforced by a further reduction in hours worked due
to the substitution effect. Because of the implicit tax rate under the AFDC pro-
gram, the take-home wage from an hour of work is reduced by two-thirds. Since
each hour of leisure now has a lower opportunity cost, the individual is moti-
vated to substitute from work towards more leisure, further reducing total hours
of work. ~
Issues in Three general goals are associated with any welfare program: income provision,
Welfare work incentives, and cost minimization. First, to the extent that protecting
Reform people from the indignities of poverty is a social goal, the welfare program
should provide a relatively high income floor or guarantee. Second, if encourag-
ing people to find work and be self-supporting is also a social goal, it implies the
program should have a relatively low implicit tax rate t so as not to penalize the
rewards from work. Third, it is also likely to be a social goal that the costs of
the program be held to “reasonable” levels.
Sheldon Danziger, Robert Haveman, and Robert Plotnick, “How Income Transfers Affect
Work, Savings and the Distribution of Income,” Journal of Economic Literature 19, no. 3 (September
1981): 975-1028.
The fundamental dilemma facing policymakers is that the three goals are im-
possible to achieve simultaneously. Achieving the first two goals expands the
cost of the program dramatically, as both the sizes of payments and the number
of people eligible for payments increase. On the other hand, holding down costs
of the welfare program requires either an income guarantee so low that it may
not provide a minimum standard of living, or a tax rate high enough to remove
any incentive on the part of recipients to find work.
This dilemma has bedeviled policymakers for over two decades as they have
wrestled with the problem of welfare reform. The reform efforts of the Nixon,
Ford, and Carter administrations centered on replacing AFDC and similar pro-
grams with a federal income maintenance program known as a negative income
tax plan.” The negative income tax plan was similar to existing welfare pro-
grams in that it had a basic guarantee level, an implicit tax rate on earned in-
come, and a break-even point. The effect of the negative income tax plan on
labor supply is thus directly analogous to that of AFDC as analyzed in Figure
2.12. What made the negative income tax plan different, however, was that it
was to be noncategorical (eligibility was open to anyone regardless of family
situation as long as the “need” criteria were met); it was to establish a common
income floor across all states; and, finally, the program was designed to encour-
age work incentives by having a relatively low tax rate. A common proposal, for
example, was for t = .5.
The negative income tax plan was considered in Congress but never passed
due to concerns over program costs and the effect of the uniform income floor on
low-wage states. Even with the existing welfare programs, however, public ex-
penditures on welfare have escalated rapidly, as shown in Table 2.1. Conse-
quently, when the Reagan administration came to office in 1981, one of its
major goals was to reduce federal spending for AFDC and other income transfer
programs.
Welfare Reform under the Reagan Administration The changes in AFDC
under the Reagan administration represented a fundamental divergence from
the direction of welfare reform taken by previous presidents.*° The most impor-
tant change was the movement away from providing work incentives for AFDC
recipients. Recipients formerly were able to deduct the first $30 of earned in-
come per month from “taxable” income (income subject to the implicit tax rate -
t) and were allowed to keep one-third of all earnings beyond that. Under the
changes enacted in 1981, the $30 + one-third monthly earnings “disregard” was
permitted only for the first 4 months of work, after which AFDC payments were
reduced dollar for dollar with increases in earned income (in other words, after
four months, t = 100%).
*See Williams, Turnbull, and Cheit, Economic and Social Security, Chapter 16, for a detailed
description.
“James R. Storey, “Income Security,” in Palmer and Sawhill, eds., The Reagan Experiment
361-392.
The Reagan administration argued that the changes in AFDC would save
approximately $2 billion in 1981 and might even increase labor supply. To
understand this, it is helpful to consider three different groups in the AFDC
population.”
First, for adults in the AFDC program who were already working zero hours,
the increase in the tax rate from .67 to 1.00 would not cause any net change in
hours of work; they worked zero hours before the change and would continue to
work zero hours at the higher tax rate. Second, adults working only a few hours
per week would find that with a tax rate of 100 percent that they could reduce
hours of work to zero and still have the same total income. This group would be
expected to reduce hours of work under the Reagan plan. The third group is the
AFDC recipients who are near the original break-even point of the program. By
increasing the tax rate, the Reagan plan reduced the break-even point and some
adults formerly eligible for AFDC no longer would be. This group would be ex-
pected to increase hours of work since they would now have less nonlabor in-
come (G = $0) and they would no longer be subject to the implicit tax rate on
earnings. The result for all three groups considered together, the Reagan admin-
istration argued, would be a reduction in program costs and a net increase in
labor supply.
Critics of the Reagan plan advanced several reasons why the 1981 welfare re-
visions in AFDC would be detrimental.”* First, they argued that the 100 percent
tax rate on earnings would cause the remaining recipients of AFDC not to work
at all. In their view, even only a few hours of work per week are important for
individual self-esteem and economic independence. They also felt the remain-
ing AFDC recipients would have a smaller chance of ever getting off the welfare
roles. A second argument of the critics was that under the Reagan plan the re-
cipients near the break-even point would not leave the program, but would re-
duce hours of work enough (or quit work altogether) to requalify for AFDC.
(Under AFDC, the recipient is also eligible for free medical service under
Medicaid. For a single mother with dependent children, the loss of Medicaid
might overwhelm the additional income that comes from working beyond the
break-even point.) Finally, critics of the Reagan plan argued that the imposition
of higher tax rates on the poor, while reducing tax rates for higher income
groups, was inequitable and seemingly at odds with the philosophy of supply-side
economics.
What is the evidence on these issues?”” With respect to program costs, it is
estimated that the changes in AFDC under the Reagan administration resulted
7For a graphic analysis see Frank Levy, “The Labor Supply of Female Household Heads, or
AEDC Work Incentives Don’t Work Too Well,” Journal of Human Resources 14, no. 1 (Winter
1979): 76-97. Also see Robert Moffitt, “Work Incentives in Transfer Programs (Revisited): A Study
of the AFDC Program,” in Ronald Ehrenberg, ed., Research in Labor Economics, 1986, vol. 8, Part
13 (Greenwich, Conn.: JAI Press, 1986), 389-439.
28Storey, “Income Security,” in Palmer and Sawhill, eds., The Reagan Experiment.
29This discussion comes from Bawden and Palmer, “Social Policy” and Appendix C in Palmer and
Sawhill, eds., The Reagan Experiment.
By 2 Chapter 2. Hours of Work
IN « THE « NEWS
A New Effort at Welfare Reform
According to a recent news report, “Congress is if welfare mothers are to have both an incentive
considering what could be pathbreaking legisla- and an opportunity to find gainful employment,
tion to change the nature of the Aid to Families provision must be made for job training, child sup-
with Dependent Children program.” The report port, and medical care. Third, both sides agree
says that for the first time in many years there that fathers must be required to take more respon-
seems to be a consensus among both liberals and sibility for the financial support of their children.
conservatives not only about the necessity of wel- Currently, less than half of the nation’s 8.8 mil-
fare reform, but also about the direction it should lion single mothers receive any child support pay-
take. As outlined in separate bills introduced in ments. Moynihan’s bill would require that states
1987 by Senator Moynihan (D-—N.Y.), Senator keep a record of both parents’ Social Security
Downey (D-—N.Y.), and Representative Brown numbers when a child is born and that an absent
(R—Colo.), the proposed reform of AFDC con- father have child support payments automatically
tains three key elements. First, as urged by conser- deducted from his paycheck.
vatives, some type of “workfare” provision must be
included that requires able-bodied welfare recipi-
ents to seek market work as a prerequisite for fur- Source: “Welfare Reform May Finally Be in the Works,” Busi-
ther assistance. Second, as emphasized by liberals, ness Week (November 2, 1987): 108-12.
in a net reduction in outlays of 14 percent relative to what would have been

spent if pre-Reagan policies were still in place. The number of families receiving
AFDC also declined, as the Reagan administration predicted. The total reduc-
tion in the case load based on the program changes was in the range of 400,000
to 500,000 families. Another 300,000 families retained their eligibility, but suf-
fered a reduction in benefits of between $150 and $200 per month. Evidence
also suggests the “recidivism rate” (the proportion of people who left the AFDC
rolls only to return a year later) did not increase substantially after the eligibility
changes, contrary to the critics’ fears.*° On the negative side, studies find little
evidence that the changes in AFDC promoted greater work effort. A second
drawback is that while the program changes did result in a significant cost sav- )
ings, they did little to cure the factors that lead to welfare dependency in the
first place. Curing these would have required substantial expenditures for family
planning, sex education in schools, and education and training programs for low
income workers—expenditures the Reagan administration was reluctant to
make. Finally, the cost savings to the taxpayer must be weighed against the hard-
ship imposed on persons who, in many respects, are at the bottom of the eco-
nomic ladder.
*See Moffitt, “Work Incentives in Transfer Programs,” and Robert M. Hutchens, “The Effect of
the Omnibus Budget Reconciliation Act of 1981 on AFDC Recipients: A Review of Studies,” in
Ehrenberg, ed., Research in Labor Economics, 1986: 351—388.
Chapter 2. Hours of Work Ped
QUALIFICATIONS The ORES ee TIE RR eres repent developed iin the! preceding
TO THE sections is relatively simple, yet it provides a numberof important insights re-
LABOR/LEISURE garding the determinants of labor supply. In some respects the assumptions of
MODEL the labor/leisure model do not always accord with actual conditions facing indi-
vidual workers in the labor market. In particular, two important assumptions of
the model are frequently violated in the real world.
Employer The labor/leisure model emphasizes the role ofindividual choice as the major
Preferences determinan t The data presented in Figure 2.1 suggest that other
of work hours.
factors are also important. While there is a great diversity in hours of work
among individuals, the single greatest percentage of workers just happened to
“choose” 40 hours per week as their work schedule. How is it that 40 hours per
week is so popular?
The. answer has to do with employer preferences and fixed work schedules. A
common feature of the majority of jobs in the labor market is that the employers
set the hours of work, with respect to both the starting and ending times each
day and the total number of hours each week. One study found that only one in
five workers had complete discretion about hours of work, while another 25 to
30 percent had some room for choice.*' For the majority of full-time jobs, how-
ever, 40 hours per week is the standard workweek set by employers. This stan-
dard reflects the influence of such things as the necessity of paying overtime for
work over 40 hours per week and the fact that continuous process industries such
as steel or chemicals can divide the workweek into three 8-hour-per-day shifts.
Employer-set work schedules represent aninstitutional constraint on individ-
ual choice that may prevent some workers from reaching their most preferred
. This situation is illustrated in Figure
2.13. The budget constraint is AB, assuming a wage rate of W, per hour. Also
shown are two indifference curves, |, for person A, and I for person B. If free to
choose his or her optimal work schedule, person A would choose 50 hours per
week (point V), and person B would choose 30 hours per week (point X). What
if each person’s employer requires a workweek of 40 hours? This combination of
income and hours of work is shown as point Z.”” The effect of the fixed 40-hour-
i
in). Individual A is made worse off because he or she desires to work longer hours
and would willingly sacrifice 10 more hours of leisure to obtain the additional
Jonathan Dickinson, “Labor Supply of Family Members,” in James N. Morgan et al., eds., Five
Thousand American Families —Patterns of Economic Progress, vol. 1 (Ann Arbor, Mich.: Institute for
Social Research, 1974): 177-250. On the subject of fixed work schedules, also see Joseph Altonji
and Christina Paxton, “Job Characteristics and Hours of Work,” in Ehrenberg, ed., Research in Labor
Economics, 1986; and William T. Dickens and Shelly Lundberg, “Hours Restrictions and Labor Sup-
ply,” Working Paper 1638 (Cambridge, Mass.: National Bureau of Economic Research, 1985).
2With an employer-mandated work schedule of 40 hours per week, the budget constraint in Fig-
ure 2.13 degenerates from AB to two points, point Z and point A.
FIGURE 2.13 Fixed Work Schedules and Desired Hours of Work

If left free to choose, person A would maximize utility by working 50 hours (point V), and person B would maximize utility by
working 30 hours (point X). S hould the employer require a fixed work schedule of 40 hours (point Z), however, both persons suffer
a decline in utility since point Z is on a lower indifference curve for each. Note that the indifference curves I, and /, represent
the preferences of two different people and, thus, it is possible for the two curves to intersect, whereas indifference curves for
one individual never cross each other.
Income (Y)
0 100 Hours of leisure—>

100 50 40 30 OHours of work (H)*—
income. Person B is also made worse off, but for the opposite reason, since he or
she would gladly sacrifice some income to have 10 more hours of leisure.
In this situation both individuals have several courses of action open to them.
One option is to find a new occupation that allows flexibility in working hours.
A person with an office job, for example, who feels dissatisfied with the rigid
9-to-5 schedule can quit and enter a different line of work such as sales or self-
employment where there is the freedom to set one’s own hours. Many people
cannot switch to an entirely new occupation, but they do have a second op-
tion—to quit their present employer and search for another firm that does pro-
vide the desired work schedule. In certain fields with large numbers of alter-
native employers and types of work that do not require rigid production shifts
this is a viable option. Examples are dental hygienist, store clerk, bookkeeper, —
and carpenter. Flexible hours are particularly important for working women with
children, helping to explain why women are concentrated in certain occupa-
tions, such as the first three listed above.
offers a solution to rigid work schedules for many work-
ers, there are also a large number who cannot easily change employers, either
because similar jobs with the desired work schedules are not available at other
firms in the area, or because factors such as seniority or pension rights inhibit
mobility. In this case, the constraints on choice in the labor market confine
workers, . In the long run,
this situation may be alleviated, however, if the dissatisfaction with existing
work schedules is strong enough and is felt by a sufficient number of people. In
Chapter 2. Hours of Work iD
that case, a firm would find that it could increase its profits by changing from the
40-hour work schedule since the dissatisfied workers would presumably be willing
to work for somewhat lower wages in order to gain their desired hours of work.
A third option that is available at least to individual A (the underemployed
worker) is to moonlightby taking on a second job. Although the primary em-
ployer may provide only 40 hours of work per week, a 10-hour part-time job
during the weekend or before or after regular hours would allow person A to
reach the desired income/hours of work position (point V). In 1985, 5.7 million
wage and salary workers held multiple jobs; their total hours of work averaged 51
per week, composed of 37 hours on the primary job and 14 hours on the second-
ary job.
There are several drawbacks to moonlighting, however, that may limit its use
by some workers. One drawback is that the second part-time job frequently pays
less than the primary job, in effect rotating the budget constraint to the left after
SNR REGRETS, The option of moonlighting may, therefore, overcome
part of the disequilibrium caused by fixed hours of work, but it still may not allow
the worker to completely reach the most preferred income/hours of work com-
bination. A second possible problem is that the additional commuting costs and
irregular hours of the second job may make it much less preferable to the alter-
native of additional hours on the primary job.
The option of moonlighting is not relevant for individual B, whose problem is
rs. If there are not alternative
employers offering work schedules with fewer hours, is there a way for the over-
employed worker to obtain the preferred hours of work? One answer involves
deliberate tardiness or absenteeism on the worker's part. Although the employer
in Figure 2.13 mandates a 40-hour workweek, by calling in sick or coming to
work late person B can still reduce hours of work toward the desired level at
point X. Such a course of action is not free of cost for the employee, who risks
disciplinary action including loss of the job. Absenteeism also imposes clear-cut
costs on the firm due to lost production and inflated labor cost. Should the cost
of absenteeism become sufficiently large, the firm will be motivated by its desire
to maximize profits to change its work schedule to meet more closely the prefer-
ences of its employees. ~
A final option is available to workers who are either overemployed or under-
employed because of fixed work schedules. Rather than individually searching
elsewhere for a firm wi e desirable work schedule -
. Through the
threat of astrike they could induce the firm to change the work schedule toward
that preferred by the workers.
How serious a problem are fixed work schedules? Available evidence indicates
that the majority of workers are generally satisfied with their total hours of work
per week. One recent survey, for example, found that 67 percent of workers were
satisfied with their number of work hours, while 25 percent desired more work
3John F. Stinson, Jr., “Moonlighting by Women Jumped to Record Highs,” Monthly Labor Re-
view 109, no. 11 (November 1986): 22-25.
hours and 8 percent desired fewer hours.** Rigid work schedules do appear to
cause a greater amount of g
. Data reveal that workers cite inconvenient hours as a much big-
ger problem than excessive hours.** The implication is that workers want more
freedom in such things as reporting for work and rearranging their work sched-
ules to fit personal or family needs. **
To respond to these demands, companies have tried several innovations. One
is a syste m
of “flexitime” where employees have a range of hours from, say,
7:00 a.m. to 9:00 a.m. when they can report for work, and a similar range when
they can leave work. While flexitime is becoming increasingly popular, it still
covers only a small fraction of the workforce (about 12 percent of full-time work-
ers in 1985).*” Another tactic, tried by firms in the chemical and petroleum in-
dustries, is to provide workers with bi by switching
from 8-hour shifts to 12-hour shifts.** These longer shifts allow 3 to 4 days off
from work at one time, rather than the typical 2-day weekend.
Salaried A second qualification to the labor/leisure model concerns workers who are paid
Professional a lump-sum salary rather than a wage rate per hour. The workers most likely to
and Managerial be affected are in white collar occupations, particularly professional workers such
Workers as lawyers, engineers, and teachers, as well as persons in a wide range of man-
agerial occupations.
The hours of work decisions for an hourly worker and a salaried worker, re-
spectively, are illustrated in Figure 2.14.Ltis‘assumed
both individuals have the
set of
indifference
same curves, 1,, I;. The major difference between them con-
«cerns the budget constraint. The budget constraint for the hourly worker is AB,
the slope of which is equal to the marginal change in income for each additional
hour worked, that is, the wage rate W,. Income for the hourly person can thus
range from zero at zero hours of work to a maximum of $W, + 100 = Y,, if all
possible hours are worked (again assuming 100 maximum work hours per week).
The budget constraint for the salaried worker, on the other hand, is shown by
the
right-angle line CDE. Below some minimum hours of work per week, for
instance H,, the individual would be fired by the employer, and earnings would
be zero (point C). Above this minimum level of work effort, the worker receives
See Susan Shank, “Preferred Hours of Work and Corresponding Earnings,” Monthly Labor Re- .
view 109 (November 1986): 40—44.
* Janice Neipert Hedges, “Job Commitment,” in Jack Barbash et al., eds., The Work Ethic—A
Critical Analysis (Madison, Wis.: Industrial Relations Research Association, 1983), 43—60.
One piece of evidence in support of this argument comes from a study by Steven Allen, “An
Empirical Model of Work Attendance,” Review of Economics and Statistics 63, no. 1 (February 1983):
77-87. He found among a large sample of workers no relationship between the length of the work-
week and absences from work, but a strong, positive relationship between lack of flexibility in sched-
uling hours and absenteeism.
“Earl F. Mellor, “Shift Work and Flexitime: How Prevalent Are They?” Monthly Labor Review,
109, no. 11 (November 1986): 14-21.
“See Herbert Northrup, James Wilson, and Karen Rose, “The Twelve-Hour Shift in the Petro-
leum and Chemical Industries,” Industrial and Labor Relations Review 32, no. 3 (April 1979):
312-326.
Chapter 2 Hours of Work a,
FIGURE 2.14 Hours of Work for an Hourly and a Salaried Worker

The budget constraint for
the hourly worker is AB, Income (Y)
and the desired hours of
work are H, (point V).
Should this person be Yy
converted to a salaried
worker earning Ys, the
budget constraint becomes
CDE. Below H, hours of
work, the person loses the
job and income is zero;
above H,, the person
receives Ys weekly income
regardless of how many
hours are worked. Since
the marginal wage is zero,
the salaried worker
maximizes utility by
working the least number
of hours possible, H, (point Hours of leisure —>
D). Hours of work (H) <—
a even up
to the maximum possible of 100 (point E). Because income is invariant to the
ero
e.
The equilibrium level of hours of work for the hourly worker is H,, and in-
come earned is Y; (point V) where the budget constraint AB and indifference
curve I, are just tangent. If this worker were now to be converted to a salaried
worker and guaranteed the same income of Y; as he or she had been earning,
what would happen to the optimal level of hours of work? The individual would
maximize utility by selecting the point on the budget constraint that touches the
highest indifference curve possible, in this case 1. Hours of work for the salaried
Empirical evidence shows just the opposite for most salaried professional and
managerial workers, however. Instead of working fewer hours than the standard
40-hour week, many of these salaried workers put in quite long hours. A recent
survey of female business executives found, for example, that 76 percent worked
50 hours or more.” What can account for this apparent discrepancy between
theory and fact? Economists have advanced two possible explanations—one is
more sociological in nature, the other more economic.”
39See “Executive Women Find It Difficult to Balance Demand of Job, Home,” The Wall Street
Journal (October 10, 1984): 33.
“This discussion is drawn from Bevars Mabry, “Income-Leisure Analysis and the Salaried Profes-
sional,” Industrial Relations 8, no. 2 (February 1969): 162-173. The reasons why some workers are
paid on an hourly basis while others receive a salary is discussed in Chapter 6.
Differences in Tastes The first explanation for salaried professional and

managerial workers’ s is that they have a greater
“taste ” According to this
for work. view, professional and managerial jobs offer a
wide range of opportunities for creativity, self-expressio n, and ego-gratification.
, but is an activity willingly taken on by the in ividual to

obtain these psychic rewards.*! Several pieces of evidence can be marshalled in
support of this view. In one national survey, 87 percent of workers in manage-
ment positions stated that they “liked their job very much,” while only 59 per-
cent of hourly workers stated such a view.” Also of interest are the observations
of a bank economist who recently completed several years of service with an
agency of the federal government. He noted that as a result of the temporary
nature of the positions held by many upper level political appointees and the
limited opportunity for bonuses or promotions in the civil service system, there
is frequently little financial incentive to work long hours. Despite this, in his
words, “the top people work long and intensely because of a genuine liking of
their jobs. They are attracted by the public service opportunities and the excite-
ment present in the political policy making process.”®
The influence of these positive tastes for work on labor supply is illustrated in
Figure 2.15. The budget constraint is assumed to be ABC; below H, hours of
work, income is zero; above H, hours, the person is paid astraight salary of Yo.
Two different sets of indifference curves are also shown. The first is the set of
three curves labeled I,, I,, and I,. The distinctive aspect of these indifference
curves is that they have a horizontal segment up to H, hours of work. This im-
plies that work is no longer a source of disutility—the individual considers work
he object of the worker is to maxi-

mize utility, and to do so he or she seeks to reach the highest indifference curve
possible, which is I, given the budget constraint ABC. There is no unique tan-
gency point; rather, a series of tangency points extend from point B to X.The
i ; the person could work as many
as H, hours or as few as H,.
ever, would require additional compensation, since beyond that point work
; Sere
It would be possible to obtain a unique comin number of hours of work if
a stronger assumption were made about the desirability of work. Work may be so
valued by the individual relative to alternative uses of time that over some range
of hours it is actually preferred to leisure. This would give the indifference curves
a U shape, as illustrated by the broken curve I; (beyond point X, I} merges into
the upward-sloping solid line I,). The tangency point between I; and the budget
“For an elaboration of this theme, see Mark Lutz and Kenneth Lux, The Challenge of Humanistic
Economics (Menlo Park, Calif.: Benjamin/Cummings, 1979).
* Cited in “Workers Today Sacrifice Less for Job,” Jackson, Michigan, Citizen Patriot (October 14
1984): El.
* Robert Dederick, “A Primer on Surviving in Washington,” Business Economics 19, no. 5 (Oc-
tober 1984): 7.
FIGURE 2.15 Hours of Work for a Salaried Worker When Work and Leisure Are Perfect Substitutes
One explanation for why salaried workers put in long hours on the job is because they enjoy their work. If work is not a source
of disutility, then the indifference curves /,, /, and /3 are horizontal (up to a maximum hours of work of H,). Utility is maximized
where the indifference curve /, is tangent to the budget constraint ABC. Equilibrium hours of work are between H, and Hp. If
work is actually preferred to leisure up to H, hours, the indifference curve will resemble the broken line I,. In this case, the
salaried worker will choose to work exactly H, hours (point X).
,
Income (Y)
H, HH Hours of leisure —>

constraint ABC occurs at H, (point X), giving rise to long hours of work for the
salaried worker.
Difference in Reward Structures A second explanation for the long hours
worked by salaried professional and managerial workers rests on a more eco-
nomic argument that stresses the difference in reward structures facing hourly
and salaried workers. In this view, it is not love of work that induces professional
and managerial workers to put in 50 or 60 hour weeks, but rather the financial
rewards they hope to receive in the form of future salary increases and job pro-
motions. On the surface, hourly workers appear to have more incentive to work
longer hours since they generally receive overtime pay (150 percent of the usual
hourly wage) for all hours worked over 40 per week, unlike the salaried worker
who often receives zero additional compensation for each extra hour. The fal-
lacy of this view is that it neglects the long-run nature of the reward system for
salaried workers. Additional hours of work this year may not affect this year’s in-
come for the salaried worker, but there is a definite link in the reward structures
of most companies between work effort this year and future earnings and promo-
tions. The reason many people are attracted to these salaried jobs, it is argued, is
precisely because the potential for earnings growth far outdistances what an
hourly worker can hope to make even at overtime rates.
The nature of this argument is illustrated in Figure 2.16. Shown there is the
horizontal budget constraint ABC; below H, hours of work, income is zero, and
above H, hours of work, the person’s salary in the current year is Y,, regardless of
hours worked. Also shown are two representative indifference curves, I, and I,.
Based on this simple model, it would seem, as was shown in Figure 2.14, that the
salaried person's equilibrium hours of work would be the minimum amount H,
FIGURE 2.16 Hours of Work for a Salaried Worker Given Future Earnings from Work
A second explanation for why salaried workers work long hours is that of the anticipation of future economic rewards. Focusing
of H, (point
only on current year income, the budget constraint for the salaried worker is ABC, yielding equilibrium hours of work
B). If additional hours of work in the current year result in salary increases in future years, however, the budget constraint becomes
ABD. At H, hours, for example, total compensation received for this level of work effort is Yo, made up of ¥, income in the anes
year and Y, — Y; income in future years in the form of salary increases. Given the budget constraint ABD, equilibrium hours o
work are H, (point X), an amount higher than Hj.
income (Y)
Y3
H> H, Hours of leisure —>

(point B), where the indifference curve I, and the budget constraint ABC just
touch each other. In fact, the salaried person works much longer hours than
this, say H,. The key to this paradox is that while additional hours of work in
the current year will not, by assumption, result in any increase in earnings above
Y,, these extra hours of work will lead to an increment in income in future years,
Thus, if hours of work
are H,, total income received by the salaried worker over the life of the job for
these H, hours is comprised of two components: the Y, amount that is received
in the current year, and the salary increases of Y, — Y, amount received in future
years as a result of this year’s work effort.** Similarly, if all 100 hours were
worked, total income earned would rise to Y; (point D), composed of Y, current
income and Y, — Y, income received in the future.
Given the new budget constraint ABD, it is now obvious why, in this view,
salaried professional and managerial workers work long hours. The equilibrium
“Since the Y, — Y; income is received in future years, it should be discounted to express it in

present value terms. Because we have not yet introduced the concept of present value, this consid-
eration is abstracted from in the present discussion.
FIGURE 2.17 The Long-Term Decline in Hours of Work

The figure for equilibrium
hours of work in 1900 was Income
2 (
y)
53 per week (point V).
The rise in real wages
between 1900 and 1987 is
represented by the rotation
of the budget constraint
from AB to AC. Given the
indifference curve /,,
hours of work per week in
1987 fall to 40 (point X),
suggesting that the income
effect engendered by the
rise in wages outweighed
the substitution effect. The
amount of income given up
to purchase extra leisure is
Y3 aa Y.
53 40 Hours of leisure —>

in Figure 2.16 occurs at point X where the indifference curve I, and budget con-
straint ABD are just tangent, giving rise to an observed level of work of H,
hours, not H,. In this case, it is not because salaried workers prefer to work any
more than hourly workers do; rather they consciously trade off current work for
the expectation of higher future income.
Which of these competing explanations is correct? Economists would most
likely emphasize the second one; sociologists would probably tend to emphasize
the first. The empirical evidence cannot clearly show one view or the other to be
correct. The two views do show how market, institutional, and sociological
forces work together to give rise to the pattern of hours of work observed in the
labor market, and that there is considerable room for controversy regarding the
exact role and importance of each of these factors in the labor market.
~
THE TIME-SERIES Figure 2.1 shows that average hours of paid work in the United States have de-
PATTERN OF clined substantially over the 20th century, from approximately 53 hours per
HOURS OF WORK week in 1900 to slightly under 40 hours per week in 1987. The labor/leisure
model provides a consistent explanation for the cross-sectional pattern in hours
of work. Can the same be done for the time-series pattern?
In general the answer is yes, although with a few qualifications. The applica-
tion of the labor/leisure model to the time-series pattern of hours of work is rela-
tively straightforward.” Figure 2.17 shows the budget constraint facing the aver-
age working person in 1900 as AB, the slope of which reflects the real wage rate
*® See H. G. Lewis, “Hours of Work and Hours of Leisure, ” Proceedings of the Ninth Annual Meeting
(Madison, Wis.: Industrial Relations Research Association, 1956): 196-206.
earned per hour of work in 1900. Individuals in 1900 also had certain prefer-
ences regarding income versus leisure, represented by the indifference curves L,
and I,. Given the budget constraint and the indifference curves, the equilibrium
hours of work in 1900 for the average working person would be 53 hours where
the budget constraint AB is just tangent to the indifference curve I, at point V.
Between 1900 and 1987 the average level of real wages in the economy in-
creased nearly 400 percent, causing the budget constraint to rotate sharply to
the right from AB to AC. Assuming people in 1987 had the same general prefer-
ences for income versus leisure as in 1900 (admittedly a big assumption), and
thus the same set of indifference curves, the new equilibrium level of hours of
work would occur where the budget constraint AC is tangent to the indifference
curve I, at point X. Asa result of the rise in real wages, hours of work per week
fall from 53 to 40, but weekly income rises from Y, to Y,.
The increase in real wages over time will cause both an income effect and a
substitution effect, the income effect from higher wages leading to a reduction in
hours of work, and the substitution effect leading to an increase in hours of
work. Since hours of work have declined substantially in the last 80 years, the
inference is that for the average person, the income effect has dominated the
substitution effect as people have, on balance, used part of their higher income
to purchase more leisure as well as more market goods. The exact amount of
income forgone to buy additional time off from work is readily identifiable in
Figure 2.17. If workers in 1987 continued to work 53 hours per week, weekly
income would have been Y;; the difference of Y, — Y,, therefore, represents the
amount of goods and services sacrificed to obtain additional leisure. One study
estimated Y; — Y, to be 15 percent of the growth in potential income per
person.*°
This time-series conclusion dovetails with the empirical evidence found in
cross-sectional studies. Cross-sectional studies have found that the labor supply
curve for men is negatively sloped as the income effect from higher wages domi-
nates the substitution effect. Since men have historically been the largest group
in the labor force, the decline in hours of work in the last 80 years suggests that
over time the income effect for males has also been stronger than the substitu-
tion effect, a result entirely consistent with the cross-sectional evidence.
While the labor/leisure model thus provides a consistent and appealing expla-
nation for the long-term decline in hours of work, two important questions re-
main about this process. The first concerns the mechanism by which preferences
for greater leisure have actually been translated into shorter working hours; the
second issue concerns the sharp difference between the reduction of work hours
prior to World War II and that which has taken place since.
The Process of As real wages have increased over time, the labor/leisure model suggests that
Hours Reduction people have chosen to take part of this gain in income in the form of additional
leisure. How has this preference for additional leisure been realized? This ques-
“Cited in Janice Neipert Hedges and Daniel E. Taylor, “Recent Trends in Worktime: Hours Edge
Downward,” Monthly Labor Review 103, no. 3 (March 1980): 3-11.
tion again involves the institutional constraint placed on individual choice by

employers’ fixed work schedules.
At the turn of the century, the working day was frequently 9 or 10 hours; in
the steel industry a 12-hour day was standard. While workers may desire shorter
working hours, employers—then and now—typically resist a reduction in work-
ing hours because it typically raises costs as capital is less efficiently utilized,
hiring costs are increased and, frequently, the demand for shorter hours is ac-
companied by a demand for higher wages in order to maintain the same level of
weekly income. What force or forces induced employers to reduce the workweek
from 53 hours in 1900 to 40 or so in 1987?
The Role of Market Forces One answer is the market forces of demand and
supply. Economic theory suggests that employers will continue to require long
working hours as long as the monetary benefits from doing so exceed the costs.
Maintaining longer work schedules than desired by employees will, however,
raise costs for firms in several ways. First, employees who remain with the firm
will likely be less productive because of factors such as low morale, increased
fatigue, and greater absenteeism. Second, the process of labor mobility will
penalize firms with longer-than-desired work hours since workers will quit those
firms and seek work elsewhere, making it necessary for these firms to pay higher
wages to attract a sufficient supply of labor. In this way, if the outcome dictated
by institutional forces (employer preferences) diverges too far from that desired
by individual workers, market forces are set off that raise costs to employers until
they find it serves their own self-interest to reduce hours of work.
There is historical evidence that part of the secular decline in hours of work is
a result of this process. The 8-percent reduction in hours of work achieved dur-
ing the labor shortages of the World War I period is one example. Other histori-
cal examples indicate, however, that the competitive pressures caused by indi-
vidual labor mobility were not strong enough to overcome the institutional
power of employers. One such case is the steel industry, which adopted the
8-hour day in 1924 only after massive strikes, public boycotts, and the personal
intervention of President Harding.*’
The Role of Institutional Forces ~Market forces are effective in inducing em-
ployers to cut hours of work only during periods of near full employment, be-
cause only then are firms’ costs of production threatened by inability to attract
workers. During periods of significant unemployment, however, the shortage of
available work leads workers to prefer a job with longer-than-desired hours to no
job at all, which short-circuits the competitive pressure on the employer to re-
duce hours of work. In these situations, the preference of workers for shorter
hours has been expressed in a collective form through the countervailing institu-
tional force of union bargaining demands and government legislation.
A collective decision to say “no” to long hours of work (that is, to strike) may
raise costs sufficiently that the employer finds it more profitable to avoid the
47Marion Cahill, Shorter Hours (New York: Columbia University Press, 1932): 206-216.
strike and reduce hours of work. While the proportion of the work force that was
unionized remained relatively small (about 10 percent) up to the mid-1930s, his-
torical evidence suggests that unions had a significant effect in reducing hours of
work in the industries that they organized as well as among unorganized employ-
ers through the threat of unionization.** Likewise, paid holidays and vacations
were first won in unionized industries in the 1940s and then later adopted by
nonunion employers. In the postwar period, unions have continued to push for
shorter hours, primarily as a means to create additional jobs.”
A second institutional force leading to a reduction in hours of work has been
government legislation. Government legislation, as with union bargaining de-
mands, represents a collective expression of preferences for shorter hours of
work. Rather than through the strike threat, however, government legislation
induces employers to reduce hours in two ways. It can require under penalty of
law that hours of work shall not exceed a certain number. This approach was
used in 1912 in the Eight Hours Act, which required the 8-hour day in govern-
ment contract work. Legislation can also raise the cost of long work schedules
through such requirements as overtime pay. This approach was used in the Fair
Labor Standards Act (FLSA) passed in 1938. The FLSA is the most important
piece of protective labor legislation in the United States, establishing for the
first time nationwide overtime and minimum wage standards. The FLSA re-
quires that employers pay a worker covered under the act 13 times the employee’s
base wage rate for all hours worked over 40 per week.
These three factors, one economic and two institutional, have in combina-
tion brought about the long-term decline in hours of work. Most economists
would give primary emphasis to market forces as the major impetus inducing em-
ployers to reduce hours of work, although unions and government legislation
have also played a role.
The Slower Decline The second issue concerning the long-term decline in the workweek that de-
of Work Hours in serves Closer examination is that hours of work declined more slowly in the post-
the Postwar Period World War II period than earlier in the century. Between 1900 and 1940, for
example, the workweek was reduced by about 10 hours; from 1950 to 1987,
hours of work declined by only another 4 hours. It has been argued that even
this four-hour decline in the reported workweek is largely a statistical artifact
due to the growing number of part-time and female workers in the labor force, a
fact which by itself would result in a secular decline in the reported average
hours of work. As an example of this downward bias, John Owen found that
when work hours of only nonstudent males were considered, the workweek actu-
*See Paul Douglas, Real Wages in the United States, 1890-1926 (Boston: Houghton-Mifflin,
1930): 112-118; and Richard L. Rowan, “The Influence of Collective Bargaining on Hours,” in
Clyde Dankert et al., eds., Hours of Work (New York: Harper and Row, 1965): 17—35.
* Sar Levitan and Richard Belous, Shorter Hours, Shorter Weeks: Spreading the Work to Reduce Un-
employment (Baltimore: Johns Hopkins University Press, 1977).
*F, Ray Marshall, “The Influence of Legislation on Hours,” in Dankert et al
ankert et al., eds., Hours of
Work, 36-53.
ally declined little at all over the post-1945 period, from 41.6 hours in 1947 to
41.3 hours in 1977.°!
How can this divergent pattern in hours of work be reconciled? If rising real
wages led to a decrease in the workweek prior to World War II, why hasn't
this same trend continued in the postwar years? There are several possible
explanations.
The Growth of Paid Time Off According to some economists, leisure time has
continued to grow as in the past—the problem is that there is a growing gap
between scheduled or paid hours of work, and actual hours that employees are on
the job.” The primary cause of this gap is the growth in paid time off from work.
Before 1940, paid holidays and vacations were practically unknown. Over the
intervening 40 years, they have come to account for a sizable block of leisure
time. By 1986, the average American worker with 5 years of job tenure annually
received at least 10 paid holidays and 10 days of paid vacation.» Thus, while the
number of hours per week that workers are paid for has not declined much in
recent years, there has been a significant reduction in actual hours on the job.
Taking the growth of paid time off into account, one study estimated that hours
of work per year for full-time production workers declined from 2,236 in 1947 to
1,955 in 1978.
Increase in Leisure over the Life Cycle Broadening the focus to annual hours
of work still understates the true growth of leisure time. The more accurate mea-
sure is lifetime hours of work. For example, the life expectancy of men increased
from 46.3 years in 1900 to 70.0 years in 1980, an increase of 24 years. Of this
24-year gain, one study found that 6.7 years were devoted to extra work, leaving
a total of 17.0 years of additional time free for other activities.” These extra
years of free time have been used primarily to delay entrance into the labor force
to allow continued schooling and at the end of the life cycle for retirement. For
women, life expectancy increased by 29.3 years between 1900 and 1980. Unlike
men, however, women split this extra time into 23.1 additional years of market
work and 6.2 years of time free from work. These trends have resulted in two
different changes in the pattern of work and leisure. The first is that work histo-
ries of men and women are converging; in 1900 men worked 32.1 years on aver-
age, while women worked, on average, 6.3 years. In 1980, the gap had narrowed
considerably to 38.8 years and 29.4 years, respectively. The second change is
that for both men and women, total leisure time over the life cycle has grown
Mohn D. Owen, Working Hours (Lexington, Mass.: Lexington Books, 1979): 10-12.
2 Herbert Northrup and Theresa Greiss, “The Decline in Average Annual Hours Worked in the
United States, 1947-1979,” Journal of Labor Research 4, no. 2 (Spring 1983): 95-114.
Bureau of Labor Statistics, Employee Benefits in Medium and Large Firms, 1986, Bulletin 2281
(Washington, D.C.: G.P.O., 1987).
Northrup and Greiss, “The Decline in Average Annual Hours Worked in the United States,
1947-1979,” Table 1.
55Shirley J. Smith, “Revised Worklife Tables Reflect 1979-1980 Experience,” Monthly Labor Re-
view 108, no. 8 (August 1985): 23-30.
quite substantially, dwarfing the gains in leisure from a shorter workweek and
more paid time off.
Other Factors A variety of other factors have also been suggested as possible
causes of the levelling off of the workweek. According to one analyst, the De-
pression and World War II forced people to forgo a variety of consumption
needs; after the war, therefore, workers were far more interested in catching up
on purchases of consumer goods and housing than in further reductions of work
time. This catching up process also had a longer-run dynamic to it because of
the “baby boom” that followed the war. Children impose financial costs on a
family for 20 years or more; thus, the costs of raising and educating the baby-
boom children made further reductions in work hours and income unattractive
for most American households.
Two other reasons for the stability in weekly hours of work have been ad-
vanced. The first is that the increase in benefits and coverage of the Social Secu-
rity program led workers to change their life cycle pattern of labor supply by re-
tiring earlier, but working more hours during their prime earning years.’ A
second and quite different explanation is that after World War II the preferences
for income versus leisure changed among the American work force. Prior to the
1940s, preferences were such that rising wages led to a dominant income effect,
leading in turn to a decline in desired work hours. Some economists have ar-
gued, however, that the development of the modern “consumer society,” with
its emphasis on the acquisition of material goods, has led to a shift in preferences
that gives more weight to income and less to leisure.** The stability of weekly
hours in the postwar period in this view is explained by a change in the shape of
the indifference curves in Figure 2.17 such that the income and substitution
effects engendered by the continuing rise in wages just happen to offset each
other.
SUMMARY Hours of work in the economy exhibit both a distinct cross-sectional and time-
series pattern. At a point in time, there is a considerable variation in weekly
hours, ranging from the 4 percent of the work force who work less than 14 hours
a week to the 7 percent who work more than 60 hours a week. The greatest
proportion of people, however, work exactly 40 hours per week. Over time,
average work hours have declined significantly in the economy, from about 53
hours per week in 1900 to slightly under 40 hours per week in 1987.
The pattern of hours is influenced by individual choice and institutional con-
straints. The role of individual choice is explicated by the labor/leisure model,
which shows that the individual maximizes utility by choosing the hours of work
*Owen, Hours of Work.

‘Richard V. Burkhauser and John Turner, “A Time Series Analysis on Social Security and Its
mca the Market Work of Young Men,” Journal of Political Economy 86, no. 4 (August 1978):
1-715.
*Levitan and Belous, Shorter Hours, Shorter Weeks: Spreading the Work to Reduce Unemployment.
where the wage rate is equal to the marginal rate of substitution. The la-
bor/leisure model also shows that, in response to a change in the wage rate, a
person may either increase or decrease hours of work, depending on whether the
income effect or the substitution effect is larger.
The long-term decline in hours of work in the twentieth century is explicable
in terms of a dominant income effect arising from the secular rise in wages.
While individual choice is an important determinant of work hours, so too are
fixed work schedules mandated by employers. If fixed work schedules conflict too
greatly with individual preference, then market forces are set off that induce em-
ployers to change their scheduled hours. In times of less-than-full employment,
however, market forces may prove ineffective in this task, and workers will then
turn to collective action in the form of union bargaining power or government
legislation to achieve the desired change in work hours.
The concept of income and substitution effects has many applications to
issues of public policy. For example, a decrease in the income tax rate sets off a
substitution effect that pulls a person toward more work and an income effect
that pulls toward less work. Whether the Reagan tax cuts lead to more or less
labor supply in the economy depends on whether the substitution or income
effect is quantitatively larger. A second public policy example is the labor supply
impact of a welfare program such as AFDC. In this case both the income and
substitution effects work together to cause a reduction in desired hours of labor
supply.
GLOSSARY
Aid to Families with Dependent Children Fair Labor Standards Act (FLSA) — Passed in
(AFDC) — A welfare program begun in 1935 1938, this law sets minimum wage and overtime
that provides income to poor families with chil- standards for employers covered by the act.
dren who are in need because of the death, in- Implicit tax rate (t)—-The amount by which a
capacitation, or absence of one of the parents. welfare payment is reduced for each additional
Backward-bending supply curve — A labor sup- dollar of earnings.
ply curve that has a positively sloped segment at Income effect —— The change in hours of work re-
low wage rates and a negatively sloped segment sulting from a change in income, holding the
at high wage rates. wage rate constant.
Break-even point (BE )— The level of earned in- Income guarantee (G) — The amount of welfare
come for a family at which it no longer qualifies payment received by a family if it has zero earn-
for a welfare payment. ings from work.
Budget constraint—The line that shows all Income maintenance program— An income
the combinations of hours of work and income transfer program which bases eligibility and the
available to a person, given the wage rate and size of payments on financial need.
amount of nonlabor income. Indifference curve —A curve that plots all the
Cross-sectional data— Data collected at a point combinations of income and leisure that yield
in time for a number of separate individuals. the same level of utility.
Marginal rate of substitution (MRS) — The rate payments on loss of income from an identifiable
at which a person is psychologically willing to problem and the amount of past contributions
trade leisure for income. into the program.
Negative income tax plan— An income mainte- Substitution effect— The change in hours of
nance program considered by the Nixon, Ford, work resulting from a change in the wage rate,
and Carter administrations. The plan was non- holding income constant.
categorical, established uniform benefits stan- Supply curve of labor—The relationship be-
dards across states, and had a relatively low im- tween the wage rate and the hours of work sup-
plicit tax rate. plied to the market.
Social insurance program — An income transfer Time-series data— Data which show the change
program which bases eligibility and the size of over time in a statistical series.
REVIEW QUESTIONS
Cp Use the labor/leisure model to analyze the is independent of income, such as a property tax)
effect on hours of work of a law requiring employ- and a proportional income tax (the tax rate t is
ers to pay time-and-one-half for hours worked over constant regardless of earnings). Identify the in-
40 per week. come effect and substitution effect (if any) caused
a. First, draw a budget constraint for some- by each tax.
one earning $5.00 per hour who receives 3. If every employer in the labor market man-
$100 a week in nonlabor income. What dates a fixed work schedule, is there still room for
now happens to the budget constraint choice by individuals concerning their hours of
with the time-and-one-half requirement? work per week? What economic forces would in-
b. Given your answer to (a), draw an indif- duce employers over the long run to change fixed
ference curve so that, given the original work schedules to match more nearly the prefer-
budget constraint, the person is working ences of workers? Does this result depend at all on
40 hours per week. What now happens to how competitive the labor market is, say as mea-
hours of work given the new budget con- sured by the number of employers in the market?
straint? Draw an indifference curve to 4. Use the labor/leisure model to illustrate the
show this. Have hours of work increased impact of increasing the implicit tax rate t in the
or decreased as a result of the overtime AFDC program, from t = .67, to t = 1.00 on the
law? Explain why. labor supply decision of the following three groups:
c. In words, define the income and substitu- (a) those people working zero hours; (b) those
tion effects. Next, use the graph to iden- people who are receiving some benefits, but who
tify the income and substitution effects. are also working in the market a few hours per
Which is stronger? week; and (c) those people who are receiving some
d. If the wage had been increased for all benefits from the program, but who are also earn-
hours worked rather than just for those ing a sufficient amount to put them close to the
after 40 per week, would the change in eak-even point.
hours of work be any different? Show this Use the labor/leisure model to explain the
in the graph and explain why or why not. ong-term decline in hours of work per week dur-
2. Use the labor/leisure model to analyze the ing the 20th century. What role, if any, did in-
effect on labor supply of alump-sum tax (a tax that stitutional forces have in this process?
ADDITIONAL READINGS
Barbash, Jack, et al. The Work Ethic—A Critical ply and recent empirical findings. It contains a dis-
Analysis. Madison, Wis.: Industrial Relations cussion of the importance of institutional constraints
Research Association, 1983. An excellent series on workers’ voluntary choices concerning hours of
of articles on the changing role of work in society. work. Also see the other papers in this volume.
Provides a review of historical trends in hours of Kniesner, Thomas J. “The Full-Time Work Week
work, the effect of social programs and attitudinal in the United States, 1900-1970.” Industrial
changes on desired work hours, and a discussion of and Labor Relations Review 30, no. 1 (October
the future of the work ethic. 1976): 74-84. Reviews the trend over the 1900 to
Hart, Robert. Shorter Working Time: A Dilemma 1970 period in weekly hours and attempts to explain
for Collective Bargaining. Paris: OECD, 1984. why hours of work have levelled off in the post—
Examines the attitudes and policies of unions with World War II period. Investment in more years of
regard to worksharing, paid time off, and shorter education by workers is identified as an important
work hours. factor.
Hausman, Jerry A. “Labor Supply.” In How Taxes Pencavel, John. “Labor Supply of Men: A Sur-
Affect Economic Behavior, edited by Henry J. vey.” In Orley Ashenfelter and Richard Layard,
Aaron and Joseph Peckman. Washington, D.C.: eds. Handbook of Labor Economics. Amsterdam:
The Brookings Institution, 1981: 27—92. A so- North-Holland, 1986: 3-102. Provides an ex-
phisticated econometric study of the effect of income haustive review of the theory of labor supply and the
taxes on labor supply. Finds that replacing the pro- empirical research on the determinants of the labor
gressive income tax with a proportional income tax supply of men. The labor supply of women is exam-
would substantially increase labor supply. ined in a companion paper by Mark Killingsworth
F. Thomas Juster. “Preferences for Work and- and James Heckman.
Leisure.” In F. Thomas Juster and Frank P. Perlman, Richard. Labor Theory. New York: John
Stafford, eds. Time, Goods, and Well-Being. Ann Wiley and Sons, 1969. A thorough theoretical re-
Arbor, Mich.: Survey Research Center, 1985: view of the labor/leisure model and various exten-
333-351. Examines people’s feelings about work sions of it to issues such as overtime hours and
and finds that work yields higher levels of intrinsic moonlighting.
satisfaction than most leisure time activities. Robins, Philip K., et al. A Guaranteed Annual In-
Kalachek, Edward. “Workers and the Hours Deci- come: Evidence from a Social Experiment. New
sion.” In Robert L. Clark, ed. Work Time and York: Academic Press, 1980. Provides a thorough
Employment. Washington, D.C.: National Com- discussion of the Seattle and Denver income mainte-
mission for Manpower Policy, 1978: 175-197. nance experiments and a review of the estimated la-
A very readable summary of the theory of labor sup- bor supply response to the programs.
APPENDIX 2A
Estimating a Labor Supply Curve
with Linear Regression
Chapter 2 gives a brief, heuristic explanation of how economists estimate a labor
supply curve from cross-sectional data. This appendix provides a deeper look at
this subject, albeit at a very nontechnical level.
A supply curve of labor shows the relationship between the wage rate and
hours of work, holding other factors constant. To estimate such a supply curve,
economists typically use cross-sectional data obtained from large surveys of sev-
eral thousand people. These surveys provide a wide range of information on
each person, such as hours of work per year, wage rate, annual income, spouse's
income, race, sex, years of schooling, occupation, and so on. One approach to
estimating the labor supply curve is to plot in a scatter diagram the combination
of annual hours of work H and wage rate W for each person in the survey. This is
illustrated by the hypothetical scatter of points in graph (a) of Figure 2A.1.' To
derive the supply curve of labor, the next step is to pass a straight line through
this scatter of points so that it fits the data as closely as possible.
Given that H is the dependent variable in the labor supply function and W is
the independent variable, the equation for such a straight line can be written as
H=a+ bW. (2A.1)
The coefficient a is the intercept term; it shows that when W = 0, H = a. The
coefficient b is the slope term; it shows the change in H for a one-unit change in
W; that is, b = AH/AW. The survey data provide the values for H and W; the
job of the researcher is to estimate the values of the coefficients a and b. Let a
and 6 stand for these estimates. With the estimates d and 6, it is possible to draw
a straight line through the scatter of points, such as the one illustrated in graph
(a). The equation for this line is
H=a+ bw, (2A.2)

where H is the predicted value of H, given the actual value of W. Thus, as
shown in graph (a), if W = W,, the actual value of H is H,, and the predicted
value given by the labor supply function is H,.
The statistical technique used by economists to obtain the coefficients @ and 6
is known as linear regression. Linear regression calculates the @ and 6 that result
in the smallest sum of the squared values of the vertical distance between H and
H.’* This can be expressed as
‘For expositional convenience, H is measured on the vertical axis and W on the horizontal axis
rather than the reverse, as is usually done in drawing a labor supply curve (as in Figure 2.9).
*A detailed description of linear regression is provided in Damodar Gujarati, Basic Econometrics
(New York: McGraw-Hill, 1978). A concise and less technical description of linear regression that is
well suited for classroom use is David L. Sjoquist, Larry D. Schroeder, and Paula E. Stephan, Under-
standing Regression Analysis: An Introductory Guide (Beverly Hills, Calif.: Sage University Papers
1986).
Appendix 2A Estimating a Labor Supply Curve with Linear Regression 91
FIGURE 2A.1 Estimating a Labor Supply Curve with Linear Regression

sani
- ©.
- ; 6 2 . : (6)
$2. $4 $6 $8 $10 $12 Ww
N
= ye > (ah, Gi MOY, (2A.3)
where N is the number of observations. The deviations between H and H are
squared; otherwise, distances above the line would be canceled by distances be-
low the line. Minimizing the sum of squared deviations results in a line that
passes through the scatter of points as closely as possible.
The regression line estimated from Equation 2A.3 is a bivariate regression;
that is, it is a relationship between only the two variables H and W. Economic
theory predicts, however, that an individual’s desired hours of work are influ-
enced by other factors besides the wage rate. One important factor, for example,
is the person’s level of nonlabor income Y. Since it is of interest to know the
effect of changes in both W and Y on hours of work, economists typically esti-
mate labor supply curves with an expanded regression technique known as mul-
tiple regression. Multiple regression estimates the linear relationship between the
dependent variable and a group of independent variables. Thus, if W and Y are
the only two independent variables, the labor supply function to be estimated is
H =a+t bW + cy, (2A.4)
where the coefficient c shows the effect on H of a one unit change in Y, holding
the value of W constant (c = AH/AY). Multiple regression estimates the co-
efficients a, 6, and é that minimize the sum of squared deviations between H
and H:
iMz (H, — H,)? = (Feige cy,)2, (2A.5)

1 i Mz
1
An example of such a regression line might be

H = 1,500 + 50W — .1Y. (2A.6)
92 Appendix 2A Estimating a Labor Supply Curve with Linear Regression
This regression line is shown in graph (b) of Figure 2A.1 as H(W),. If W = 0

and Y = 0, annual hours of work would be 1,500 (point A). Keeping Y = 0,
each dollar increase in the wage W would cause annual hours of work to increase
by50 hours. Thus, if W = $6.00 and Y = 0, predicted hours of work would be H
= 1,500 + 50 (6.00) — .1(0)= 1,800 (point B). An increase in the wage to
$8.00 would cause hours of work to increase further to 1,900 (point C). What
happens to annual hours of work if nonlabor income Y is no longer $0 but
$2,000? From Equation 2A.6, H = 1,500 + 50(8.00) — .1(2,000) = 1,700
(point D), yielding a new regression line H(W),. A change in the wage rate
causes a movement along the labor supply function, while a change in nonlabor
income causes a shift in the labor supply function.
There are a number of other factors besides the wage rate and level of non-
labor income that affect hours of work and, accordingly, should be included in
the regression equation. Let the letter Z stand for this list or vector of variables.
The complete labor supply function to be estimated, therefore, can be repre-
sented as
H=a+bW+cY+ dZ, (2A.7)
where d is the vector of regression coefficients that correspond to the variables

in Z.
The Standard Error of the Regression Coefficient Once estimates of 6 and

the other regression coefficients are obtained, there is next a question about
their reliability. The reliability of the regression coefficient b is measured by the
standard error of the regression coefficient G,,. The standard error provides an idea
of the precision with which the regression coefficient 6 is able to estimate the
true value of b. The following concrete example may make this concept clearer.
In Equation 2A.6 it is assumed that 6 = 50. If &, = 3.0, there is (roughly
speaking) about a 68 percent chance that the true value of b will be in the inter-
val b + 6, that is, within 47 and 53. Likewise, there is about a 95 percent
chance that b will be in the interval 6 + 26, (44 to 56). This range is fairly
narrow, suggesting that 6 is a very reliable estimate of b. If, on the other hand,
6, = 75, this implies that there is about a 95 percent chance that the interval
—100 to 200 contains the true value of b. In this case 6 is a very unreliable
estimate of b, so unreliable, in fact, that it is quite possible that b = 0 (implying
no relationship between W and H).
A common way to judge whether 6 is sufficiently reliable that it can be as-
sumed that b # 0 is to calculate a t-ratio. The t-ratio is measured by dividing the
regression coefficient by its standard error. For 6, the t-ratio is t, = 6/é,. Asa
handy rule of thumb, whenever the t-ratio equals two or more it can be con-
cluded with a 95 percent degree of confidence that 6 is “statistically significant,”
meaning that the true value of b is indeed different from zero. Thus, in the ex-
ample above, for b = 50 and &, = 3.0, t, = 16.6; if &, = 75, however, t, = .66.
In the latter case, although 6 = 50, the t-ratio indicates that this estimate is
sufficiently unreliable that we cannot reject the possibility that b = 0. Finally,
although the discussion has been in terms of the regression coefficient 6, it is
possible to calculate a standard error and t-ratio for each of the other regression
coefficients estimated using Equation 2A.7.
The Coefficient of Determination Another statistic frequently encountered

in regression analysis is the coefficient of determination, R?. It measures the pro-
portion of the overall variation in the dependent variable that is explained by
the regression equation. The R? can range in value from 0 (the regression ex-
plains none of the variation in H) to 1.0 (all the variation is explained). The
coefficient of determination is often used as a measure of the explanatory power
or “goodness of fit” of the regression. While a low R? means that much of the
variation in the dependent variable is unaccounted for by the independent vari-
ables in the model, this does not mean that the regression equation is funda-
mentally flawed. Particularly in studies using cross-sectional data, much varia-
tion in behavior is random or a result of factors incapable of being measured. In
this case, the R? may be quite low, although the estimated regression coefficients
on the independent variables may all be statistically significant.
Calculating Labor One of the major objectives of estimating a labor supply function is to determine
Supply Elasticities the size of the income and substitution effects. As given in Equations 2.2 and
2.3, the substitution effect is AH/AW|Y, and the income effect is AH/AY|W.
These expressions relate the absolute change in H to the absolute change in W
or Y. For many purposes, it is more convenient to express these in percentage
terms, or what are called elasticity estimates. Three labor supply elasticities are
frequently reported in empirical work.’
The first is the gross or uncompensated wage elasticity of labor supply ey, w. It is
defined as:
dH
A she lel Ne
enw = yr = oy 7). (2A.8)
W
The uncompensated wage elasticity is the percentage change in H for a given
percentage change in W. The change may be either positive or negative, de-
pending on whether the substitution effect is less than or greater than the in-
come effect. The elasticity e,, yw can be calculated as 6(W/H) where 6 is the
estimated regression coefficient from Equation 2A.7 (recall that b = @H/aW),
and H and W are the sample means of the variables H and W.
The second elasticity is the compensated wage elasticity e;; wy; it measures the
percentage change in H for a given percentage change in W, holding income
constant. This elasticity measures the size of the substitution effect in labor sup-
ply. It is defined as:
aoe _ yw 2. (2A.9)
See Mark Killingsworth, Labor Supply (Cambridge, England: Cambridge University Press,
1983), Chapter 3.
94 Appendix 2A Estimating a Labor Supply Curve with Linear Regression
This elasticity can be obtained from the estimated regression coefficients by cal-
culating 6(W/H) — ¢cW.
The third elasticity is the income elasticity ey, yw- It measures the percentage
change in H for a given percentage change in Y, holding the wage constant.
This elasticity, therefore, measures the income effect in labor supply. It is de-
fined as:
Olay
en, ylw — ces : (2A.10)
Based on the regression coefficients obtained from estimating Equation 2A.7,

the income elasticity can be calculated as ¢(Y/H) where Y is the sample mean of
the income variable.
AN EXAMPLE OF A To illustrate the concepts discussed above, an actual example of a labor supply
LABOR SUPPLY function estimated with linear regression is examined. The study was done by
FUNCTION Jane Leuthold. The dependent variable was the annual desired hours of work in
1969 of white married women with husband present. The independent variables
are briefly described below and are listed in Table 2A.1. Table 2A.1 also reports
the estimated regression coefficients, t-ratios, and the coefficient of determina-
tion.* The data used by Leuthold to estimate the regression came from the Na-
tional Longitudinal Survey (NLS). The number of women in her sample was
1,543.
In estimating a labor supply function, the first task is to decide what indepen-
dent variables should be included in the regression. Economic theory suggests
what these variables might be. Based on the labor/leisure model, differences in
annual hours of work among married women would be expected to be related to
the after-tax wage, the family’s nonlabor income, and the wife’s preferences for
work versus leisure. As discussed in Chapter 3, the husband’s earnings, the num-
ber of children in the family, the wife’s level of education, and the husband’s
attitudes toward his wife working, among other things, should also have an influ-
ence on the desired hours of work of married women. Table 2A.1 shows that
Leuthold has included variables for all of these factors in her regression model,
although some (particularly the attitude variables) do not totally capture their
theoretical counterpart in the theory.
The interpretation of the regression results in Table 2A.1 is relatively
straightforward. The regression coefficient on the wife’s wage rate, for example,
is 104.97. This means that the desired hours of work of the married women in
the sample increased, on average, by about 105 hours per year for each dollar
increase in the after-tax wage rate. This increase would indicate that for these
workers the substitution effect outweighed the income effect, making the labor
supply curve upward sloping. The uncompensated wage elasticity e}; y, there-
fore, would be positive. (Leuthold did not report the sample means H and W, so
the exact value of e,; y and the other labor supply elasticities cannot be calcu-
*To save space, not all of the independent variables included in the regression are reported in
Table 2A.1. Some of the variable names have also been altered slightly.
TABLE 2A.1 An Example of a Labor Supply Function Estimated with Linear Regression
aa aaa
Dependent Wife’s Husband's Edu- UES
Variable Constant Wage Rate Earnings Age cation A(1) A(2) C(1) C(2) R?
re
Wife's hours 1286 104.97 —.026 1.20 69 —19.47 266.06 —118.64 —110.61 .383
(4.67) (3.70) (—3.80) (.24) (.08) (—.40) (6.94) (—3.04) (—6.14)
Note: t-ratios are in parentheses.
Variable Description
Hours: Wife's annual desired hours of work, calculated as usual work hours per week times weeks worked per year plus weeks
looking for work.
Wife's wage rate: After-tax real average hourly earnings of wife.
Husband's earnings: Husband's previous year after-tax real annual earnings.
Age: Wife's age in years.
Education: Years of school completed by wife.
A(1): Attitude variable. Coded = 1 if “respondent felt it was all right for a woman to work if she desired and her husband agrees.”
Coded 0 otherwise.
A(2): Attitude variable. Coded = 1 if “the respondent's husband favored his wife's working.” Coded 0 otherwise.
C(1): Number of children less than 6 years of age.
C(2): Number of children in age group 6 to 13.
source: Jane Leuthold, “The Effect of Taxation on the Hours Worked by Married Women,” /ndustrial and Labor Relations Review 31, no. 4 (July 1978):
520-526.
lated.) Since the t-ratio is 3.70, the regression coefficient is statistically signifi-
cant from zero.
The second variable of interest is the husband's after-tax earnings. The coeff-
cient is —.026, indicating for each additional $1,000 of husband’s earnings, the
wife’s desired hours of work fall by 26 per year. Since additional earnings by the
husband increase the income available to the wife but do not change her wage
rate, this should lead to a pure, negative income effect on the wife’s desired labor
supply. The negative sign on the regression coefficient, and the fact that the
coefficient is statistically significant, supports this prediction of the theory.
Several other aspects of the regression results also deserve mention. Some of
the regression coefficients have t-ratios less than two and, thus, are not statis-
tically significant. The regression coefficient on the age variable, for example, is
1.20, meaning that desired hours of work increase by an average of 1.2 per year
with each additional year of age. This estimate is so imprecise, however, that
the true value may actually be zero. A second feature of interest is the regression
coefficients on the attitudinal variables A(1) and A(2). These are dummy vari-
ables that take on only one of two possible values: 0 or 1. The second attitudinal
variable A(2) is “respondent’s husband favored his wife’s working.” If the answer
was “no,” the variable A(2) was given a value of 0; if the answer was “yes,” it
was given a value of 1. The regression coefficient on this variable is approxi-
mately 266. What this means is that women who answered yes to this question
worked, on average, 266 more hours per year than women who answered no—
for ano answer, the effect of this variable on labor supply is 266(0) = 0, and for
a yes answer it is 266(1) = 266.
A third point of note is that the R? statistic is .383, meaning that the regres-
sion equation was able to account for 38 percent of the variation in the desired
work hours among the women in the sample.
VW o@
- 7
= = °
is. fgrveae=! &y ‘
a4 or =
we ; oe by
a ] iu ,
a a P] > J an? ee
7 es oy ~~ ane “Gai
; vt 2 Ses Cin orl 1 pepe

D Pp ps “ ras
ae ?¢ >
mel at “wa vee
sAee wo nt ee
Ta io FH
_ 4 — - : < oe
CHAPTER
Labor Force Participation
A second dimension of labor supply is labor force participation. Not only do

individuals have to make a choice of how many hours to work, they make a
simultaneous decision of whether to work at all. As with hours of work in the
last chapter, this chapter first highlights recent trends in labor force participa-
tion, then examines theoretical explanations for these trends, and finally turns
to a review of empirical evidence on labor force participation.
DEFINITION AND Before examining the trends in labor force participation, it is necessary to under-
MEASUREMENT OF stand some of the concepts to be discussed in this and later chapters and how
THE LABOR FORCE labor force data are collected.
Prior to and during most of the Depression, the federal government did not
regularly and systematically collect data on employment, unemployment, and
other vital economic statistics. Without these data it was difficult for policy-
makers to gauge either the extent of economic hardship or the direction of
change in the economy. As a consequence of this difficulty, Congress authorized
a joint program administered by the Census Bureau and Bureau of Labor Statis-
tics (BLS) to collect and analyze labor force statistics. A major source of such
data is the Population
Current Survey (CPS), commonly referred to as the
“household” survey. The CPS is conducted each month to obtain comprehen-
sive data on the labor force, the employed, and the unemployed. The informa-
tion is collected by a special group of over 1,000 highly trained interviewers who
sample about 60,000 households across all 50 states.' Even though only 1 out
of every 1,300 American households is interviewed, the figures obtained from
the CPS on national employment and unemployment are quite accurate. The
chances are 9 out of 10, for example, that the unemployment rate calculated
from the survey will be within one-tenth of one percentage point of the true rate.
Table 3.1 shows several key measures of labor force activity collected from the
CPS for various years between 1950 and 1987. The basic purpose of the labor
A detailed discussion of the sample design, interview process, and accuracy of the CPS is given
in National Commission on Employment and Unemployment Statistics, Counting the Labor Force
(Washington, D.C.: G.P.O., 1979). Also see the “Explanatory Notes” in any issue of Employment
and Earnings.
97
98 Chapter 3 Labor Force Participation
TABLE 3.1 Employment Status of the Noninstitutional Population, Selected Years, 1950-1987
Labor Force
Employed Unemployed
Civilian Barcent
Noninsti- Percent Resident Nonagri- of Not in

tutional of Popu- Armed Agri- cultural Labor Labor
Population Number lation Total Forces Total culture Indusiries Number Force Force
1950 106,164 63,377 og, 60,087 1,169 58,918 7,160 51,758 3,288 5.2 42,187
1960 119,106 71,489 60.0 67,639 1,861 65,778 5,458 60,318 3,852 5.4 47,617
1970 139,203 84,889 61.0 80,796 2,118 78,678 3,463 75,215 4,093 48 54,315
1980 169,349 108,544 64.1 100,907 1,604 99,303 3,364 95,938 7,637 7.0 60,806
1981 171,775 110,315 64.2 102,042 1,645 100,397 3,368 97,030 8,273 75 61,460
1982 173,939 111,872 64.3 101,194 1,668 99,526 3,401 96,125 10,678 9.5 62,067
1983 175,891 113,226 64.4 102,510 1,676 100,834 3,383 97,450 10,717 gip 62,665
1984 178,080 115,241 64.7 106,702 1,697 105,005 3,321 101,685 8,539 74 62,839
1985 179,912 117,167 65.1 108,856 1,706 107,150 3,179 103,971 8,312 71 62,744
1986 182,293 119,540 65.6 111,303 1,706 109,597 3,163 106,434 8,237 6.9 62,752
1987 184,490 121,602 65.9 114,177 1,737 112,440 3,208 109,232 7,425 6.1 62,888
Note: Numbers are in thousands.
source: Bureau of Labor Statistics, Employment and Earnings 5, no. 1 (January 1988): Table A-1.
force data are to measure the amount of labor input available for production in
the economy. Thus, of the total population of 243 million in 1987, people who
were less than 16 years of age or were confined to prisons, mental hospitals, and
so on were immediately excluded from consideration as potential labor force par-
ticipants, since they are presumed unable to work.’ Subtracting this group from
the total population yields the noninstitutional population, shown in the far left
column of Table 3.1. Those 184.5 million people were potentially available for
work, but not all decided to work or look for work; some chose instead to go to
school, stay at home and care for children, or retire, for example.
In 1987, 62.8 million people were included in this category.

The remaining part of the population comprises the labor force. These people
were either working or looking for work. In 1987, there were 121.6 million
people in the labor force. One of the most important labor force statistics is the
lahassforca pesticination saveit measures the people in the labor force as a per-
cent of the noninstitutionalized population. In 1987, the labor force participa- |
tion rate was 65.9 percent.
The labor force is itself comprised of two groups: the employed and unem-
ployed. The employed group includes anyone serving in the armed forces or
working at a job in the civilian economy for pay for at least one hour a week.
Also counted as employed are two other groups of people: those who work 15
hours or more a week without pay in a family business, and those who have a
paying job but are not currently at work because of illness, bad weather, a strike,
or personal reasons. The unemployed group is counted in the labor force because
*Prior to 1967 the lower age limit for inclusion in the labor force was 14.
Chapter 3 Labor Force Participation 99
even though these people are not working, seekin

they are g
work and are thus
available as labor input for the economy. The unemployment rate in 1987 was
6.1 percent; it is measured as the number of unemployed persons divided by the
The proper definition of these various labor force concepts has been the sub-
ject of much debate. In 1978, President Carter appointed a national commission
to review these definitions and the methods by which the data are collected. Of
all the issues dealt with by the commission, probably none was more contro-
versial than the measurement
of unemployment. As currently defined, to be
counted as unemployed an individual has to et ———__
i sree
ed. The
criteria for “looking for work” is defined as making some effort (for example,
answering a want ad, going for a job interview, or asking friends for job leads) to
find work in the previous four weeks. If the person is out of work and desires a
job, but has macl¢qganmeCHSRSERERATeT AT to find employment, the
CPS interviewer no longer counts the person as unemployed, but as not in the
labor force. These people are called discouraged workers, implying that they
have given up even looking for work because of poor job prospects. In 1987 there
were an average of 1.1 million discouraged workers. A number of critics of the
current BLS definition of unemployment argue that by excluding discouraged
workers, the BLS biases the unemployment rate significantly downwards. If dis-
couraged workers had been included as part of the unemployed group in 1987,
the unemployment rate would have been 6.9 percent instead of 6.1 percent.
The presidential commission, after much debate, recommended that the un-
employment rate remain as currently defined.’ A major argument against count-
ing discouraged workers as unemployed was that, to be reliable the data on
unemployment have to be based on objective differences in behavior (such as
whether a person makes an explicit effort to find a job) rather than only on the
subjective desire of wanting a job. A second consideration was that many dis-
couraged workers have only a very marginal attachment to the labor force. A
recent study, for example, found that in 1978 more than one-half of the discour-
aged workers had not looked for work even once over the preceding 12 months.*
PATTERNS IN The labor force participation rate measures the percentage of the eligible popula-
LABOR FORCE tion that is working or seeking work. It is clearly an important dimension of
PARTICIPATION labor supply to the economy.
Shown in Figure 3.1 for the years 1950 through 1987 is the labor force partici-
pation rate for three groups: men, women, and all demographic groups together.
In 1950, 59.7 percent of the noninstitutional population was in the labor force;
by 1987 this rate had increased to 64.7 percent. During the first two-thirds
of this period, the total labor force participation rate was constant, hovering
3Robert L. Stein, “National Commission's Recommendations on Labor Force Statistics,” Monthly

Labor Review 103, no. 4 (April 1980): 11-21.
4Paul O. Flaim, “Discouraged Workers: How Strong Are Their Links to the Job Market?” Monthly
Labor Review 107, no. 8 (August 1984): 8-11.
FIGURE 3.1. Trends in Labor Force Participation Rates, 1950-1987
Percent in
labor force
100
90
80
70
60
50
Female
40
30
20
10
0
1950 1955 1960 1965 1970 1975 1980 1985 1987
source: Bureau of Labor Statistics, Employment and Earnings 35, no. 1 (January 1988): Tables A—1, A-2.
around 60 percent. In the 1970s, however, the participation rate moved up-
wards fairly noticeably.
The constancy inthetotal participation rateinthe 1950s and 1960s masked
significant developments in participation for men and women, respectively. In
1950, there was a clear division of labor between the sexes—the great pro-
portion of men (86.8 percent) were in the labor force, while only one-third
(33.9 percent) of women were. This disparity has shrunk considerably in the
intervening 30 years; the male labor force participation rate had declined to
It is evident from Figure 3.1 that the most dynamic change in labor force par-
ticipation in the post-World War II era involves the dramatic increase in the -
proportion of women working in the labor market. In a 30-year period, the labor
participation rate for women increased by two-thirds with the result that today
over one of every two women are working or seeking work. This development
has had profound social and economic consequences and is one of the central
topics on which this chapter focuses.
THE DECISION As with hours of work, each individual is faced with a choice of how to allocate
TO WORK time: ime i ivities?
As a first step in analyzing the labor force participation decision, it is assume
that total discretionary time can be used for one of two mutually exclusive ac-
tivities, work or leisure. The next section, however, introduces a more realistic
N
FIGURE 3.2 A Corner Solution and Nonparticipation in the Labor Force

Given the budget constraint ABC and the shape of the indifference curves, utility is maximized by working zero hours
(point 8).
As long as the slope of the budget constraint is less than or equal to the reservation wage (the slope of the broken line), a corner
solution will occur and the person will choose not to participate in the labor force.
Income (Y)|_
100 Hours of leisure>

100 ~ 0 Hours of work (H) =<
and
leisure.
The labor/leisure model presented in Chapter 2 to analyze hours of work can
also be used to analyze the labor force participation decision, as shown in Fig-
ure 3.2. It is assumed there are 100 hours ofdiscretionary time per week that can
be allocated between work and leisure. The budget constraint ABC represents
the possible combinations of income and leisure available to the individual,
given both the wage rate of W, and Y, nonlabor income. Finally, there is a series
of indifference curves I,, I,, and I, that represent the individual’s preferences for
income versus leisure. To maximize utility, the individual should choose the
combination of income and leisure that allows him or her to reach the highest
indifference curve possible. Chapter 2 (Figure 2.5) shows this occurring where
the i , an outcome
economists call an interior solution sifice the point of tangency occurs within the
i If the indifference curves are relatively
steep, or the budget constraint is relatively flat, however, no interior solution
may be possible. This situation is illustrated in Figure 3.2. Given the budget
constraint ABC, the highest level of utility is reached at point B where the bud-
get constraint and the indifference curve I, intersect. T
. At point B, the person maximizes utility by working zero hours and
is classified as not in the labor force. This person devotes the entire amount
of discretionary time to nonmarket activities, whether school, housework, or
pure leisure.
Why did the individual depicted in Figure 3.2 decide not to participate in the
, labor force? Recall that the slope of the budget constraint reflects the value of
the wage rate—the amount of income that can be earned by “trading in” one
hour of leisure. The slope of the indifference curve, on the other hand, measures
FIGURE 3.3. The Effect of a Change in the Wage Rate on Participation

The original value of the wage rate is Wj, giving rise to the budget constraint ABC and a corner solution with the indifference
to
curve /; at point B. At point B the person is out of the labor force. An increase in the wage to W, rotates the budget constraint
ABD, and an interior solution occurs at point X on indifference curve /,. The person now chooses to participate in the labor force
and work H, hours.
Income (Y)
Y;
0 H, 100 Hours of leisure—>

100 0 Hours of work (H)<*—
the marginal rate of substitution, the subjective valuation of how much extra
income the individual would have to be given to be induced to give up one hour
of leisure. The slope of the indifference curve at the point of zero hours of work
is of particular significance and is called the reservatio n it measures the
wage;
amount of money the person would have to begiven to beinduced to work the
first hour. The value of the reservation wage is shown in Figure 3.2 by the slope
of the broken line, which is just tangent to I, at point B where hours of work are
zero. Whenever the market wage (the slope of the budget constraint) is less than
the reservation wage (the MRS at zero hours of work), hours of work will be zero
since the utility loss from giving up even one hour of leisure to participate in the
labor force would be greater than the utility gained from the income earned from '
market work.
Participation and Figure 3.3 illustrates how the decision to participate in the labor force is influ-
Changes in the enced by changes in the market wage rate. The initial equilibrium is at point B
Market Wage Rate where the indifference curve I, intersects the budget constraint ABC, and the
individual opts not to participate in the labor force. What happens to labor force
participation if the wage rate rises to W,? The rise in the wage rotates the budget
constraint from ABC to ABD, and a new equilibrium point is reached at H,
hours of work (point X) where ABD is just tangent to the indifference curve I.
Because the person is now working at a paying job, he or she is classified as in the
labor force. The reason for this switch in behavior is that the market wage (the
i ge (the MRS at zero hours
of work on I;), and the monetary gain from working that first hour exceeds the
utility loss from giving up that hour of leisure. Each extra hour of work continues
to increase utility until point X is reached, where the wage W, is equal to the
marginal rate of substitution.
hapter i shows that a rise

in the wage rate may cause a person already ri in the market to either work
more hours or fewer hours, Sean on file relative size of the income and
SESRSACEER, iherearank ae a rise in the wage cannot give rise to a

negative income effect as long as hours of work (and earnings from work) are
zero. A rise in the market wage will continue to give rise to a positive substitu-
tion effect, however, since every hour of nonmarket time now becomes more
expensive. For a person who is out of the labor force, a sufficiently large increase
in the market wage must at some point make the market wage greater than the
reservation wage, leading the individual to seek gainful employment. One pre-
diction of the labor/leisure model, therefore, is that labor force participation
should be more sensitive to changes in wage rates than should hours of work, a
prediction that is confirmed by most empirical studies.°
Participation The next issue concerns the effect of changes in nonlabor income on the deci-
and Changes sion to participate in the labor force. This case is illustrated in Figure 3.4. It is
in Nonlabor Income assumed that the person has zero nonlabor income and the market wage is W,,
giving rise to the budget constraint AB and equilibrium hours of work of H,
(point V) where the budget constraint AB is tangent to the indifference curve
I,. The person chooses to participate in the labor force since the value of the
market wage (the slope of AB) exceeds the reservation wage (the slope of the
broken line tangent to I,).
2»
er in a orl
asumptonthat
consent she The person
still chooses to participate in the labor force, however, since the market wage
(the slope of CD) exceeds the reservation wage (the slope of the broken line
tangent to I,). If nonlabor income is increased further, the person will at some
point finally decide to work zero hours in the market and will drop out of the
labor force. Given Y, nonlabor income, for ee the budget constraint is
5Michael C. Keeley, Labor Supply and Public Policy (New York: Academic Press, 1981): 24.
FIGURE 3.4 The Effect of an Increase in Nonlabor Income on Participation one

in
With zero nonlabor income, the budget constraint is AB, and hours of work are H, (point V). The person chooses to participate
the labor force since the market wage (the slope of the budget constraint) is greater than the reservation wage (the slope of the
dashed line drawn tangent to /;). With nonlabor income of ¥,, the budget constraint becomes ACD, and hours of work decline to
H, (point X). Since the market wage still exceeds the reservation wage, the person chooses to remain in the labor force. When
nonlabor income is increased to Y>, the budget constraint becomes AEF, and the person drops out of the labor force. At this point
the market wage no longer exceeds the reservation wage.
Income (Y)
Y2
Y;
H, H, OO Hours of leisure —>

The conclusion is that the ili ici

, other things equal. The reason for this
can be seen in Figure 3.4. Given the assumption that leisure is a normal good, it
is evident by comparing the respective slopes of the broken lines that the reser-
vation wage of the individual increases as successively greater amounts of non-
labor income are attained. At a sufficiently high level of nonlabor income the
reservation wage will finally equal the market wage, and the person will drop out
of the labor force.
A HOUSEHOLD The labor/leisure model of hours of work was first used to analyze the participa-
MODEL OF LABOR tion decision of men, since they were by far the largest group in the labor force.
SUPPLY Because adult males have traditionally spent the bulk of their time in continu-
ous, full-time market work with relatively little time spent on various forms of
nonmarket work (for instance, child rearing and housework), it did not seem to
stretch reality too far to couch the time allocation decision in terms of work
and leisure.
As married women became agrowing proportion ofthe labor force, however,
economists realized that the simple labor/leisure model is deficient in two impor-
tant respects. First, it ignores the family context in which the labor supply deci-
sions of husbands and wives are made. The husband and wife do not indepen-
dently decide on each person’s allocation of time to work and leisure, but rather
pool their resources and make a joint labor supply decision to maximize the util-
ity of me entire ele ee ain ATS NIN RRENEI a
The cebti ae ofthe ee ees amen is that it does not realisti-

cally represent the set of choices open to men and women with respect to their
use of time. Rather than a strict dichotomy between labor and leisure, married
women have traditionally had three alternative uses of time: market work, non-
market or home work, and leisure. Given the gradual evolution in the social
perception of family roles, this three-way allocation of time has become increas-
ingly relevant for men as well in recent years. Thus, to adequately explain trends
in labor supply it is necessary to expand the theory to take into account not only
shifts in time between labor and leisure, but also between market work and non-
market work.
Studies by Jacob Mincer and Marvin Kosters, among others, have broadened
the labor/leisure model to take into account both of these objections to the
simpler theory.° The decision-making unit in this household model is the family,
composed of a husband, wife, and possibly one or more children. Both the hus-
band and wife (and children who are old enough to — have three uses. of
) ) e.
The object of the family is to allocate the time of each family member to these
three activities to maximize the total utility of the group. The attainable com-
bination of market goods, nonmarket or home goods (such as home-cooked
meals, a clean house, and child rearing), and leisure are constrained by thewage
and he qa See TemEEE | over which the oe has discretion.

The first issue to consider is the effect of changes in nonlabor income on the
family’s allocation of time. As before, leisure is assumed to be a normal good. An
increase in the amount of nonlabor income available to the family will, there-
fore, set off a negative income effect,leading to a reduction in hours of work on
the part of all or some of the family. A large inheritance or a substantial amount
of capital or property income, for example, might lead the husband to work
fewer hours of overtime or the wife or teenaged son or daughter to withdraw from
the labor force altogether. In this three-way household model a greater demand
for leisure does not necessarily imply an equal decrease in hours of market work
as in the two-way labor/leisure model; more hours of leisure could be obtained by
reducing work in the home instead of work in the market.
The allocation of time of each family member is also influenced by the wage
rate each person can earn in the market. A change in the wage of person i in the
SJacob Mincer, “Labor Force Participation of Married Women,” in H. Gregg Lewis, ed., Aspects
of Labor Economics (Princeton, N.J.: Princeton University Press, 1962): 63—105; Marvin Kosters,
“Income and Substitution Effects in a Family Labor Supply Model,” P-3339 (Santa Monica, Calif.:
Rand Corporation, 1966).
family will affect both the individual’s own labor supply decision and the labor
supply decision of another family member, person j, through three different
and conceptually distinct channels. First, a change in the wage of person 1 (W,)
will lead to a positive substitution effect on the hours of work of person i (H;).
me
effect. Both of these are exactly the same concepts as developed in Chapter 2.
In the i
“cross-sub
labor/leisure effect. This measures the effect of
model—the stitution”
a change in the in the family.
To make the discussion more concrete, assume the family is composed of only
a husband and wife, and that initially the husband works in the market full-time
while the wife devotes her time to work in the home. Also assume that, due to
economic growth, the market wage that the wife can earn increases. This will
have several repercussions on the family’s allocation of time. With respect to the
wife’s own labor supply decision, the rise in the wage increases the opportunity
cost of each hour spent in home work and leisure, pulling the wife toward par-
ticipation in the labor force. Only when the market wage rises enough to exceed
her reservation wage, however, will she seek employment outside of the home.
When the wife was out of the labor force, the rise in her market wage led to a
fromitute
positive substitution effect that motivated her to subst nonmarket uses
of time to et
mark work. Once she is in the labor force, a further increase in the
wage sets off both an income effect and a substitution effect for the wife. For a
given number of hours of work, a rise in her wage leads to a higher amount of
income and a negative income effect pulling her toward additional hours of lei-
sure. A rise in the wage also raises the price of nonmarke t leading to a
time,
positive substitution effect that pulls her towards market work and away from
home work and leisure. The net effect of this increase in the wife’s wage on her
hours of work, therefore, is indeterminant— it depends on the relative sizes of
the income and substitution effects.
The increase in the wife’s wage also leads to a second income effect in the
household model that was not present in the simple labor/leisure model. When
the wife decides to work, not only her income increases, but also the income
Assuming his wage rate remains constant, higher earnings of the wife equiva>
are
ausing him to demand |
greater leisure and, thus, to offer less time to market work and home work.’ This
result is symmetric: if the earnings of the husband increase, the greater is the
income available to the wife, and the lower is the probability that she will
participate in the labor force. (This corresponds to the situation illustrated in
Figure 3.4.) This prediction accords with the observed tendency for married
‘The validity of this prediction is clearly attested to in remarks made recently by the president of
the American Society for Personnel Administration. He said, “When a family had a single wage
earner, that man often had to accept many kinds of unattractive work obligations. Now men are
saying no to overtime, to heavy travel, to unpleasant working conditions. But the freedom to put
lifestyle first comes only with the fact that the man has a second income, namely his wife's.” Quoted
in “Men Lose Freedom if Women Lose Ground,” The Wall Street Journal (February 2, 1987): 24.
Chapter 3. Labor Force Participation 107
women’ labor force participation rate to fall as the income of the husband be-
comes greater.
Finally, a rise in the wife’s wage also leads to a
cross-substitution
effect on her
husband's supply of labor.* As stated earlier, the cross-substitution effect is the
change in hours of work of person j given a change in the wage rate of person i,
holding income of the family constant.
AH,
Cross-substitution effect= AW, 7 =10 (3.1)
The sign of the cross-substitution effect can be either

positive or negative. Just
as a rise in the price of good X, will cause consumers to change the quantity they
demand of good Y, so too will a rise in the price of leisure for the wife cause a
change in quantity of market work offered by the husband (and vice versa), quite
independent of any change in family income. The effect of a rise in the price of
good X on the demand for good Y depends on whether the two goods are sub-
. The same is
true in the theory of labor supply. Holding income constant, a rise in the wife’s
wage leads her to work more through the substitution effect. If the wife’s and
husband's time in the market are substitutes, as she works more he will work less
(for example, he will spend more time cooking and cleaning at home), leading
to * depaveacaiscsulstinution, chege It is possible, however, that her market
work and that of her husband may be complements—as she works more in the
market so does he (for example, leisure time is only enjoyed if both share it to-
gether and, thus, as the wife reduces her amount of time devoted to leisure so
does the husband). In the case of complements, the cross-substitution
effect
is positive.
Theory cannot predict whether the cross-substitution effect will be positive or
negative. The limited amount of evidence on the subject indicates that the labor
supply of prime-age males is largely independent of the wife’s wageate (the hus-
band’s cross-substitution effect is zero), while the wife’s cross-substitution effect is
negative— attsinereasesinatheshusbandismwagesandlabonsupply (holding income
constant) causes the wife to work less.”
The household model illustrates the complex interrelationships that exist be-
tween a change in one person's wage rate and income and the resulting change in
the individual’s own hours of work and the hours of work of other family mem-
bers. The household model of labor supply also provides several key insights into
different facets of labor force participation.
The Sexual Division of Labor The household model helps explain the clear
division of labor between husband and wife in the “traditional” family of 30
or 40 years ago. It suggests that the specialization of men in market work and
women in home work reflected a rational allocation of time by the family to
8See Orley Ashenfelter and James Heckman, “The Estimation of Income and Substitution Effects
in a Model of Family Labor Supply,” Econometrica 42, no. 1 (January 1974): 73-85.
°Mark R. Killingsworth, Labor Supply (Cambridge, England: Cambridge University Press, 1983):
109.
those activities for which each person had a comparative advantage. For the
husband and wife to achieve an optimal allocation of time, each should work an
additional hour in the market as long as the wage rate exceeds the value of that
dnosumsicmuorslatonivadieialiaialciae In the context of the economic and social
environment of 1950, the great majority of married women did not participate
in the labor force because, in this view, the wage rate they could get from market
work was relatively low while the value of their time at home was much higher
due to factors such as the large number of children in the family and the un-
availability of time-saving household appliances and ready-to-eat food. Married
men, on the other hand, found it economically worthwhile to devote their time
to market work, reflecting the relatively high wage that they could earn and
their actual or perceived low productivity at child care, housework, and so on.
The Convergence in Participation Rates The household model also explains
why this traditional division of labor between the sexes has been breaking down.
As shown in Figure 3.1, over the last 30 years the labor force participation rates
of men and women have slowly converged. Why has the labor force participa-
tion rate for females risen so sharply and that for men fallen?
In terms of the household model, the key to understanding the rise in female
labor force participation is the increase in labor market opportunities for women.
Over the post-war period, n considerably,
. A major insight of the household

model is to show that this substitution effect for women takes two different
forms. A rise in the wage rate makes leisure more expensive, motivating a sub-
stitution from leisure to market work. A rise in the wage also makes time devoted
to home work more expensive, motivating a substitution from nonmarket to
market work. If the woman is already working, this rise in the wage also sets off a
negative income effect, pulling her toward more leisure and less work.
The household model predicts that the potential for substitution not only
from leisure to market work but also from home work to market work should
make the
men. Evidence from time-budget studies does reveal just such a substitution.
One study found that women who were full-time homemakers devoted approxi-
mately twice as many hours per week to household work as did wives who held
paying jobs.'° Part of the rise in female labor force participation, therefore, has —
taken the form of a replacement of home work with market work: These studies
have also found that women who choose to work also reduce their leisure time as
well, particularly for passive leisure activities such as sleeping, reading, and
watching television. One estimate is that. the working wife has about 8 hours less
leisure time per week than the full-time homemaker."!
“Donna Hodgkins Berardo, Constance Shehan, and Gerald Leslie, “A Residue of Tradition:
ns or and Spouses’ Time in Housework,” Journal of Marriage and the Family 49 (May 1987):
81-390. ;
See Martha S. Hill, “Patterns of Time Use,” in F. Thomas Juster and Frank P. Stafford, eds.,
Time, Goods, and Well-Being (Ann Arbor, Mich.: Survey Research Center, 1985): Table 7.3.
Chapter 3 Labor Force Participation 109 \
If women’s labor force participation has increased over time as a reaction to

e?
The household model also offers an explanation for this apparent inconsistency.
on the labor supply of men, the income effect pulling men toward more lei-
sure and the substitution effect pushing men toward additional work. The crucial
difference is that the size of the substitution effect for men is likely to be much
smaller than for women. Since men have traditionally devoted little time to
home work, they cannot substitute from nonmarket work to market work as can
women, and, thus, the substitution effect for men is limited to a trade-off of lei-
sure for market work.
The secular decline in male labor force participation rates reflects the fact that
for men the negative income effect from higher wages outweighed the substitu-
tion effect, a substitution effect which was considerably smaller than that for
women. Part of the decline in male labor force participation rates may also have
been due to “ened Uavsajesulbicionsellect between the rise in women’s
wages and husbands’ hours of work. As women work more, assuming that the
time of husbands and wives in home work are substitutes, part of the decline in
male labor force participation would reflect a reallocation of time for men from
market work to home work. Is this what has happened?
The bulk of the evidence points to the contrary. One study found that in 1965
married women spent an average of 4.4 hours per day on home work, compared
to only 1.6 hours for married men. By 1975, hours of home work for women had
fallen to 4.0 per day, while men’s remained at 1.6 hours.'? More recent data for
the early 1980s show some increase in the proportion of home work done by
husbands, particularly among younger families where the husband and wife are
both college-educated and employed in professional occupations." It still ap-
pears to be true today, however, that in most two-earner families, an unequal
sharing of market work, home work, and leisure remains. The best evidence
available continues to be a 1976 survey of the time allocation of married men
and women. The husbands worked significantly longer hours at market work
than did their wives, 2,211 hours per year versus 1,431. The wives, however,
spent 1,561 hours per year in home work, compared to only 487 for their hus-
bands. Total hours of work, therefore, were 2,992 for wives, but only 2,728 for
husbands.“
2Frank P. Stafford, “Women’s Use of Time Converging with Men's,” Monthly Labor Review 103,
no. 2 (December 1980): 57-59.
3See F. Thomas Juster, “A Note on Recent Changes in. Time Use,” in Juster and Stafford, eds.,
Time, Goods, and Well-Being, 313-332.
'4John W. Graham and Carole A. Green, “Estimating the Parameters of a Household Production
Function with Joint Products,” Review of Economics and Statistics 66, no. 2 (May 1984): Table 2. Also
see Graham Staines and Joseph Pleck, The Impact of Work Schedules on the Family (Ann Arbor,
Mich.: University of Michigan, 1983). Further evidence is provided in a survey of 1,870 adults con-
ducted in 1987 by The New York Times on hours of home work done by men and women. It was
found that in families where both the husband and wife work 90 percent of the women still per-
formed the shopping and 86 percent did all or most of the cooking. According to one expert, women
average 30 hours of household chores per week and men average 4 to 6. See “Women: Out of the
House but Not out of the Kitchen,” The New York Times (February 24, 1988): 1, 18.
The Becker Model The household model represents a significant generalization of the theory of
of Time Allocation choice. This generalization is carried one step further in an important study by
Gary Becker.!° According to Becker, leisure time does not yield utility in and of
ime i sed
. Whether such an
experience is watching a movie, eating a meal, or raising children, each activity
requires some combination of time and market goods.
The important insight from Becker's theory is that as the cost of time changes
so will the way in which people produce the experiences that yield utility. In
particular, as wage rates increase, individuals are motivated not only to sub-
stitute time from nonmarket uses to market work, substitute
but also to away
from activities that are relatively time-intensive, such as cooking a meal at
home, to less time-intensive but more income-intensive activities, such as din-
ing out at a restaurant. ;
Becker's model has considerable relevance to explaining the secular rise in
women’s labor force participation. The household model explained the increase
in women’s participation in terms of a substitution from home work (and possibly
leisure) to market work. This shift has been possible, however, only by a funda-
mental change in family life-style, ranging from having fewer children to use
of labor-saving appliances such as microwave ovens and dishwashers. Becker's
model can be used to analyze the reasons for this change in life-style. The rising
wage rate for women over the last 30 years gave a natural incentive to substitute
goods purchased in the market for time in producing a particular activity (for
instance, to substitute nursery school for mother’s time in child care). Thus,
Becker's model points out an additional substitution effect neglected by the house-
hold model, the substitution of market goods for the time of the wife in house-
hold
production.It is this substitution effect that has allowed married women to
substantially reduce home work and increase market work.
Empirical Evidence
Age/Participation Profiles
The household model of labor supply can explain a number of important pat-
terns or trends in the labor force participation rates of various demographic.
groups in the population. In this section, attention focuses on the cross-sectional
pattern of labor force participation rates among persons classified by sex, race,
and marital status.
Several facets of the cross-sectional pattern in labor force participation rates
are illustrated in Figure 3.5. Both diagrams present what is known as an age/par-
ticipation profile—a snapshot at a point in time of the labor force participation
° Gary Becker, “A Theory of the Allocation of Time,” Economic Journal 75 (September 1965):
493-517. Also see Reuben Gronau, “Home Production—A Survey,” in Orley Ashenfelter and
— Layard, eds., Handbook of Labor Economics, vol. 1 (New York: North-Holland, 1986),
—304.
FIGURE 3.5
Age/Participation (a) By Sex and Marital Status
Profiles by Sex, Labor force
Marital Status, and participation rate
Race, 1986 100 Married men
90
80 ¢ :
¢ Single women *
70 porta”
on
~~
s
S
60 Married women
50
20
10
16-17 18-19 20-24 25-29 30-34 35-44 45-54 55-64 65+ Age
(b) By Sex and Race

Labor force
participation rate
100 White males
90 Black males
oy oF tm on eooa.
White females ’
60
50
16-17 18-19 20-24 30-34'35-44

25-29 45-5555-64 65+ Age
source: Bureau of Labor Statistics, unpublished data.

ibsM74 Chapter 3 Labor Force Participation
rates of individuals of different ages. Graph (a), for example, shows the age/par-
ticipation profile for men and women, respectively, classified by marital status,
for the year 1986. Several features stand out in the data shown there. First, both
for married and single groups, men have a higher labor force participation than
womenIn 1986, 71.7 percent of single (never married) men 16 years of age
do
and over were in the labor force, while 65.1 percent of single women were.
Similarly, 78.4 percent of married men (spouse present) and 55.8 percent of
married women were in the labor force.
A second feature is that marriage accentuates the divergence in the participa-
tion behavior of men and women. P 5, men’s and women’s
O Dd DatlO ATE d
patio ates TO arried onsiderably ative to sing 5

A third feature of the participation rates shown in graph (a) is their pattern
over the life cycle of each group. The pattern for single men and women is
nearly identical with participation rising sharply from the teenage years to a peak
in the midlife years, followed by an accelerating decline as the individuals ap-
proach age 65.'° The life cycle pattern of married men and women, however,
differs markedly in several respects. T
over the life cycle than for single men; even in their teenage years, 94 percent of
married men participate in the labor force, and not until the 55- to 64-year-old
age group do participation rates fall much. Much the same pattern is evident ina
comparison of the profiles for single women and married women. While the par-
ticipation rate for single women increases sharply up to age 30, reaching a peak
of 84 percent, the participation rate for married women levels off abruptly at a
much lower rate of 68 percent during the years associated with childbearing and
raising a family.
Graph (b) of Figure 3.5 shows the age/participation profile in 1986 for men
and women classified by race. Several interesting features of participation stand
out. First, white men participate more in the labor force than do black men, and
this relationship is true for every age group. Second, much the opposite situation
is true for women, with
: . Finally, the life cycle
patterns of participation rates for white and black women differ significantly;
white women’s participation rates drop significantly during the prime childbear- .
ing years (ages 25 to 34) while a participation
rate for black women continues
The household model of time allocation offers several explanations for this
marked contrast in the labor force participation of these various demographic
groups.
“It is important to realize that the actual age/participation profile of persons in relatively young
cohorts in 1985 (say 20 to 24 years old) will probably look quite different from those shown in Fig-
ure 3.5, which are based on the experience of cohorts born in much earlier years. This is particularly
true for women. Evidence on this point is given in James P. Smith and Michael P. Ward, “Time-
Series Growth in the Female Labor Force,” Journal of Labor Economics 3, pt. 2 (January 1985):
559-590.
Marital Status The household model suggests that each person should allo-
cate an additional hour to home work as long as its value exceeds the monetary
value received from an hour of work at a paying job. Why, as shown in Fig-
ure 3.5, is there a much greater division of labor between market and nonmarket
work for married people than single people?’
Perhaps the most:fundamental difference between the average single, never-
married person and the family unit of husband and wife is th
dren. One important reason for the division of labor in a married couple is
because the of
‘time spent at home work relative to market work. Children, particularly pre-
schoolers, require substantial amounts of time for caring, feeding, and training
and, thus, an hour of time devoted to home work will increase the individual
and family’s well-being by more, on average, than the same hour devoted to
home work by a single person.
The second aspect of the time allocation decision for a married couple is to
choose the
work. Traditionally, the husband specialized in market work and the wife spe-
cialized in home work. The household model suggests that this decision re-
flected, first, the higher wage rate the husband could earn from an hour of
market work relative to the wife and, second, the steateieastyaleGisnesesing!
productivity
of the wife at various home tasks, particularly child care. Thus,
; ties (for example,

bearing and raising children) that enhance the value of time spent at home work
while, at the same time, allowing a degree of specialization in market and non-
market work unavailable to the single person.
This dual consideration of returns from work versus productivity in the home
also helps to explain another feature of married women’s labor force participa-
tion. As noted in Figure 3.5, the life cycle pattern of participation for married
women is much lower than that for single women and does not exhibit the peak
in participation rates in the 25 to 29 age group. This pattern is clearly associated
with the timing of the birth of children, as well as the total number of children.
The household model predicts that the labor supply of the wife should decrease
as her time becomes more valuable in the home, and it is no accident that
(white) married women’s participation falls during the prime childbearing years.
Race The household model also provides a logical explanation for the differ-
ences in labor force participation rates among whites and blacks in the popula-
tion. The two features of most importance noted in Figure 3.5 were the smaller
participation rates for black men relative to white men, but higher participation
rates for black women relative to white women.
Because of differences in quantity and quality of education, geographic loca-
tion, and discrimination, among other reasons, the average wage rate available
These issues are discussed in Sandra L. Hofferth and Kristen A. Moore, “Women's Employment
and Marriage,” in Ralph E. Smith, ed., The Subtle Revolution (Washington, D.C.: The Urban Insti-
tute, 1979): 99-124.
a4 Chapter 3 Labor Force Participation
to black men is significantly less than for white men. The household model pre-
dicts that, given their smaller return for an hour's work, black men would sub-
stitute toward relatively more nonmarket work, and possibly leisure, than white
men, leading to lower labor force participation. Particularly for black men with
low levels of schooling, this prediction seems to accord with observed behavior.
Studies of black men living in urban ghetto areas, for example, find a much
lower attachment to the labor force (as officially defined), both because they are
more likely to engage in work in the “underground” economy (such as street
hustling) and because they tend to drift into and out of jobs on a fairly fre-
quent basis. '®
Black women, on the other hand, would be motivated to have a higher labor
force participation rate for several reasons. ! One reason is that in black married
couples, given that the husband’s income is on average lower than his white
counterpart’s, there isless ofanegative income effect on the wife’s decision to
work than for a white woman. Second, it is often considered that the forces of
(GSP Trocaer ORTON ReeFa GRR ORE dekewonten than black
men, making the return from work relatively greater, increasing labor force par-
ticipation for black women relative to black men. Finally, the labor force partici-
pation of black women does not decrease during the childbearing years as it does
for white women. One explanation for this is that a proportion
greater of black
mothers are single parents (49.9 percent in 1985 versus 15.0 percent for white
mothers) and, thus, fewer black women than white women can specialize in
child rearing since fewer of their husbands are present. A second reason is the
greater proportion of extended families among blacks, resulting in better access
to relatives to help with child care. A third factor is economic necessity—given
the lower income of the husband (or absence of a husband), the black mother
has little choice but to work.
Policy Application
Labor Force Growth during Recessions
The growth in the labor force exhibits both a steady, long-term upward trend
and a short-term, cyclical movement associated with the business cycle. This.
cyclical aspect of labor force growth raises an important question that has been a
source of debate in labor economics for over 50 years—wh the size of the
ether
j j _20
The answer to this question depends on the strength of what has become known
“See, for example, Elliot Liebow, Tally’s Corner (Boston: Little, Brown and Co., 1967); and Ken
Auletta, The Underclass (New York: Vintage Books, 1983).
"See Phyllis A. Wallace, Black Women in the Labor Force (Cambridge, Mass.: MIT Press, 1980).
**Jacob Mincer, “Labor Force Participation and Unemployment: A Review of Recent Evidence,”
in Robert A. Gordon and Margaret S. Gordon, eds., Prosperity and Unemployment (New York: John
Wiley and Sons, 1966): 73-112.
as the added worker effect'and the discouraged worker effect. The issues in-
volved in this debate can be readily understood using the household model of
labor supply.
The Added Worker Effect The added worker effect leads to a temporary ine
crease in the size of the labor force during recessions and depressions. The rea-
soning is relatively simple. Consider a family where the husband works while the
wife and teenaged children are engaged solely in nonmarket activities (such as
housework and school). Next, assume the economy enters a recession and the
husband is laid off from his job. Since the unemployed are counted as part of the
labor force, as long as the husband actively searches for work, his movement
from the employed group to the unemployed group does not cause any net change
in the size of the labor force.
The household model predicts, however, that the unemployment of the hus-
band will have a definite impact on the labor SSR eainGaCete: wife and
teenaged children. The loss of the husband’s earnings results in a large decrease
in income for the entire family, setting off a pure negative income effect for the
other family members. Assuming leisure is a normal good, the reduction in fam-
ily income will lead to a decreased demand for leisure and greater supply of labor
as the wife or teenaged children seek work to supplement the family income.
From the wife’s point of view, for example, the situation is exactly the opposite
of that shown in Figure 3.4; as her nonlabor income is reduced her reservation
wage declines and at some point the wife finds it advantageous to join the labor
force. The added worker effect, therefore, represents the temporary entrance
into the labor force of “secondary” workers (the wife or teenaged children) in
response to the unemployment of the primary head of household. If the added
worker effect were the only development to take place, the labor force would
grow more rapidly during economic downturns, further exacerbating the im-
balance between labor demand and supply.”
The Discouraged Worker Effect The positive impact on labor force growth
of the added worker effect is offset by the simultaneous occurrence during re-
cessions of the discouraged worker effect. Figure 3.6 illustrates the discouraged
worker effect. Suppose the individual earns a wage of W, per hour, the budget
constraint is AB, and hours of work per week are H, (point X). Assume now
that a recession begins and the person is laid off from his or her job. Even though
hours of work temporarily fall to zero (point A), it is a disequilibrium position—
given the wage rate of W, and the preferences for income versus leisure, the
individual desires H, hours of work and will search for new employment offering
those hours. As long as the worker does continue to search for new employment,
there is again no net change in the labor force.
Does it pay to keep searching for work, however? To answer this, the unem-
ployed worker compares the financial return obtained from devoting an hour of
21 For empirical evidence on the added worker effect, see Tim Maloney, “Employment Constraints
and the Labor Supply of Married Women: A Reexamination of the Added Worker Effect,” Journal of
Human Resources 22 (Winter 1987): 51-61.
FIGURE 3.6 The Discouraged Worker Effect

(point X).
Given the wage W,, the budget constraint is AB and the person chooses to participate in the labor force, working H, hours
Should the person be unemployed, the monetary return from an hour spent in the labor force falls to z7W, where 7z is the _
probability of
probability of finding another job. Assuming a < 1.0, the budget constraint rotates to the left from AB to AC. If the
(the slope
finding another job is small enough, the person's reservation wage will exceed the expected value of the market wage
of AC) and he or she will drop out of the labor force as a discouraged worker (point A).
Income (Y)
Slope=— W,
Slope=— 1 W,
Hours of leisure ->

Hours of work (H) =<
time to job search relative to the value of time in nonmarket activities. These
activities might include, for example, painting the house, looking after the chil-
dren, or just consuming leisure. When this person had a job, the wage W, was
high enough that it exceeded the reservation wage, and he or she chose to work
in the market rather than pursue these other nonmarket activities. Now that the
person is unemployed, however, each hour spent in the labor force does not pay
W,, but some lesser amount. In statistical terms, this lesser amount is equal to
E(W), the expected value of the wage. In equation form, this expected value is
equal to:
E(W,) = 7W,, (3.2)

where 77 is the probability of finding another job paying W,. If W, = $6.00 and .
there is a 50 percent chance of obtaining such a job (7 = .5), the average return
the unemployed worker can expect from this hour in the labor force is only .5
($6.00) = $3.00. For someone without a job, the opportunity cost of not work-
ing is only E(W,), not W,. This is represented in Figure 3.6 by rotating the
budget constraint downward to AC, with a slope of —7W,. If the probability of
finding a job is sufficiently low, the expected value of the wage will fall below the
person's reservation wage, and he or she will give up the job search altogether
and drop out of the labor force, shown in Figure 3.6 by the corner solution at
point A, where the budget constraint AC and indifference curve I, intersect.
The person pictured in Figure 3.6 is a discouraged worker because he or she de-
sires to work at the wage of W, but has given up any job search because of the
low probability of finding such a job. This discouraged worker effect causes the
labor force to contract during recessions as unemployed workers temporarily (or
in some cases permanently) withdraw from the labor force. This withdrawal
causes the published unemployment rate to understate the true amount of job-
lessness during recessions.
The added worker effect leads to an expansion in the labor force during peri-
ods of recession, and the discouraged worker effect leads to a contraction of the
labor force. Is the impact of one larger than the other? Empirical research has
consistently found that the discouraged worker effect dominates in quantitative
terms the added worker effect, causing the labor force to shrink in recessions and
to grow in periods of economic recovery.”? While both occur simultaneously dur-
ing a period of recession, the added worker effect occurs only in a relatively small
percentage of families in which both the primary earner is laid off and other
family members are able to work, but are not currently in the labor force. The
discouraged worker effect, on the other hand, is likely to be more widespread,
since the probability of finding work will decrease for all the unemployed. Rising
unemployment will also cause some people such as students or housewives to
postpone entry into the labor force until better economic conditions return, fur-
ther slowing the growth of the labor force. Finally, it should be noted that dis-
couraged workers are most likely to be found in labor force groups that either
have alternative uses of their time in nonmarket pursuits that are relatively pro-
ductive (for example, housewives) or face particularly poor job prospects in the
labor market (such as less-educated black males).
CHANGES IN One of the most important labor supply developments of the last 30 years has
PARTICIPATION been the gradual convergence in the labor force participation rates of men and
OVER TIME women due to the secular decline in male participation rates and the sharp secu-
lar increase in female participation rates. The basic economic explanation for
this phenomenon was discussed previously in the development
of the household
theory of labor supply. Several additional facets of the time-series change in
male and female participation rates deserve further attention, however. The de-
cline in male participation rates is discussed first, followed by the rise in female
participation rates.
The Decline in Between 1950 and 1987 the real wage rate in the nonagricultural economy in
Male Participation t while over the same period the labor force participation
Rates rate of men declined from 87 percent to 76 percent. The secular decline in male
participation rates would suggest, therefore, that the ceCAERSMNERINSNENIE (/
from this increase in the wage rate has dominated the positive substitution
effect, leading men to cut back on market work and increase their hours of non-
work, leisure time.
2William Bowen and T. A. Finegan, The Economics of Labor Force Participation (Princeton, N.J.:
Princeton University Press, 1969).
118 Chapter 3. Labor Force Participation
FIGURE 3.7
Age/Participation Labor force
participation rate
Profile for Men, 1950
and 1987 100
90
80
70
60
16-17 18-19 20-24 25-34 45-54 55-64 654+ Age

source: Bureau of Labor Statistics, Employment and Earnings 35, no. 1 (January 1988): Table A—4; Handbook of
Labor Statistics, 1983, Table 4.
Further insights into this process are provided in Figure 3.7, which shows the
age/participation profile for men for the years 1950 and 1987. The decline in
male labor force participation over the last 30 years is largely concentrated in
three groups. The first is young men (ages 16 to 24), the second is prime-age
males in the 45 to 62 age bracket, and the third, and most important, is among
men
retirement
at age (62 plus). Among men age 65 and over, for example, the
decline in participation was quite dramatic, from 46 percent in 1950 to 16 per-
cent in 1987.
Economists have advanced several reasons to account for the long-run decline
in the participation rates of these groups of men. Three of the most important
are considered here.
The Life Cycle Allocation of Time One defect of the labor/leisure model and
the expanded household model is they are static or “one-period” models—they
do not consider the optimal timing of labor, leisure, and nonmarket activities
over the multiperiod life cycle of an individual. In recent years, therefore, econo-
mists have broadened the theory of labor supply even further by constructing
dynamic models of labor supply that are capable of predicting the optimal alloca-
tion of time over the entire adult life of an individual.’®> While these models tend
” The most accessible introduction to dynamic labor supply models is Killingsworth, Labor Sup-
ply, chapter 5.
A
to be highly mathematical and beyond consideration here, they do yield some
common sense implications about the reasons behind the decline in male par-
ticipation rates.
Why, for example, are fewer men working beyond age ‘65? Two related facts
seem to explain this in a life cycle context. The first is that the life cycle pattern
of men’s ising rapidly from midlife to
later life and then declining as skills become obsolete and on-the-job training
slackens. Because of this pattern of earnings, the opportunity cost of additional
hours of leisure in the prime work years (ages 24 to 54) is much higher compared
to later years in life. The implication is that an individual has a strong economic
incentive to Substitute work for leisure during midlife when the rewards from
work are greatest,
lowest.** The second fact of importance is the long-run rise in real wages due
is
to economic growth. The rise in wages over time causes the inverted V age/
earnings profile to shift upward over time for successive cohorts (people of the
same age group), setting off an income effect and substitution effect on lifetime
labor supply much as in the static household model.” Assuming that for men,
the income effect of this wage increase outweighs the substitution effect, when is
the optimal time to> TELS aueenel leisure? nC edic
~ OPPO
is lowest. eek retirement, eee is the ees response to an ee

demand for leisure.
Life cycle considerations also help explain the decline in the labor force par-
ticipation rate among young men in the 18 to 24 age group. For these men, the
smaller amount of time spent at work has little to do with leisure; rather, it re-
flects a ,
As compared to 1950, today’s jobs require
a much higher level of education. Since education is often a full-time activity,
the worker faces the question of when in the life cycle to schedule work and
when to schedule time out of the labor force for education.
Not unexpectedly, the optimal time to invest in human capital is at the begin-
ning of the life cycle for two important reasons.” The first is that the oppor-
tunity cost of devoting time to education is lowest when an individual is young.
Since wage rates rise asaperson ages because of experience and on-the-job train-
ing,itmakes economic sense to invest in additional human capital at a young
age when the earnings penalty is the lowest. A second consideration is that the
return on an investment in education is also greater the earlier in the life cycle it
Empirical evidence in support of this prediction is presented in an interesting study by Frank

Stafford and Greg Duncan, “The Use of Time and Technology by Households in the United States,”
in Juster and Stafford, eds., Time, Goods, and Well-Being, 245-288, particularly Figure 10.1.
25 As discussed in Killingsworth, Labor Supply, a wage change due to a shift in the age/earnings
profile gives rise to both an income effect and a substitution effect, while a wage change due to a
movement along a given profile gives rise only to a substitution effect since total lifetime remains
unchanged. An important implication is that cross-sectional estimates of income and substitution
effects may be seriously biased because they confuse these two sources of change in wages.
See Karl E. Ryder, Frank P. Stafford, and Paula E. Stephan, “Labor, Leisure, and Training over
the Life Cycle,” International Economic Review 17, no. 3 (October 1976): 651-674.
Chapter 3 Labor Force Participation
is made. The direct monetary outlay for a college education will be the same for
a person whether he or she is 25 or 45. The younger person, however, will have
making the rate
of return on the human capital investment much greater.
Social Security and Private Pension Plans A
important
second reason for
the decline in participation rates among older men is the growth in the coverage
and benefits from Social Security and private pension plans.
The Social Security program was instituted in 1935. Although the legislation
originally provided only for retirement benefits for workers at age 65, over the
years it has been broadened considerably to provide other forms of income pro-—
tection, principally in the form of disability insurance, survivors’ benefits, and
hospital and medical insurance for the elderly. Both the number of people eli-
gible for Social Security retirement benefits and the dollar amount of the bene-
fits have increased substantially in the last two decades. In 1985, 91 percent of
the population aged 65 or over received Social Security payments, compared to
79 percent in 1962. Likewise, between 1972 and 1985 the average monthly
benefit received by a retired worker increased by 47 percent in real terms.”’
Most research on the subject has found that the availability of Social Security
benefits is an important influence leading workers to retire earlier.** The reasons
for this are illustrated in Figure 3.8. The budget constraint for a worker aged 65
in the absence of a Social Security law is assumed to be AB, reflecting a wage
rate of W, and zero nonlabor iincome. Under the Social Security program, how-
ever, the budget constraint becomes ACDEB.A worker aged 65 or over now
qualifies for an annual aeyteneof $G (point C) if he or she retires. (The exact
amount depends on previous contributions into the program.) Just as with Aid
to Families with Dependent Children, there is also an implicit tax rate t on addi-
are :
tional labor earnings. The Social Been Sere ee an Cane test that re-
is of
pt1987hen maximum earnings level was $8,160.
he implicit tax rate, ore, is t = .5 and the break-even point where an
RET receives zero caren is rec nama Between
points C and D the new budget cr
constraint pos HSsameOe: asS the pie)
tax rate.
What impact does the Social Security program have on desired hours of work? —
Not surprisingly, the provision of additional nonlabor income, coupled with an
implicit tax rate on earnings from work, provides strong incentives for older
“Social Security Administration, Social Security Bulletin: Annual Statistical Bulletin, 1986 (Wash-
ington, D.C.: G.P.O., 1986).
*°See Gary Burtless and Robert A. Moffitt, “The Effect of Social Security Benefits on the Labor
Supply of the Aged,” in Henry J. Aaron and Gary Burtless, eds., Retirement and Economic Behavior
(Washington, D.C.: Brookings Institution, 1984): 135-171; Roger Gordon and Alan Blinder,
“Market Wages, Reservation Wages and Retirement Decisions,” Journal of Public Economics 14, no. 2
(October 1980): 277-308; and Gary S. Fields and Olivia S. Mitchell, Retirement, Pensions, and So-
cial Security (Cambridge, Mass.: MIT Press, 1984).
FIGURE 3.8 The Effect of Social TS) on raher Force Participation

Before Social Security, the i
budget constraint is AB, Bes “ieame
and hours of work are H, per year (Y)
(point V). With Social Se -
Security, the budget : |
constraint becomes —
ACDEB. As drawn, the
effect is to cause the 7
person to reduce hours of
work to H, (point X). It is
possible, however, that a
corner solution would
occur at point C and the
person would drop out of
the labor force.
Hours of leisure >

workers to reduce their supply of labor. Without the Social Security program,
the equilibrium hours of work is H, (point V), given by the original budget con-
straint AB and the indifference curve I,. With the program, the budget con-
straint becomes ACDEB and a new equilibrium occurs at H, ae X) where I,
and ACDEB are tangent. The impact ot gra rk
from H, to H,. Although the new eirivisseners hours of vistain ene oe 8 oc-
curs at point X on the line segment DC, there are several other possibilities.
One would be for the indifference curve I, to be tangent to the segment ED;
another would be for the new equilibrium to be a corner solution at point C. In
the latter case, the person would retire from work altogether and drop out of the
labor force. Evidence shows that about one-half of new Social Security recipi-
ents do completely withdraw from the labor force, while the other half continue
to work at a paying job, generally for fewer hours than before.”
Social Security is one important cause of earlier retirement. The growth of
employer-financed pensions is another. In 1985, 39 percent of retired workers
received benefits under some type of private or government pension plan, com-
pared to only 14 percent in 1962. Research has found that private pension plans
have a stronger impact on the retirement decision of workers than does Social
Security.*° Two factors account for this. The first is that private pension plans
generally condition the provision of benefits on a complete cessation of work at
the firm, unlike Social Security, which allows a person to work and draw benefits
29 Alan Fox, “Income Changes at and after Social Security Benefit Receipt: Evidence from the
Retirement History Study,” Social Security Bulletin 47, no. 9 (September 1984): 3-23.
. Richard V. Burkhauser and Joseph F. Quinn, “Is Mandatory Retirement Overrated? Evidence
from the 1970s,” Journal of Human Resources 18, no. 3 (Summer 1983), Table 2.
simultaneously. The second reason is that the benefit structures of most private
pension plans contain stronger incentives for early retirement than does Social
Security. Under Social Security, a person can retire at age 62 or 65 or later. The
incentive to retire at age 62 is the 3 years of additional benefits that can be
obtained; the drawback is that the monthly payment is smaller. Whether the
worker is financially ahead to retire early at age 62 or wait until age 65 depends
on two factors: the degree to which the monthly Social Security payment is
larger at age 65, and the life expectancy of the recipient. Studies that have
examined this issue reach contradictory conclusions, although the evidence in-
dicates that there is a large financial penalty to postponing retirement beyond
age 65.*!
What about private pension plans? Here too the worker must balance the
benefits of retiring earlier against the cost of a smaller monthly payment. In the
case of private pensions, the evidence is unambiguous—the total value of pen-
sion benefits declines as the worker postpones retirement. One study estimated
that a worker who postpones retirement from age 64 to age 65 sacrifices $1,200
to $1,400 of pension benefits. ** One reason why private pension plans are struc-
tured this way is that employers would rather have workers voluntarily choose
retirement rather than be forced to quit when they reach the mandatory retire-
ment age. (Under the 1978 amendments to the Age Discrimination in Employ-
ment Act, a worker can not be forced to retire before his or her 70th birthday. )
The Growth in Disability Benefits The third group of men for whom partici-
pation rates have declined noticeably is prime-age men between the ages of 45
and 62. This phenomenon is primarily concentrated among black men. The par-
ticipation rate for black males aged 45 to 54, for example, fell from 93 percent in
1954 to 84 percent in 1987, more than twice the decline for white males.
The factors responsible for this trend are not entirely understood. One pos-
sible reason is that poor employment prospects for older, less-educated black
males causes a greater proportion to drop out of the labor force as discouraged
workers. A second factor found to be important in several studies is the greater
incidence of health problems among older black males.*? A third factor identi-
fied by Donald Parsons as a major explanation is the rather substantial expansion
and liberalization of coverages and benefits in the disability insurance program of
Social Security. ** Between 1968 and 1986, for example, the number of recipients
*'See Alan S. Blinder, Roger H. Gordon, and Donald E. Wise, “Reconsidering the Work Disin-
centive Effects of Social Security,” National Tax Journal 33, no. 4 (December 1980): 431-442;
Richard V. Burkhauser and Joseph F. Quinn, “The Effect of Pension Plans on the Pattern of Life
Cycle Compensation,” in Jack E. Triplett, ed., The Measurement of Labor Cost (Chicago: University
of Chicago Press, 1983): 395-415; and Olivia S. Mitchell and Gary S. Fields, “The Economics of
Retirement Behavior,” Journal of Labor Economics 2 (January 1984): 57-83.
* Burkhauser and Quinn, Table 12.2. Also see Lawrence Kotlikoff and David Wise, “The Incen-
tive Effects of Private Pension Plans,” in Zvi Brodie, John Shoven, and David Wise, eds., Issues in
Pension Economics (Chicago: University of Chicago Press, 1987), 283-336.
LL. Wayne Plumly, “The Declining Labor Force Participation Rate of Prime-Age Males,” unpub-
lished Ph.D. dissertation (Atlanta: Georgia State University, 1983).
“Donald O. Parsons, “Racial Trends in Male Labor Force Participation,” American Economic Re-
view 70, no. 5 (December 1980): 911-920.
grew from 2.3 to 4.3 million while average benefits per month increased by 110
percent in real terms. Given that older black workers have a greater incidence
of health problems, the expansion and liberalization of the disability program
would be expected to cause the greatest decline in participation among that
group. Parsons also argues that older, less-educated black males have the most
incentive to attempt to qualify for disability benefits since the level of benefits
relative to their potential earnings from work are frequently much higher than
for white workers with better-paying jobs. His statistical analysis of labor force
participation rates for black and white workers found considerable evidence that
a signifcant portion of the relative decline in the participation rate of black
males is due to the greater availability of disability benefits. A more recent study
on this subject also found a positive link between the generosity of disability
benefits and the probability of nonparticipation in the labor force by prime-age
males, although the magnitude was far smaller than that estimated by Parsons.
The Increase The most dramatic change in labor supply in the United States has been the
in Female tremendous increase in the number of working women. At the turn of the cen-
Participation Rates tury only 5 million of the 28 million Americans in the labor force were female.
One-quarter of these women were teenagers; only 15 percent were married. As
recently as 1950, fewer than 19 million of the 64 million labor force participants
were women. Over the past 35 years, however, women, and in particular mar-
ried women, entered the labor force at an unprecedented rate. Between 1950
and 1987, for example, six out of ten additions to the labor force were women,
with the result that women now comprise 45 percent of the labor force. Equally
significant, g, in-
n.
The dynamics of change in the pattern of female labor force participation are
apparent from the age/participation profiles for women for the years 1940, 1960,
and 1987, as shown in Figure 3.9. Two features stand out. The first is that the
rise in SecGRRE EACTeoh SEO TAEGETEAA SOE REDEEMER
65, reflected in the general upward shift in the age/participation profile for 1987
relative 10 1540. Ganieua
SESlasheeinmnacieensainaiialy
tion for women in the childbearing»years. In 1940, the age/participation profile
had a single peak in the age group 20 to 24 followed by a sharp and continuous
decline that coincided with the time of marriage and childbearing. By 1960, the
age/participation profile had changed considerably, and began to resemble an
inverted W. Women still dropped out of the labor force with marriage and child-
bearing, but the difference between 1960 and 1940 was that many of the women
in 1960 were reentering the labor force once their children were older and in
school. This pattern in 1960, while revolutionary at the time, was only a mile-
post in an ongoing transformation. By 1987, the sharp dip and rebound in fe-
male participation rates had itself disappeared, repla ced
by a pattern of stable
Robert Haveman, Barbara Wolfe, and Jennifer Warlick, “Disability Transfers, Early Retire-
ment, and Retirement,” in Aaron and Burtless, eds., Retirement and Economic Behavior, 65—93.
FIGURE 3.9
Age/Participation Labor force
Profile for Women, participation rate
1940, 1960, and 1987 100
90
80
70
20
10
16-17 18-19 20-24 25-34 35-44 45-54 55-64 65+" Age
source: Bureau of Labor Statistics, Employment and Earnings 35, no. 1 (January 1988): Table A—4; and Department
of Commerce, Historical Statistics of the United States: Colonial Times to 1970, Part 1 (Washington, D.C.: G.P.0.,
1975): 132.
The household model of labor supply pointed out the important role that the
played in bringing about this secular in-
crease in labor force participation. Several other developments have, however,
also played crucial roles in this process.”
Fertility For biological and cultural reasons, childbirth and child rearing
have been the primary responsibility of the female parent. Given the tremen-
dous time demands that raising children entails, married women could not have
entered the labor force as they have in the last 2 decades without a fundamental
Nearly every empirical study has found a strong negative relationship between
wife.” This negative relationship is particularly pronounced for children of pre-

school age. A powerful influence leading to greater market participation on the
part of women in the United States, therefore, has been the significant decline
** A thorough review of the causes of the increase in female labor force participation over the 20th
century is provided in Smith and Ward, ““Time-Series Growth in the Female Labor Force.” They
conclude that 60 percent of the secular increase in female participation rates is due to the rise over
time in the real wage (and the induced decline in fertility), with other factors accounting for
the rest.
See, for example, David Shapiro and Lois B. Shaw, “Growth in the Labor Force Attachment of
Married Women: Accounting for Changes in the 1970s,” Southern Economic Journal 50, no. 2 (Oc-
tober 1983): 461-473.
e. In the mid-1950s, the fertility rate

(average number of births per woman) was 3.5, a figure which then declined
steadily, reaching 1.8 in 1986.
Several factors have been responsible for the decline in fertility rates. One is
the greater number and reliability of contraceptive devices, as well as greater
social acceptance of family planning. A second factor that has been given much
attention by economists is the growing “price” of children.** The price or cost of
a child for the family is the direct cost of food, clothing, and education, and the
SERS EAPC faM HPA ORUREAIREL » Since the real wage
that can be earned at work has increased substantially over the last 3 decades,
the relative price of time-intensive activities such as raising children has also
grown considerably, causing families to consciously cut back on the number of
children they desire. In this view, the decline in fertility and the rise in female
labor force participation reflect a simultaneous response to the ever-growing eco-
nomic rewards from market work.
A second shift in fertility patterns concerns the postponement of childbirth
on the part of many women until they are past their mid-20s. In 1970, 19 per-
cent of first births were to women 25 or older; by 1985 this figure rose to 40
percent. The impetus behind this trend is an attempt by women to accom-
modate both
careers and children. As documented in an interesting study by
Marianne Ferber and Bonnie Birnbaum, a woman in her 20s who desires both a
career and children faces a significant time “squeeze” —there is not enough time
available to do both simultaneously.*° One option is to have children first and
then seek a job. Another is to reverse the order and begin a career first, postpon-
ing children until later, even if this means dropping out of the labor force for a
period of time. Research shows that the second option is the superior one in that
it promises higher future earnings than the first.*' As women move into upper
management and professional positions, even the second option (postponing
children until several years of job experience are accumulated) is perceived by
many as prohibitively expensive in terms of career advancement. In a survey of
over 700 executive women in 1984, for example, only 36 percent under the age
of 40 had any children.”
on
Gary S. Becker, “An Economic Analysis of Fertility,” in Universities-National Bureau Commit-

tee for Economic Research, Demographic and Economic Change in Developed Countries (Princeton,
N.J.: Princeton University Press, 1960): 209-231.
In 1980, it was estimated that the direct costs of raising a child from birth through college
ranged from $58,200 to $85,000. If the mother also stayed out of the labor force until the child was
aged 15, the amount of forgone earnings was estimated to be $130,000 for a mother with an aver-
age level of education and $189,000 for a mother with a graduate degree. See Allyson Sherman
Grossman, “Working Mothers and Their Children,” Monthly Labor Review 104, no. 5 (May 1981):
49-54. A more detailed analysis is provided in Lawrence Olson, Costs of Children (Lexington,
Mass.: Lexington Books, 1983).
40Marianne Ferber and Bonnie Birnbaum, “One Job or Two Jobs: The Implications for Young
Wives,” Journal of Consumer Research 7, no. 3 (December 1980): 263-271.
41See Ellen Van Velsor and Angela M. O’Rand, “Family Life Cycle, Work Career Patterns, and
Women’s Wages at Midlife,” Journal of Marriage and Family 46, no. 2 (May 1984): 365-373.
2“Eyecutive Women Find It Difficult to Balance Demands of Job, Home,” The Wall Street Journal
(October 30, 1984): 33.
A third aspect of the fertility participation relationship that has changed dra-
matically over time is the greater propensity of women. to continue working
. While married couples are having
fewer children and having them later in life, childless couples are still a distinct
minority in the population. According to survey results, 80 percent of women
between the ages of 18 and 34 say they expect to have between one and three
children.*? A fundamental shift has occurred, however. When these women do
have children, they are less likely than before both to drop out of the labor force —
; 44
Thi is vividly illustrated by comparing the participation rate for married

women with children of different ages. It remains true today as in the past that
the younger are the children in the family, the less likely the wife is to partici-
pate in the labor force. In 1985, 64.2 percent of married women whose youngest
child was 16 were in the labor force, compared to only 49.4 percent for married
women whose youngest child was under 1 year of age. What is remarkable, how-
ever, is how many of these mothers with preschool children are remaining in the
labor force compared with before. In 1970, 30 percent of married women with
children under the age of six were in the labor force; in 1985 this proportion had
jumped to 54.3 percent. This growth in the participation rate of mothers with
preschool-age children has necessitated fundamental shifts in child care arrange-
ments.It is estimated that nearly three-fourths of all one- to five-year-olds are
now enrolled in some form of group child care program.*
Education A second factor often cited as a cause of greater labor force partic-
ipation among women is the rise over time in educational attainment.* At a
point in time there is a strong, positive relationship between years of education
completedsandeprobabiliqwofiparticipationsinetheraborforee. In 1986, the par-
ticipation rate of women with less than a high school education was 31.8 per-
cent, compared to 74.4 percent for college graduates. Since the median years of
schooling completed by women have risen steadily over time, the logical in-
ference is that there should also be a concomitant increase in the participation
rate of women.
There are several reasons for the positive correlation between education and
labor force participation. First, higher education is often undertaken as an in-
vestment in the sense that a person willingly suffers the large direct costs:(tui-
tion) and opportunity cost (forgone earnings from work) of a college education
with the anticipation that these costs will be recouped in the form of higher
*See Kristin A. Moore and Sandra L. Hofferth, “Women and Their Children,” in Smith, ed.,
The Subtle Revolution, 125-157.
“See Shapiro and Shaw, “Growth in the Labor Force Attachment of Married Women”; and Bu-
reau of Labor Statistics, Families at Work: The Jobs and the Pay, Bulletin 2209 (Washington, D.C.:
G.P.O., 1984).
* Sandra Hofferth and Deborah Phillips, “Child Care in the United States, 1970-1995,” Journal
of Marriage and the Family 49 (August 1987): 559-571.
*See Ralph E. Smith, “The Movement of Women into the Labor Force,” The Subtle Revolution,
1-29.
earnings and occupational attainment after graduation. To reap this return on

education, however, requires a sustained period of participation in the labor
force. A second reason why education is positively correlated with labor force
participation is that earnings from work progressively tise with educational at-
ii tial ts £ d aa
education may increase the probability of participation as it changes an individ-

ual’s tastes or attitudes with respect to the desirability of home work versus mar-
ket work.
Not only have increased years of education among women encouraged greater
labor force participation, so too has the-shift over time in the fields of study
pursued by women in high school and college. Twenty years ago, the majority of
women in college majored in fields such as English, history, or education, while
men were more likely to choose subjects such as engineering, business, or pre-
law. Although there is still a marked difference between men and women in the
choice of a college major, it has certainly narrowed.* In 1965, for example, only
2 percent of engineering degrees and 23 percent of business degrees were awarded
to women; in 1984 these figures had risen to 14 percent and 43 percent, respec-
tively. Young women, therefore, are increasingly pursuing fields of study in school
that prepare them for full-time jobs in the labor market.
A final aspect is the influence that the sharp rise in divorce rates has on the
motivation of women to invest in additional education. Marriage encourages
a division of labor where each spouse ializes i in which
he or she has a comparative advantage. The benefit of specialization is that the
total amount of market and nonmarket goods that can be obtained by the family
is larger than if both persons did an equal share of each activity. For instance,
in years past it was not unusual for the wife to work to put the husband through
school, with the expectation that he would then support her as she raised the
children full-time. In a world without divorce, this type of specialization works
to the benefit of both partners since each shares the other's “production.” The
greater the probability of divorce, however, the less incentive either partner has
to invest in the other because of the danger of suffering a capital loss should the
marriage fail.°° Why are more young women obtaining additional schooling and
majoring in marketable fields such as business or law? Given the probability that
of all first marriages of persons between the ages of 25 and 35 will end
in divorce, it is increasingly important for both spouses to have an independent
source of earnings as insurance against marital dissolution.
‘7 Increased years of schooling could also increase the productivity of time spent in home work and
leisure, making it less likely that a person would participate in the labor force. Shapiro and Shaw,
“Growth in the Labor Force Attachment of Married Women,” find that once the higher wage de-
rived from additional schooling is controlled for, women with more education have a lower proba-
bility of being in the labor force.
“See Arthur E. Blakemore and Stuart A. Low, “Sex Differences in Occupational Selection: The
Case of College Majors,” Review of Economics and Statistics 66, no. 1 (February 1984): 157-162.
This is shown in Gary S. Becker, “A Theory of Marriage: Part I,” Journal of Political Economy
81, no. 4 (July/August 1973): 813-846.
59See Alan King, “Human Capital and the Risk of Divorce: An Asset in Search of a Property
Right,” Southern Economic Journal 49, no. 2 (October 1982): 536-541.
Protection of Living Standards During the 1970s and early 1980s, the great-
est increase in labor force participation was among women in the 25 to 34
age group. Many economic and social factors spurred these young women into
the labor force. One of the most important was the attempt by young married
couples to maintain and advance their standard ofliving inthe face oftwo eco-
nomic squeezes. *
One squeeze on young married couples was the stiff competition for jobs as the
baby
generation
boom poured out of school and into the labor market. Between
1960 and 1980, the number of people aged 18 to 35 nearly doubled, as a result of
the rise in birthrates after World World II. A natural outcome of this rather large
and sudden increase in labor supply was a depressed labor market in the 1970s,
particularly for entry-level positions. One study estimated that the growth in
cohort size between 1967 and 1975 depressed the weekly earnings of college
graduates 13 percent. The second squeeze on young married couples was the
, particularly for housing. Even
as earnings were held down by the sharp competition for jobs, the level of con-
. Instead of growing
steadily year after year, the real annual income of the average family was 5 per-
cent lower in 1985 than in 1973.
The impact of these two economic events on the labor supply decision of mar-
ried couples can be analyzed with the household model of labor supply. If only
the husband worked, the slow growth in his earnings, coupled with the rapid rise
in prices, would have resulted in a substantial decline in the family’s standard of
living, particularly in comparison to two-earner families. This decline in real
income would have caused a pure income effect on the nonworking wife, leading
her to cut back on leisure in favor of market work. The entrance of women to
the labor force, however, also increased
the competition for jobs and further de-
pressed wage rates in entry-level positions. Since real wages for many jobs held
the price of nonmarket activities, setting off asubstitution effect on women’s la-
bor supply from market to nonmarket uses of time.”* The fact that a record num-
ber of women did enter the labor force during this period suggests, therefore,
that the desire to advance family income dominated the adverse effect on female
participation of a lower real wage per hour of work.
The substantial increase in the number of working women not only provided -
the means for many families to increase their standard of living, it also had a far-
reaching effect on the distribution of income among families. Twenty-five years
ago, there was a strong inverse relationship between husband’s income and the
wife’s probability of participation in the labor force. Because women from low
income households were the most likely to work, this helped to equalize family
*!See Valerie Kincade Oppenheimer, Work and the Family (New York: Academic Press, 1982).
* Finis Welch, “Effects of Cohort Size on Earnings: The Baby Boom Babies’ Financial Bust,” Jour-
nal of Political Economy 87, no. 5, pt. 2 (October 1979): 565-597.
Between 1972 and 1980, real weekly earnings for hospital workers, workers in department
stores, and computer programmers, for example, declined by —13 percent, —31 percent, and —23
percent, respectively.
Chapter 3. Labor Force Participation 129
income—poorer families had two earners whose combined income would more
nearly match the income of the single earner in the more well-to-do family. Re-
search finds that in more recent years, the presence of working wives still has an
equalizing effect on the distribut of income,
ion although the effect is not as
strong as before.** One reason is that the negative relationship between the
husband's income and the wife’s probability of being in the labor force has dimin-
ished as a greater proportion of wives from high-income families are now work-
ing. A second is that these wives, like their husbands, tend to have above-
average earnings.
Increased Job Opportunities for Women A fourth factor cited by many ana-
lysts as an important cause of the increase in female participation rates is the
> One form this has taken is the
substantial growth in the number of jobs in traditional “female” occupations in
the economy. It remains true today that women workers are concentrated in
relatively few occupations. In 1986, for example, 69 percent of women were em-
erical workers, nurses and
hygienists). While this range of occupations is fairly narrow, many of them have
grown rapidly in size in the post-war period. Perhaps the best example is in the
clerical fleld, where employment more than doubled between 1960 and 1987.
The second source of greater job opportunities for women is the gradual open-
ing up of a wide range of male-dominated occupations. Due to a combination of
factors including increased job commitment on the part of women, affirmative
action and antidiscrimination laws, and changes in social attitudes, employment
of women in many of these occupations has increased significantly. In 1960, for
example, only 3 percent of lawyers and 15 percent of managers were women. By
1986 the proportion of women lawyers and managers had risen to 18 percent and
37 percent, respectively.
Changing Social Attitudes A final consideration in explaining the rise in fe-
male labor force participation is the srone REP REP aCe concerning the
appropriateness of married women~working outside of the home, particularly
when they have small children. Prior to World War II, there were quite strong
social taboos against married women working outside of the home. In a poll
taken in 1937, only 18 percent of the respondents answered in the affirmative to
the question, “Do you approve of a married woman earning money in business or
industry if she has a husband capable of supporting her?” By the early 1960s,
attitudes had softened to the extent that only 46 percent of husbands said they
54See Francis W. Horvath, “Working Wives Reduce Income Inequality in Distribution of Family
Earnings,” Monthly Labor Review 103, no. 7 (July 1980): 51-53; David Betson and Jacques Van Der
Gaag, “Working Married Women and the Distribution of Income,” Journal of Human Resources 19,
no. 4 (Fall 1984): 532-543; and Joyce R. Shackett and D. J. Slottje, “Labor Supply Decisions, Hu-
man Capital Attributes, and Inequality in the Size Distribution of Earnings,” Journal of Human Re-
sources 22 (Winter 1987): 82—100.
55 This factor is given particular stress in Valerie Kincaid Oppenheimer, The Female Labor Force in
the United States (Berkeley, Calif.: Institute of International Studies, 1970).
disapproved when asked the question “Do you think it is a good thing for a wife
to work or a bad thing?” Probably the greatest transition in social attitudes
concerning married women and work took place during the 1970s. In 1968, only
14 percent of women aged 20 to 24 felt it was all right for a woman to work “if
she prefers to work, but her husband doesn't particularly like it.” Ten years later,
the response to this same question gained a 53 percent yes response.”’
These statistics suggest that social attitudes have changed significantly regard-
ing women and work outside of the home. There is a considerable controversy,
however, over the extent to which this change in social attitudes is a cause or a
. Some researchers
argue that the underlying determinants of the rise in female participation are
economic in nature, and changing attitudes (for example, the women’s move-
ment) simply provide an ideological rationale for a process that is already in mo-
tion.** Others have found evidence in statistical studies that changes in attitudes
do play an important, independent role in explaining why more women are en-
tering the labor force.”
Perhaps the clearest evidence that social attitudes influence family labor sup-
ply decisions comes from a comparison of labor force participation trends infor-
eign countries. Figure 3.10 shows the trend in the female labor force participa-
tion rate between 1960 and 1987 for four countries: Sweden, the United States,
Germany, and Japan. The participation rate of women in Sweden and the United
States increased sharply over this period (from 46 percent to 62 percent in Swe-
den between 1960 and 1987 and from 38 percent to 55 percent in the United
States), while in Germany and Japan the female participation rate
net decline. In Japan, the participation rate for women declined from 52.7 per-
cent in 1960, to 44.8 percent in 1975, after which it began to slowly rise, reach-
ing 47.6 percent in 1987. In Germany, the female participation rate fell from
41.2 percent in 1960 to 37.6 percent in 1981, and then it too began to rise
modestly, reaching 39.7 percent in 1987.
Economists have explained the rise in female participation in the United
States as a result of the positive substitution effect on women’s labor supply set off
by the secular
rise in real wages. Since real wages have increased considerably in
Germany and Japan over the last 30 years, why didn’t female participation rates
also increase in these countries? There are several explanations. One is that
the aggregate data mask conflicting trends among demographic subgroups. In
*°Oppenheimer, The Female Labor Force in the United States, 39-52.

‘Frank L. Mott, “Women: The Employment Revolution,” The Employment Revolution (Cam-
bridge, Mass.: MIT Press, 1982): 14-17.
*Oppenheimer, Work and the Family, 28-31; and Bowen and Finegan, The Economics of Labor
Force Participation.
* Mott, “Women: The Employment Revolution”; and Marianne Ferber, “Labor Market Participa-
tion of Young Married Women: Causes and Effects,” Journal of Marriage and the Family 44, no. 2 (May
1982): 457-468.
®See Wolfgang Franz, “An Economic Analysis of Female Work Participation, Education, and
Fertility: Theory and Empirical Evidence for the Federal Republic of Germany,” and Haruo Shimada
and Yoshio Higuchi, “An Analysis of Trends in Female Labor Force Participation in Japan,” Journal
of Labor Economics 3, no. 1, pt. 2 (January 1985), S218—S234 and S355—S374.
FIGURE 3.10 “<7 A ca

Trends in Female A gnocticvca
Participation in Four participation rate
Countries, 1960-1984 ( 100-
20 - Germany
1960. 1065. 7 1070. «1075 «=. 1980 1986

source: Bureau of Labor Statistics, unpublished data.
Japan, part of the drop in the female participation rate was due to the declineof
agriculture, a sector in Japan that had traditionally employed a large number of
women. Similarly, inGermany much of the decline in labor force participation
has been among young single women, while the participation rate among mar-
ried women has risen significantly.
Although these factors explain part of the divergent behavior in participation
rates among these four countries, they cannot explain it all. An additional fac-
tor that must be considered is the social attitudes in each country concerning
ing i e. While
these attitudes have undergone a virtual revolution in the United States in the
post-war period, in Germany and Japan they have historically been more rigid
and uncompromising. In Japan, for €xample, many companies have had a formal
or informal policy that a female employee must resign when she marries.°' Simi-
larly, social attitudes in Germany have always discouraged women from working
outside the home. Unlike the United States, even during the height of World
War II Germany did not utilize married women for work in defense plants.”
Rising wages in Germany and Japan did not result in a large-scale movement of
married women into the labor force, therefore, because, in part, social attitudes
regarding women’s place in society would not accommodate it. This conclusion
61See Alice H. Cook and Hiroko Hayashi, Working Women in Japan: Discrimination, Resistance,
and Reform (Ithaca: New York State School of Industrial and Labor Relations, Cornell University,
1980); and “Japan’s Secret Economic Weapon: Exploited Women,” Business Week (March 4, 1985):
54-55.
® See George Katona, et al., Aspirations and Affluence (New York: McGraw-Hill, 1971): 135-144.
IN « THE « NEWS
Dual Earner Families and Labor Mobility
Over the last 3 decades the “Ozzie and Harriet” manner. Traditionally, the United States has been
family of a working husband, a housewife, and sev- blessed with an extremely mobile work force, a
eral kids has gone from the rule to the exception in fact that has been a great source of strength rela-
American life. One important reason is the rising tive to other industrialized countries. Although
number of dual-earner families as more and more the United States still has an edge in worker mo-
married women join their husbands in marching bility, the proportion of families moving each year
off to work each day. The proliferation of dual- has declined fairly sharply over the last 2 decades,
earner families has had many economic and so- from about 21 percent in 1965 to 17 percent in
cial ramifications. One that has gained a growing 1985 (a 20-percent drop). Equally signficant, the
amount of attention in the press is the adverse im- decline in geographic mobility is most pronounced
pact that the dual-earner family has had on labor among adults in their 20s and 30s, the group that
mobility, particularly geographic mobility. in earlier years was the largest source of migra-
The willingness and ability of large numbers of tory flows.
people to move from one area of the country to A number of factors have worked together to
another in response to changing economic oppor- cause this drop-off in geographic mobility. One is
tunities and conditions is essential if the labor the steep price of housing, particularly if the fam-
market is to function in a smooth and efficient ily is contemplating a move from a community
merely confirms, however, what was learned from the labor/leisure model—the
decision to participate in the labor force is determined both
preferences
by (the
A BARGAINING Although the household model is the theoretical tool used by most economists
MODEL OF FAMILY to explain the pattern of family labor supply, it is not the only one. An alter-
LABOR SUPPLY native, more institutionally oriented model emphasizes the rolebargaining
of
Ss.
The household model explained the division of labor in the traditional fam-
ily of 30 to 40 years ago as a rational response to the different productivities
and rewards available to the husband and wife, respectively, in market work
and home work. Some economists have argued that social norms and women’s
inequality in bargaining power in the family were a more important cause of
women’s low involvement in market work than were rational considerations of
comparative advantage.®’ In this view, the reason relatively few women partici-
pated in the labor force in the years prior to World War II was not through
“Marianne Ferber and Bonnie G. Birnbaum, “The ‘New Home Economics’: Retrospects and
Prospects,” Journal of Consumer Research 4, no. 1 (June 1977): 19-28; Clair Brown, “An Institu-
tional Model of Wives’ Work Decisions,” Industrial Relations 24, no. 2 (Spring 1985): 182-204; and
Barbara Bergman, The Economic Emergence of Women (New York: Basic Books, 1986).
SS
a a
with a depressed economy (and real estate market) the career and income gains from moving for one
to one which is growing. A second is that many spouse may be more than offset if the relocation re-
companies have cut back on the number of corpo- sults in much lower earnings or career prospects for
rate relocations because of the high costs involved the other spouse. A recent Business Week article,
(an average of $35,000 per employee). A third im- for example, cites the case of a 27-year-old vice-
portant consideration is the growing number of president of corporate finance whose company
two-income families. asked him to move from New York to Los Angeles.
The proportion of married-couple families with His wife objected, however, saying, “I’m not mov-
two earners has increased from 41 percent in 1970 ing unless I get an incredible opportunity. I love
to 60 percent in 1985. A major explanation for my job. All our friends and family are in New
the decline in mobility rates is that the costs of York.” She did get an attractive job offer in Los
moving from one community to another are po- Angeles and they decided to make the move. The
tentially much greater for the dual-earner family decline in mobility rates in recent years suggests,
than for the family with only one working spouse. however, that many other dual-income families
For the single-earner family, the nonworking have decided to remain where they are.
spouse (usually the wife) gives up relatively little
financially in making a move, while her husband Sources: “American Workers Don’t Get around Much Any-
more,” and “Move Me, Move My Spouse: Relocating the Cor-
may be able to considerably advance his career porate Couple,” Business Week (October 28 and December 16,
and income. For the two-earner family, however, 1985): 94-96, 57-60.
choice, but lack of choice. This lack of choice emanated from several sources.
One was custom and tradition which made it socially unacceptable for married
women to pursue careers outside of the home. A second was discrimination on
the part of educational institutions, employers, and unions, which prevented
women from acquiring training and employment outside of the narrow range of
low-wage female occupations. A third cause of lack of choice was women’s un-
equal legal and economic position in the family vis-a-vis their husbands. As de-
scribed in a classic work by John Stuart Mill, these three factors resulted in most
married women remaining in the home, not for reasons of comparative advan-
tage, but because it preserved the hegemony (monopoly position) of men in the
job market and the family.”
This explanation for the sexual division of labor is far different than that
offered by the household model. Several pieces of evidence can be marshalled to
support it, however. One is the experience of working women in World War II.
During the war, millions of women entered the labor force to fill jobs left vacant
by men who were inducted into the armed forces. When the war ended in 1945,
many of these women voluntarily resumed their roles as mothers and house-
. “John Stuart Mill, The Subjection of Women (London: Longmans, Green, Reader, and Syer,
1869).
IN « THE « NEWS
How Couples React When Wives Outearn Husbands
A recent article in The Wall Street Journal sum- earning power of the wife has fulfilled dreams of
marized an informal study of married couples in equity and partnership in marriage. Some hus-
which the wife earned more money than the husbands even like to brag about how their wives’ suc-
band. The results clearly point to the important cess has liberated them from traditional wage-
role that money, power, and tradition play in de- earner burdens. But for many others, a wife’s status
termining the division of labor in many American as primary breadwinner creates anxiety and stress
families. in the marriage. Divorce is a frequent result. Re-
According to the article, about one-fourth of search finds, in fact, that the higher the woman's
the nation’s working wives earn more than their earnings, the higher the chance of divorce.
husbands, and about 2 million earn at least twice Why do marriages split up when the wife out-
as much. Surprisingly, however, only about 15 per- earns the husband? An important reason, accord-
cent of these women hold professional or mana- ing to the article, is that the husband’s position as
gerial positions and less than half have college de- breadwinner and head of the family is typically
grees. A close look at the statistics reveals that in based on the fact that he earns more—usually a
about half of the cases where the wife outearns the great deal more—than his wife. Hence, when the
husband it is because the husband is laid-off, re- wife brings home the bigger paycheck, the hus-
tired, or disabled. band often feels threatened and insecure since his
The article says that for some couples, the greater role and authority in the family are called into
wives. Many others, however, desired to keep their jobs since the level of earn-
ings was much higher than in the jobs traditionally open to women. Despite this
desire, a large portion were forced to quit, and the jobs were given to men.®
A second, more recent piece of evidence concerns the division of labor in
households where the wife has a higher level of education than the husband.
One study looked at 49 families where the wife had a Ph.D., but the husband
had less than four years of college.® Based on considerations of comparative ad-
vantage, it would be hypothesized that the wife would specialize in market work
and the husband would specialize in home production. None of the men, how- |
ever, were found to be “househusbands.”
A bargaining model of family labor supply also offers quite a different perspec-
tive on the reasons for the large increase in the female participation rate over the
last 20 to 30 years. The bargaining power of each spouse in the family is a func-
® At Ford Motor Co., for example, women comprised 25 percent of the work force during the
war, but only 4 percent in 1946. See Mary M. Schweitzer, “World War II and Female Labor Force
Participation Rates,” Journal of Economic History 15, no. 1 (March 1980): 89-95. On this subject,
also see Ruth Milkman, Gender at Work: The Dynamics of Job Segregation by Sex During World War II
(Champaign, IIl.: University of Illinois Press, 1987).
Cited in Ferber and Birnbaum, “The New Home Economics,” 21.
Se
erer ee
question. As one expert said, “The more money day. A number of women interviewed for the ar-
a woman makes, the more power she has in the ticle said they resented the fact that their hus-
household.” Another remarked, “Money is the bands did not do more of the housework, but that
key to understanding authority in the family.” The they preferred to do it themselves rather than risk
result, then, is that some husbands have difficulty family turmoil by bringing up the subject.
allowing their wives to have the additional au- The article concluded that when a wife outearns
thority in the family that goes with a bigger pay- the husband, the marriage is almost always signifi-
check, while the wives are apt to believe they de- cantly altered, sometimes for the good and some-
serve a greater say in family decisions and feel times not. The couples most likely to successfully
frustrated if they don’t get it. adjust to their role reversal tend to fall in one of
Besides authority, a second issue that comes to three groups: couples with similar education levels
the fore when the wife outearns the husband is the and occupations, wealthy couples who can avoid
allocation of housework. Interviews revealed that the conflict over household chores by hiring help,
some husbands willingly do more of the household and couples who start out in marriage with the
chores such as shopping, cooking, and laundry in wife as the primary breadwinner.
recognition of their wives’ heavy time demands at
work. Others however, balk at taking on these
tasks and do relatively little of the housework,
Source: Laurie Hays, “Pay Problems: How Couples React When
even in situations where the wife works full time, Wives Outearn Husbands,” The Wall Street Journal (June 19,
but the husband remains at home for most of the 1987): 25.
tion of his or her dependency on the other: the greater the degree of dependency
of spouse X on Y, the less bargaining power X has to influence decision making
and the allocation of income and consumption in the family.®’ In this view, mar-
ried women in the family of 1950 were at a distinct disadvantage in bargaining
power in the family because they were far more dependent on their husbands
than their husbands were on them. One important reason was economic—
the range of jobs and relatively high*ates of pay open to the husband gave him
more financial independence than the wife. A second factor was the near-
inevitability of children because of the primitive nature of birth control. Why,
then, did so many married women enter the labor force in the 1960s and 1970s?
The answer is that employment afforded women the opportunity to achieve posi-
tions of equality in the family. One factor facilitating this trend was the growing
6 The link between bargaining power and dependency is developed in Samuel B. Bacharach and
Edward J. Lawler, Bargaining, Power, Tactics, and Outcomes (San Francisco: Jossey-Bass Publishers,
1981). A formal bargaining model of family labor supply is given in Marilyn Manser and Murray
Brown, “Bargaining Analysis of Household Decisions,” in Cynthia Lloyd, Emily Andrews, and Cur-
tis Gilroy, eds., Women in the Labor Market (New York: Columbia University Press, 1979): 3-26.
. See Roberta M. Spalter-Roth, “Differentiating between the Living Standards of Husbands and
Wives in Two Wage-Earner Families, 1968 and 1979,” Journal of Economic History 18, no. 1 (March
1983): 231-240.
number of white-collar and “pink-collar” jobs available to women. A second was

the improvement in contraceptive devices that allowed women to both limit
and time the birth of children.” A third was the rather marked shift in social
attitudes toward working women, coupled with the passage of antidiscrimination
and affirmative action laws. While in the household model it was largely the rise
in real wages over time that pulled married women into the labor force, from a
bargaining perspective, these women were attracted into the labor force not only
by rising wages, but also by the opportunity to establish personal independence
and equality in marriage.
SUMMARY Labor force participation involves the question of whether or not to work in the
market. The labor/leisure model shows that work in the market is optimal only if
the market wage exceeds the individual’s reservation wage. The person's reserva-
tion wage is determined, in turn, by factors such as the preferences for work and
leisure and productivity in nonmarket activities. The labor/leisure model shows
that the probability of labor force participation increases with the wage rate.
This insight is further developed by the household model of labor supply. This
model shows that a wage increase for one spouse affects not only that person’s
labor supply, but also that of the wife or husband. One channel through which
this occurs is the cross-substitution effect.
The most important trend in labor force participation in the last 30 years has
been the rapid increase in the proportion of women working outside of the
home. This development has many origins including the rise in real wages,
greater educational attainment, expanded job opportunities, the desire to ad-
vance family living standards, changing social attitudes, and the desire of
women for equality of bargaining power in the family. Concomitant with the rise
in female labor force participation has been the gradual decline in the participa-
tion rate of men. This was traced to several factors such as the desire for earlier
retirement brought on by the secular increase in incomes, the greater coverage
and benefits from Social Security and private pension plans, and the declining
participation rate among prime-age black males due to poor health and discour-
agement over job prospects.
GLOSSARY
Added worker effect — This effect causes the la- Age/participation profile—The relationship be-
bor force to grow during recessions as other fam- tween age and labor force participation rate for
ily members seek work to make up for the drop a cross-section of persons.
in family income due to the loss of job by the | Current Population Survey (CPS) — A nation-
primary earner. wide survey of 60,000 households conducted
See David M. Heer and Amyra Grossbard-Schestman, “The Impact of the Female Marriage
Squeeze and the Contraceptive Revolution on Sex Roles and the Women’s Liberation Movement in
the United States, 1960-1975,” Journal of Marriage and Family 43, no. 1 (February 1981): 49-66.
each month to obtain data on the size and char- cause some of the unemployed give up searching
acteristics of the labor force. for work.
Cross-substitution effect—The change in one Labor force participation rate— The percent of
family member's hours of work given a change in the noninstitutional population that is employed
the wage rate of another family member, hold- or looking for work.
ing total income of the family constant. Reservation wage— The wage rate at which a
Discouraged worker—A person who gives up person would just be willing to work the first
searching for work due to discouragement over hour at market employment.
the prospects of obtaining employment. Unemployment rate — The percent of the labor
Discouraged worker effect— This effect causes force counted as unemployed.
the labor force to shrink during recessions be-
a a a a a
REVIEW QUESTIONS
1. Why might a married woman with children were to allow recipients to earn as much as desired
have a higher reservation wage than a single from work with no reduction in benefits (t = 0)?
woman with no children? Show this in a graph using the labor/leisure model.
2. According to the household model, why is 6. Use the labor/leisure model to analyze the im-
the substitution effect for women likely to be larger pact of child care costs on the decision of a mar-
than the substitution effect for men? What should ried woman with children to participate in the la-
be the relationship between the husband’s income bor force.
and the wife’s probability of labor force participa-
tion according to this model? a. First, draw a budget constraint assuming
3. The household model suggests that the divi- the wage is $6 per hour and the woman
sion of labor in the traditional family where the has $200 per week of nonlabor income.
husband engages in market work and the wife en- Then draw in an indifference curve so
gages in home work is the result of rational deci- that her equilibrium hours of work are
sion making based on considerations of compara- 20 per week.
tive advantage. Critics of this theory argue that b. Next, assume if the woman works she
the division of labor between the sexes has little must pay child care costs. These costs are
to do with the innate productivities of men and of two kinds. The first is a lump sum or
women in market versus home work, but rests on _ fixed cost of $50 per week for, say, trans-
the forces of tradition and male dominance in portation of the child to a day care facil-
the family. Briefly discuss the arguments for and ity. The second is a variable cost of $2 for
against each point of view. each hour the child is at the day care fa-
4. Why does the added worker effect cause the cility. Given these two types of costs,
labor force to grow during recessions? Why does draw the new budget constraint for this
the discouraged worker effect cause it to shrink? woman.
Which effect is likely to be more important? c. Without child care costs, the woman
5. Under the Social Security law, every dollar of maximized utility by working 20 hours;
earnings over $8,160 is subject to an implicit tax with child care costs, will this woman
rate of t = .5 since benefits are reduced by 50¢ continue to work or will she drop out of
for every dollar of additional wage income. What the labor force? To determine this, draw
would be the effect on labor supply if Congress an indifference curve so that it has a tan-
gency point with the new budget con- of child care costs be different if she
straint. Is this level of utility higher or had originally been working 50 hours per
lower than if she did not work at all? week? Show this and explain why.
Show this in the graph. Would the effect
SY SS SY 3
a
ADDITIONAL READINGS
Berkowitz, Monroe, and M. Anne Hill, eds. Dis- Association, 1987. Contains 13 articles that cover
ability and the Labor Market: Economic Problems, a variety of topics such as women's occupational
Policies and Programs. Ithaca, N.Y.: ILR Press, earnings and attainment, problems of women in
1986. A collection of ten articles that focuses on the management, women and labor unions, and an
problems of disabled workers and, in particular, the international comparison of women’s position in the
impact of government programs such as Social Se- workplace.
curity, worker compensation, and vocational re- Lazear, Edward. “Retirement from the Labor
habilitation on the labor supply of the disabled. Force.” In Orley Ashenfelter and Richard
Blau, Francine, and Marianne Ferber. The Eco- Layard, eds. Handbook of Labor Economics. Am-
nomics of Women, Men and Work. Englewood sterdam: North-Holland, 1986: 305-356. Pro-
Cliffs, N.J.: Prentice-Hall, 1986. Provides an in- vides a comprehensive review of previous research
depth but relatively nontechnical review of theory on economic aspects of retirement.
and issues pertaining to the labor supply determi- Linder, Steffan. The Harried Leisure Class. New
nants of men and women, the division of labor in York: Columbia University Press, 1970. An en-
the family, and the earnings and occupational at- tertaining and insightful account of why, as people
tainment of women. earn more money, they seem to have less time to en-
England, Paula and George Farkas. Households, joy it.
Employment, and Gender: A Social, Economic, Mincer, Jacob. “Intercountry Comparisons of La-
and Demographic View. New York: Aldine De bor Force Trends and of Related Developments:
Gruyter, 1986. A comprehensive review by two An Overview.” Journal of Labor Economics 3,
sociologists of the economic and social determinants supplement (January 1985): S1—S33. Examines
of the division of labor in the household and the labor force trends in 12 industrialized countries and
evolving roles of women in the home and in the to what degree these trends can be explained with
market. the neoclassical model of labor supply.
Fuchs, Victor. “His and Hers: Gender Differences Pleck, Joseph. Working Wives/Working Husbands.
in Work and Income, 1959-1979.” Journal of Beverly Hills, Calif.: Sage, 1985. A thorough ex-
Labor Economics 4, pt. 2 (July 1986): S245- amination of the patterns and determinants of hours :
$272. Examines the changes over time in hours of home work done by men and women in dual-
of work, leisure, and income available to men and earner families.
women and finds that women's access to goods, ser- Willis, Robert J. “What Have We Learned from
vices, and leisure was actually lower in 1979 than the Economics of the Family?” American Eco-
in 1959. nomic Review 77 (May 1987): 68-81. Reviews
Koziara, Karen Shallcross, Michael H. Moskow, the economic theory of the family and its ability to
and Lucretia Dewey Tanner. Working Women. explain differences in divorce, fertility, and child
Madison, Wis.: Industrial Relations Research care across families.
APPENDIx 3A,
The Allocation of Time to Market Work,
Home Work, and Leisure: A Graphic Exposition
The labor/leisure model was extended in Chapter 3 to incorporate a three-way
allocation of time to market work, home work, and leisure. To keep the exposi-
tion relatively simple, this model was explained solely in verbal terms. This ap-
pendix develops the same model using a graphic approach for those students who
want a more in-depth treatment of this subject.'! Additional insight into the
Becker time allocation model is also provided.
THE MODEL In the labor/leisure model, an individual's utility is assumed to be a function of

goods (purchased with labor earnings and nonlabor income) and leisure time.
Becker has argued, however, that goods and time are better treated as inputs
that people use to produce various “commodities” that are the direct source of
utility.” As an example, a box of fried chicken does not yield any utility until it is
combined with an hour of time to form a commodity called a picnic lunch.
To develop this line of reasoning further, assume a single-person household.
This person seeks to maximize the amount of commodity Z, which is a combina-
tion of goods and services (X) and consumption time (L),
Z—=Z(X, DL): (3A.1)

The goods can either be purchased in the market (X,,) or produced at home
(X,), implying
NOS KG, ale Gs (3A.2)
Market goods and home goods are assumed to be perfect substitutes for each
other, and both are measured in terms of their dollar value.
Home goods are produced by work at home, according to the production
function: ~
X}, = f(H), (3A.3)

where H is hours of work in the home. (Note that this definition of H is different
from that in Chapters 2 and 3.) Since the capital stock and technology of pro-
duction in the household is assumed constant, additional hours of home work
increase the amount of home goods produced, but at a diminishing rate (i.e., the
factor input H is subject to declining marginal productivity).’
'This appendix is drawn from Reuben Gronau, “Leisure, Home Production and Work—The
Theory of the Allocation of Time Revisited,” Journal of Political Economy 85 (December 1977):
1099-1124.
2See Gary Becker, “A Theory of the Allocation of Time,” Economic Journal 75 (September
1965): 493-517.
3Students unfamiliar with the concepts of a production function and marginal product can find
both terms defined in the next chapter.
139
140 Appendix 3A The Allocation of Time to Market Work, Home Work, and Leisure
FIGURE 3A.1 The Allocation of Time to Market Work, Home Work, and Leisure
The slope of the dashed
lines at points A and B
reflect the values of the
market wage, the slope of
the concave curve ABCD
reflects the marginal
product of time devoted to
home production. Given
zero nonlabor income, the
person's opportunity set is
ABCF. \f the person's
indifference curve is
/, utility is maximized at
point B and 7, — 7, hours
of time are devoted to
home production, zero
hours to market work, and
T, — Tp hours to leisure.
Given a different set
of preferences as
represented by indifference
curve /, utility is maxi-
mized at point E with 7, —
T; hours of home work,
T; — T, hours of market
work, and 7, — 7p hours
of leisure. An increase
in nonlabor income shifts
the opportunity set to
AGJM, resulting in a new
allocation of time.
meee yest| Bo ea Ora. La Time (T)
In attempting to maximize utility, this person is bound by two constraints.

The first is the budget constraint:
Xv = WN + V, (3A.4)
where W is the person’s wage rate (assumed to be constant), N is hours of market
work, and V is nonlabor income. The second constraint is the time constraint:
T=L+H +N; (3A.5)
where T is total time available per period (e.g., week, month).

Given this framework, the issue to be determined is how the person will allo-
cate his or her time to the three activities: market work, home work, and con-
sumption time or “leisure.” The answer is given in Figure 3A.1. The concave
curve ABCD depicts the relationship between hours of time devoted to home-
work and the value of home goods produced, as given by the home production
Appendix 3A The Allocation of Time to Market Work, Home Work, and Leisure 141
function in Equation 3A.3. At point A, all time is devoted to T, hours of lei-

sure, implying H = O and Xj, = 0. At point B, hours of leisure have fallen to T,
and hours of home work have increased to T, — T,. The slope of the curve
ABCD measures AX,,/AH, the marginal product of home work or the “shadow
wage” (i.e., the implicit dollar value received from an hour of home work). The
curve becomes progressively flatter, illustrating that home work is subject to de-
clining marginal productivity.
The next step is to incorporate market work into the diagram. The relation-
ship between goods and market work is given by Equation 3A.4. Assume ini-
tially that V = 0 (no nonlabor income) and W = W,. Starting at point A,
if the person works successive hours in the market instead of at home, the
value of goods available is given by the dashed straight line budget constraint
(only a portion of which is drawn in). The slope of the budget constraint is
AX,/AN = W,. Since the return to home work (the shadow wage) is assumed
at point A to be greater than the return to market work (the market wage), a
person who decides to work should initially start out in the home. Next consider
point B. If the person chooses to work one more hour, should it be in the market
or at home? The answer is in the home since the slope of the curve ABCD at
point B still exceeds the return to market work, given by the slope of the dashed
straight line at point B. As still more hours are devoted to home work, at some
point, such as at C, it is likely that market work will begin to yield a bigger
increment of goods than home work. This is illustrated by the fact that beyond
point C, the slope of the solid straight line CF exceeds the slope of the curve
CD. The relevant opportunity set for this person, therefore, is ACF.
If it is assumed that the person also has nonlabor income of, say, V,, the effect
is to shift-out the opportunity locus to AGJM. The addition of nonlabor income
does not affect the point at which market work is more productive than home
work since neither the value of the market wage nor shadow wage is affected.
The opportunity set ACF or AGJM shows the combinations of goods and lei-
sure that are available. To determine the person’s actual allocation of time, it is
also necessary to know his or her preferences for the two items. This information
is given by Equation 3A.1, which can be used to derive a set of indifference
curves. For purposes of exposition, two alternative sets of indifference curves are
represented in Figure 3A.1, I, and I, along with I, and I,.
The person's optimal allocation of time can now be determined. Assume first
the opportunity set is ACF. If the person's indifference curve is I,, utility is
maximized at point B. This person will devote T, — T, hours to home work,
zero hours to market work, and T, — Ty hours to consumption time or “leisure.”
If the person's indifference curve is 1;, however, a different allocation of time
results. In this case, utility is maximized at point E and T; — T; hours are de-
voted to home work, T; — T, hours are devoted to market work, and T, — T;
hours are spent on leisure.‘
4If the market wage is greater than the shadow wage at point A, the diagram becomes, in effect,
the traditional labor/leisure model. In this case, the person would devote zero hours to home work.
142 Appendix 3A The Allocation of Time to Market Work, Home Work, and Leisure
FIGURE 3A.2 The Impact on the plcceuas of Time of a veh in the Marketbe
The original opportunity
set is ADG. With the
indifference curve /3 (point
F), the person's original
allocation of time would be
Ts — Ts hours of home
work, 73 — 7, hours of
market work, and 7; — Jy
hours of leisure. An
increase in the wage rate
changes the opportunity
set to ABK, giving rise to a —_ —
a=
new equilibrium at point i|

J, where leisure hours {
increase, hours of market i
i}
work increase, and time 1
spent on home work i
declines. 1
L}
!
1
L}
'
!
t
i
t
t
i
i
1
I ee
ee f i
ee
ae
:—
ee
ee
ee
SF
Next consider how the allocation of time to these three activities is affected
by changes in nonlabor income and the market wage rate. Figure 3A.1 illustrates
the impact of changes in nonlabor income. Increasing nonlabor income from
V = 0to V = V, shifts the opportunity set from ACF to AGJM. If the person ‘
had been at point E, the new equilibrium will be at point K. The result is
a decline in market work from T,; — T, to T; — T), an increase in leisure, and
no change in home work. Alternatively, the new equilibrium for the person at
point B is point H. In this case, hours of home work fall from T; — T; to
T, — Ts, while leisure time increases and hours of market work remain un-
changed at zero.
The impact of a change in the market wage rate on the allocation of time is
illustrated in Figure 3A.2. The initial opportunity set is ADG. An increase in
the market wage changes the opportunity set to ABK. The line segment BK has
a steeper slope than DG and is tangent to the home production curve further to
Appendix 3A The Allocation of Time to Market Work, Home Work, and Leisure 143
the right. Assume first that the person is initially working in the market, such
as at point F. Hours of home work are T; — T;, hours of market work are
T; — Tj, and hours of leisure are T, — Tp. The increase in the wage leads to a
new equilibrium at point J. One unambiguous prediction is that hours of home
work will decline, from T; — T; to Ts — Ts.° The impact on hours of market
work and leisure is indeterminant, however, just as in the simple labor/leisure
model. Hours of market work will increase only if the increase in leisure time
from T, — T) to T, — Tp is more than outweighed by the decrease in home pro-
duction time from T, — T; to T, — T;. This condition is more likely to occur as
the rate of substitution increases between goods and consumption time in the
person’s utility function and as the income elasticity of leisure declines, among
other things.
A second possibility to consider is a situation where the person initially en-
gages only in home production (point C). An increase in the market wage in
this case results in a new equilibrium at point H. The net result is that hours of
market work increase from zero to T; — Ty, reflecting the decision of this person
to begin to participate in the labor force. At point H, hours of home work have
declined by a like amount, while hours of leisure time have remained the same.
Leisure time could also increase or decrease, however, depending on the shapes
of the indifference curves. Finally, note that if the original tangency point at C
had instead occurred to the right of point B, the increase in the market wage
would have no impact at all on the allocation of time.
*Empirical evidence in support of this prediction is presented in Peter Kooreman and Arie
Kapteyn, “A Disaggregated Analysis of the Allocation of Time within the Household,” Journal of
Political Economy 95 (April 1987): 223-249.
CSHVASe EER
The Demand for Labor

in the Short Run
In the next two chapters, attention is focused on the demand for labor. The
demand for labor is concerned with the level of employment desired by business
firms. Important issues to be analyzed are the factors that determine a firm’s op-
timal level of employment, and how this desired level of employment responds
to changes in wages, the costs of capital and other inputs, the level of sales of the
firm, and improvements in the state of technology.
The analysis of the demand for labor is broken into two parts. This chapter
examines the demand for labor in the short run, reserving until the next chapter
\, consideration of labor demand in the long run. The distinction between the
short run and long run is of fundamental importance. In its attempt to maximize
profits, a business firm will adjust its use of labor, capital, energy, and other fac-
tor inputs to achieve the lowest costs of production. If the relative price of labor
should increase, the firm is motivated to cut back on the use of labor and sub-
stitute capital and other factor inputs in its place. Every business firm, however,
needs time to order and put in place new machinery or to build a new, more
automated plant. This initial length of time when the firm is locked into a fixed
amount of plant and equipment is defined as the short run. The long run is de-
fined as the period of time after which the firm can change not only labor, but
the amounts of capital and all other factor inputs. The distinction between the
short run and the long run is a conceptual one. In practice, the actual length of
the short run varies considerably from industry to industry. A textile firm, for
example, could quite possibly purchase and put in place new, more technologi-
cally advanced dyeing and weaving machines in several months; a new inte-
grated steel plant, on the other hand, might take 5 years or more to complete.
The first objective of this chapter is to derive the firm’s demand curve for
labor, showing, in the process, the reason for the inverse relationship between
the wage rate and the firm’s desired level of employment. The next topic is the
elasticity of demand for labor. The importance of this concept is illustrated by an
analysis of wage concessions in the auto industry and the employment effects of a
wage subsidy program. The following section analyzes the link between the firm’s
145
146 Chapter 4 The Demand for Labor in the Short Run
FIGURE 4.1. The Pattern of Employment Growth in the U.S. Economy, 1950-1987
Employment
(in millions)
120
100
Total employment
80
60
Service-producing
40
Goods-producing
20
Government
. Agriculture
(i ———_—
1950 1960 1970 1980 1987
source: Bureau of Labor Statistics, Employment and Earnings 35, no. 3 (March 1988): Tables A—1, B—-1.
level of sales and its desired level of employment, focusing on issues such as the
impact of the business cycle on employment and the employment effects of for-
eign trade. The last topic covered in the chapter concerns the use of hiring stan-
dards and other screening devices in the employee selection process as business
firms try to decide not only how many people to hire, but also whom to hire.
THE PATTERN OF The demand for labor is reflected in the level of employment among business
EMPLOYMENT firms and nonprofit organizations such as government. Several key features of
the pattern of employment in the United States are illustrated in Figure 4.1.
Shown there are the annual levels of employment in the entire economy for the
years 1950 through 1987 and the levels of employment in four major sectors:
agriculture, other goods-producing industries (mining, construction, and man-
ufacturing), government, and other service-producing industries (transportation
and communications, trade, finance, and services).
Several aspects of the pattern of employment depicted in Figure 4.1 are par-
ticularly noteworthy. The first is the trend over time in the total level of employ-
ment in the economy. To keep unemployment from rising, the economy must
generate nearly 2 million new jobs each year, both because of the entrance of
additional people into the labor force and because productivity growth allows
Chapter 4 The Demand for Labor in the Short Run 147
firms to produce the same level of output with less labor. How has the American
economy done? Since 1950, employment growth has been quite remarkable,
with the total number ofjobs having increased from 60 million to 114 million in
1987. To put this fact in perspective, consider that while the American econ-
omy was adding 25 million new jobs between 1970 and 1986, the four largest
European countries had a net job gain of only slightly more than 1 million.'
A second important feature of employment is the relative decline of agricul-
ture and the goods-producing industries as a major source of jobs in the econ-
omy. In 1950, 48 percent of jobs were in agriculture or industries that produced a
physical good of some type. Over the next 38 years, however, employment in
agriculture was cut in half, while employment in the goods-producing industries
managed a scant increase of 6 million. This performance stands out in stark con-
trast with employment in government and other service-producing industries.
Employment in government grew by over 250 percent over this period, and the
service-related industries in the private sector generated over 38 million new
jobs. As a result of these disparate trends, by 1987, nearly seven out of every ten
jobs in the American economy were in government or private sector industries
that produced a service.
A third feature of employment that deserves attention is the rise and fall of
employment over the business cycle. Even as employment increased steadily
from one decade to another, this upward march was temporarily interrupted by
seven recessions between 1950 and 1987. During each recession, the decline in
business activity caused some firms to slow their hiring while others closed plants
and laid off workers. The result was a significant slowdown in employment
growth or, when the recession was quite severe as in 1981 and 1982, an actual
decline in the number of jobs in the economy. Figure 4.1 shows, however, that
the effect of recession on employment is not spread evenly through the econ-
omy. In government and service-related industries, it is difficult to discern reces-
sion years from years of prosperity since employment in these sectors is not very
cyclically sensitive. Goods-producing industries, on the other hand, are quite
vulnerable to the cyclical swings in economic activity, as revealed by the de-
clines in employment in recession years such as 1971, 1975, and 1982.
These are some of the most important aspects of employment on which the
next two chapters focus. However, first, it is necessary to develop the basic the-
ory of labor demand.
THE MARGINAL The optimal level of employment for a business firm is contingent on a number
PRODUCTIVITY of factors, such as the cost of labor, the productivity of the work force, the level
THEORY OF LABOR of production, and the price the business firm can charge for its product. The
DEMAND theory of labor demand developed here organizes these diverse considerations
into a relatively simple model of employer decision making that has, neverthe-
less, considerable ability to explain real-world behavior.
1See Janet L. Norwood, “Labor Market Contrasts: United States and Europe,” Monthly Labor
Review 105, no. 8 (August 1983): 3—7. Data updated to 1986 by the author.
The standard model of labor demand in economics is the neoclassical marginal

productivity theory of demand. To develop this model, several simplifying assump-
tions must be made. First, it is assumed that the goal of business firms is to maxi-
mize dollar profits. Second, in order to simplify the graphic analysis, it is as-
sumed that the firm uses only two factors of production, capital and labor, to ;
produce its product. Third, it is initially assumed that the business firm operates
in perfectly competitive product and labor markets. The substantive importance
of this assumption is that both the price the firm can get for its product and the
wage rate it has to pay for labor are unaffected by changes in its individual pro-
duction and hiring decisions, and thus both variables can be treated as given.
Finally, it is assumed that wages represent the only cost of labor and that labor is
completely homogeneous. Each of these assumptions is then relaxed or critically
examined in either this or the next chapter.
How many workers should a business firm hire? Economic theory suggests a
simple answer: hire additional workers as long as ea
$1 of extra profitit
business
to
tothe
the firm. This solution is one specific example of the
basic marginal decision rule found throughout economics—continue to do some-
thing as long as the marginal increase in the benefit exceeds the marginal in-
crease in the cost from that activity. Thus, the firm that wants to maximize profits
(profits are the difference between revenue and cost) should determine the mar-
ginal increase in revenue from hiring one more worker and compare it with the
marginal increase in labor cost of that last worker. As long as additional workers
bring in more revenue than the cost of hiring them, the firm adds to its profit by
increasing employment; when the last worker added by the firm brings in reve-
nue no greater than the labor cost, hiring should stop and profits will be at a
maximum.
In deriving the short-run demand for labor, the first task is to determine the
increase
_—
in revenue to the firm from hiring an additional worker. This involves
several steps.
The Marginal Product of Labor The first step is to calculate the extra physi-
cal output or production that the firm will get by employing an additional worker.
The relationship between the amount of capital inputs (K) and labor inputs (L)
used in production and the resulting output (Q) is determined by the firm’s pro-
duction function:
= 7h) (4.1)
The production function represented in Equation 4.1 states that the amount of
output produced Q is a function f of the amount of capital K and labor L used by
the firm, given the current state of technology. While the production function
given in Equation 4.1 is simply an abstract mathematical representation, econo-
mists can determine the actual statistical relationship between inputs and out-
puts for a particular firm or industry, given the appropriate data.’
* An interesting and relatively nontechnical example is given in Leslie Cookenboo, Jr., Crude Oil
Pipelines and Competition in the Oil Industry (Cambridge, Mass.: Harvard University Press, 1955);
8-32.
EWE
TABLE 4.1 Labor Quantity Marginal Marginal Marginal Reve- Cos C
Data Necessary to Input of Output Product Revenue nue Product Wage Rate
Derive a Hypothetical L Q MP, yx. MR MRP,
(1) (2) (3) (4) (5) (6)
Firm's Labor Demand
Curve 0 0 = $2 — $6
1 by 6 2 $12 1% 6
2 4s 8 2 16 |2s 6
3 20 6 2 2s 6
4 24 4 2 8 A 6
& 27 3 2 6 4 6
6 29 2 2 4 © 6
7 30 \ 1 2 2).% 6
\
FIGURE 4.2 The avin between Total and Marginal Product Schedules
The marginal product of
each worker is given by
the slope of a line drawn
(a) SS
Total
tangent to the total product
output
curve /F,. Up to L, the
tangent lines become (Q)
steeper, indicating that TP,
production is in the area of
increasing returns; after L,, Total product schedule
however, the tangent lines
become flatter, indicating
that diminishing returns
has set in. Beyond L, the
Marginal product of labor
becomes negative. Plotting
the slope of the tangent Ly L, Labor (L)
lines in the bottom portion
of the diagram yields the
marginal product schedule (b)
MP,. After point A, labor is Marginal
subject to diminishing _ product
marginal productivity.— (AQ/AL)
Marginal product schedule
Labor (L)
MP,
Since the amount of capital is fixed in the short run, additional output can be
produced_ thebyfirm only by hiring additional workers. Holding capital con-
stant, the production function in Equation 4.1 can be used to predict the in-
crease in output that would result from these extra workers. Such data for a hypo-
thetical firm are shown both in tabular form in Table 4.1 and in graphic form in
Figure 4.2.
The top of Figure 4.2 depicts the total product curve TP, for the firm; it shows
the total amount of output produced at each level of employment. In deriving
the demand curve for labor, it is particularly important to know the increment
in production contributed by each additional worker who is hired by the firm, a

concept called the marginal product of labor (MP,). The marginal product of
labor is defined as the increase in total production (AQ) from adding one more
unit of labor (AL), that is,
MP, _= =
AQ
Geometrically, the marginal product of each worker is given by the slope of a

line drawn tangent to the total product curve at each level of labor input.
The shape of the total product curve in Figure 4.2 shows that production in
this firm is first characterized by increasing marginal productivity of labor (com-
pare the slope of the first two tangent lines), and then beyond the employment
level of L, (point A), by diminishing marginal productivity of labor (compare
the slope of the last two tangent lines). Starting from a zero level of output, the
marginal product of each additional worker at first increases, as only a larger
work force can efficiently use and operate the plant and equipment of the firm.
From the origin to L,, therefore, is known as the area of increasing returns in
production. Beyond L,, however, production enters the area of diminishing re-
turns as the mangina/ product ofeachadditional worker-econtes smaller and
smaller. The law of diminishing returns states that, holding one factor of pro-
duction constant, as additional units of another factor are added total produc-
tion will increase but, beyond some point, only at a diminishing rate. In this
particular case, diminishing returns arise because the fixed amount of plant and
equipment of the firm is gradually spread among an ever greater number of work-
ers, leaving a smaller amount of capital for each one. While diminishing returns
set in at an employment level of L,, if labor is increased beyond L, (point B)
efficiency will suffer so greatly that total output will actually decline. Beyond
point B, therefore, the marginal product of labor is negative.
It is possible to directly graph the marginal product of labor, as shown in the
lower portion of Figure 4.2. The horizontal axis measures units of labor; the ver-
tical axis measures the additional output produced by each worker. The curve
MP, is called a marginal product schedule; it shows the addition to output of each
successive worker. It is derived from the top part of Figure 4.2 by plotting the
value of the slope of the tangent lines at each level of labor input. The marginal
product schedule rises while production is in the area of increasing returns; |
beyond point A, diminishing returns set in and the marginal product schedule
begins to decline, becoming negative after point B.
The same relationships are depicted by the hypothetical data given in Table
4.1. Column 1 in Table 4.1 shows the number of workers L employed by the
firm. Column 2 shows total production Q per hour for each level of employ-
ment. Given these data, it is then possible to calculate the marginal product of
labor MP,, shown by the data in column 3. The marginal product of the fourth
worker, for example, is four units per hour—the increase in production from 20
to 24 units attributable to hiring him or her. The data in column 3 show that
production is subject to increasing returns up to the second worker, and to di-
minishing returns thereafter.
Marginal Revenue Product The marginal product data in column 3 of Table

4.1 do not yet tell the firm the extra revenue it will gain from employing an
additional worker. The marginal product of labor is expressed in terms of physi:
cal units of production, To determine the dollar worth to the firm of the mar-
ginal product of each worker, those units of production have to be multiplied by
their marginal revenue (MR). Marginal revenue is defined as the change in total
revenu (TR) received
e by the firm from producing one more unit of output; that
is, MR = ATR/AQ. For a perfectly competitive firm, as assumed here, the price
(P) received for the product and its marginal revenue are the same. Since a com-
petitive firm can sell every additional unit of productionat the same price, the
increment in revenue from selling one more unit (the marginal revenue) is ex-
actly equal to the price it is sold for.
The marginal revenue received from each unit of production for this hypotheti-
cal firm is shown in column 4 of Table 4.1. It is assumed here the product price,
and hence the marginal revenue, is $2 per unit. The marginal increase in reve-
nue for this firm from hiring additional workers can now be calculated, which is
done in column 5 of Table 4.1. Column 5 contains data on the marginal reve-
nue product (MRP,) of labor, calculated by multiplying the marginal product of
labor MP, by the marginal revenue MR. How much extra revenue does the firm
get from hiring the third worker? The third worker produces an extra six units of
production per hour, and each unit sells for $2. The increase in revenue for the
firm, then, is MP - MR, or (6)($2) = $12. Similarly, the marginal revenue
product of the fourth worker is (4)($2) = $8. Note that the marginal revenue
products of the second worker through seventh worker in Table 4.1 decline in
value. The reason is not that the last worker hired is intrinsically any less pro-
ductive, efficient, or capable than the previously hired workers. The decline is
tied directly to the lawof diminishing returns and the fact that additional work-
ers have less capital to work with.
The marginal revenue product data express the marginal benefit to the firm
from hiring additional workers. The decision rule to maximize profits is to con-
tinue to hire workers as long as the marginal benefit exceeds the marginal cost.
The final piece of information needed to determine the firm’s demand for labor,
then, is the cost of hiring an additional worker.
The cost per hour to the firm of hiring each worker isequal to the wage rate,
assuming there are no fringe benefits or other types of labor cost. It is assumed
here that the wage rate (W) is $6 per hour, shown in column 6 of Table 4.1.
Because this firm is assumed to operate in a competitive labor market, its mar-
ginal wage costs per worker are the same_no matter how many workers it hires.
The Equilibrium Level of Employment It is now an easy task to determine the
optimal level of employment for this firm. Should it hire the fourth worker? The
answer is yes since the fourth worker produces $8 worth of output and costs only
$6 per hour to hire, resulting in $2_of extra profit for the firm. What about the
sixth worker? The answer is clearly no since the wages paid to the sixth worker
($6) exceed the value of his/her production ($4). Hiring the sixth worker would
reduce profits by $2. Obviously, then, the optimal level of employment in this
example is five workers. The fifth worker contributes just as much to revenue as
he or she costs and should be hired; beyond that level of employment, however,
additional employment would reduce profits.
The equilibrium condition for a firm’s labor demand can thus be stated as:
iG =MRPs W = MR: MB. (4.2)
For a competitive firm, Equation 4.2 can also be written as:
MR
W = P-MBP,or W/P
= MP,. (4.3)
Equation 4.2 states that the firm should continue to hire labor as long as the
marginal revenue product of each successive worker exceeds the money wage
rate; when the marginal revenue product of the last worker just equals the wage,
the optimal level of employment has been reached, and hiring should stop. For
the special case of a competitive firm, the product price P can be substituted for
the marginal revenue MR, as in Equation 4.3, and after rearranging terms, yields
an equivalent statement of the firm’s optimal employment level—hire workers as
long as the marginal product of labor is greater than or equal to the real wage
(W/P).
The Short-Run The equilibrium condition given in Equation 4.2 can be used to derive the firm’s
Demand Curve short-run demand curve for labor, as shown in Figure 4.3. The short-run labor
for Labor demand curve depicts the relationship between the wage rate and the firm’s
desired level of employment, holding capital and all other factors constant.
FIGURE 4.3 A Hypothetical Marginal Revenue Product Schedule and Labor Demand Curve
The marginal revenue
product schedule MAP, MRP,
shows the additional
revenue the firm obtains
from hiring each worker.
The fourth worker brings
in $8, the fifth worker $6,
and so on. The downward
sloping portion of the
MRAP, schedule is also the
firm's short-run demand
curve for labor. Thus, if the
wage W is $6, the firm
would maximize profit by
hiring five workers (point
A) since each of these
workers has a MAP, at
least as great as the wage.
At a wage of $8, the
optimal employment level is
four workers (point B).
0 1 2 3 4 5 6 ff Labor (L)
The horizontal axis of Figure 4.3 measures employment (L); the vertical axis
measures the wage rate per hour (W). The first step in deriving the demand
curve is to plot in Figure 4.3 the data on marginal revenue product (MRP, ) from
column 5 of Table 4.1. Since MRP, is measured in terms of dollars as are wages,
these data can be plotted in the same diagram. The line labeled MRP, is the
marginal revenue product schedule. It shows the marginal revenue product of
each worker: the fourth worker's MRP, is $8, the fifth workers MRP, is $6, and
sO on.
It turns out that the MRP, schedule in Figure 4.3 _is the firm’s demand _curve
for labor (D,). The equilibrium condition in Equation 4.2 says to hire “each
additional worker as long as the increase in revenue (MRP, ) is greater than or
equal to the increase in wage cost (W). At a wage of $6 per hour, how many
workers have an MRP, greater than or equal to $6 per hour? The answer is that
the first through fifth workers do and, thus, the optimal level of employment is
five (point A). Given the market wage rate, then, the MRP, curve can be used
to determine the firm’s demand for labor.’
A second example demonstrates this in a slightly different way. What would
happen to this firm’s demand for labor if the wage rate were to increase from $6
to $8 per hour? At $8 per hour it no longer pays to hire the fifth worker since the
marginal revenue product is less than the wage rate. Employment will be re-
duced to four workers, where the equality W = MRP, again holds (point B).
Two points should be stressed about the demand curve for labor. First, a
change in the wage rate causes a movement along one particular demand curve.
Thus, if the wage rate the firm has to pay for labor rises from $6 to $8 per hour,
the change in employment is represented graphically by moving up the demand
curve D, in Figure 4.3 from point A to point B; conversely, if the wage were to
decrease to $4, the firm would move down the demand curve to point C until
the marginal revenue product of the last worker hired was just equal to the wage
rate of $4. The second point is that a change in the firm’s demand for labor due
to any other reason will cause a shiftin the demand curve to the left or right.
Two different events illustrate this.
An Increase in Product Demand ~What will happen to the firm’s-demand for

labor if the demand for its product increases? An increase in product demand in
a competitive industry will cause the price of the product to rise in the market as
the product demand curve shifts to the right along a given upward sloping supply
curve. This immediately affects the demand for labor on the part of each indi-
vidual firm since price P (and thus marginal revenue MR in column 4 of Table
4.1) is part of the marginal revenue product calculation. At a higher price (say,
$3 in Table 4.1), the MRP, of each worker increases and, as shown in graph (a)
31 the wage were $12 in Figure 4.3, the firm would maximize profits by expanding employment to
three workers rather than just one, since the second worker's MRP, is greater than the wage rate. It is
only the downward sloping portion of the MRP, schedule, therefore, that is used by the firm as its
labor demand curve.
FIGURE 4.4 The Impact on the Labor Demand Curve of

Changes in Product Demand and Market Structure
Graph (a) shows that an increase in the firm's product demand will cause the labor demand curve to shift to the right from D, to
D». At the prevailing wage of W,, the firm’s optimal level of employment would increase from L, to Lp. Graph (b) shows that the
labor demand curve for an imperfectly competitive firm (D,) will lie to the left and be steeper than the demand curve for a
competitive firm (D,). At the prevailing wage of W,, the imperfectly competitive firm will hire fewer workers (L,) than the
competitive firm (L;). It will also expand employment by a smaller amount than would a perfectly competitive firm in response to a
decrease in the wage.
(a) An Increase in Product Demand (b) Perfect and Imperfect Competition in the
Product Market
1 Lo Labor (L) Lo L, Labor (L)
of Figure 4.4, results in a shift to the right in the demand curve from D, to D).
At the prevailing wage of W,, employment can be expanded from L, (point A)
to L, (point B) since, at the higher product price, the MRP, of additional work-
ers hired beyond L, will now exceed the wage cost W, until employment reaches
L,. A decrease in product demand would cause the labor demand curve to shift
to the left for the opposite reasons.
Imperfect Competition in the Product Market The data in Table 4.1 and the
derivation of the demand curve for labor in Figure 4.3 assume the firm is only
one among many firms operating in a perfectly competitive product market.
What happens to the demand curve for labor if this firm is in an imperfectly
competitive product market (such as monopolistic competition, oligopoly, or
monopoly)?
A competitive firm has a horizontal product demand curve and can sell all its
output at a constant price. Thisisreflected in Table 4.1 by the same price and
marginal revenue of $2 per unit in column 4 no matter how much the firm
produces. An imperfectly competitive firm, however, faces a downward slopin
product demand curve, implying that to sell more output it has to lower, the
price. In this situation, the marginal revenue from selling a unit of output is less
than the price it is sold for.* Equally important, if the imperfectly competitive
firm desires to sell additional units of output, it has to further lower the>price,
causing marginal revenue to decline further. P>m
The fact that price is greater than marginal revenue for the faeicety com-
petitive firm has two separate effects on its demand curve for labor. First, since P
is always greater than MR, the marginal revenue product (MR: MP,) of the last
worker hired in a competitive firm (worker L, in graph b) will be more than that
of the same worker in an imperfectly competitive firm. This would cause the
labor demand curve for the imperfectly competitive firm, shown as D, in graph
(b) of Figure 4.4, to lie to the lefeof that of the competitive firm, D,, at the
prevailing wage. At W,, therefore, employment in the imperfectly competitive
firm would only be L; (point B) compared to L, (point A) for the competitive
firm. The second effect of imperfect competition in the product market is to not
only reduce the demand for labor but to also make the demand curve D, steeper
relative to D,.’ The reason for this is that for the imperfectly competitive firm its
MRP, schedule declines for two reasons, first, because the MP, of each worker
declines due to diminish
minishing returns in production, as for the competitive firm
and, second, because its marginal revenue from extra production also declines,
unlike that of the competitive firm. “, Ly
The Market For many purposes it is important to know not only the individual firm’s demand
Demand Curve curve for labor, but also the demand for labor on the part of all firms in the labor
for Labor market. The labor market in question may encompass a local area, an industry,
or the nation as a whole, depending on the type of labor being analyzed. The
derivation of the market demand curve for labor is illustrated in Figure 4.5.
It is assumed there are_N individual perfectly competitive firms in an industry.
Graphs (a) and (b) show the labor demand curves D, and Dg, for two of these
firms, Firm A and Firm B. At the wage of W,, the demand for labor in each
individual firm can be determined from its respective marginal revenue product
schedule; employment in Firm A would be L,,, in Firmava B it be L,,, and
so on for the other N — 2 firms. The total demand for labor by all the firms in the
market could then be found by_adding up the employment demand of each indi-
vidual firm—lL,, + Ej, +... + Biy = Liy, shown as point X in graph (c).
If the wage were to fall to W,, it would be tempting to repeat the same exer-
cise. Employment in Firm A, given the demand curve D,, would be L,,4, em-
ployment in Firm B would be L;,, and so on for the other firms, yielding a total
market demand for labor of L,,4 + Ly, +... Liy = L2y, shown as point Y in
graph (c). Connecting points X and Y would, presumably, trace out the market
demand curve for labor. Unfortunately, however, this reflects an error in reason-
ing. It was implicitly assumed in deriving point Y that the price of the product
4 As an example, assume a firm can sell 10 units at $10 apiece, yielding a total revenue of $100, or
by lowering the price to $9, it can sell 12 units, bringing in $108. Even though the price of the 12th
’ unit is $9, its marginal revenue is only $4, (i.e., MR =ATR/AQ= $8/2= $4); thus, MR < P.
5A more precise meaning of “steeper” is more “inelastic,” a term to be defined shortly.
FIGURE 4.5 Derivation of the Market Labor Demand Curve

The lines D, and Dg in graphs (a) and (b) show the labor demand curve for each individual firm. At the wage W,, total employment in
the market—graph (c)—is L,4 + Lig +... + Liw= Lim (point X). A fall in the wage from W, to W, in any one firm would cause a
movement down its MAP, schedule, increasing employment from L;, to Lz, in Firm A, for example. If the wage falls to W, for a//
firms, however, the resulting expansion in employment and output will cause a decline in the price of the product, shifting each firm's
MRP schedule to the left, such as D4 and Dj. At the wage W,, therefore, total labor demand is L34 + L3g +... + Liv = Loy (point
Z). Connecting points X and Z yields the market labor demand curve.
(a) Firm A (b) Firm B (c) The Market

Ww Ww
Ww, W,
We W,
Lia Log Lor L Liglog Lop L Li Low Low b
and, thus, the marginal revenue of each additional unit of output are gonstant.
This is a correct assumption for an individual competitive firm since its level of
output is so small relative to the entire product market that it can expand or
contract employment and output with no effect on the product price. What is
true for an individual firm will not be true, however, if the wage drops to W, for
all N firms. As all N firms expand employment, total output in the product mar-
ket will increase, and to sell this extra output the price will have to fall. As the
price falls, the marginal revenue product of each worker is reduced, causing the
MRP, schedule of each individual firm to shift to the left, such as from D, and
D, to Di and Dg. Consequently, given a decline in the wage to W, in the entire
market, the total market demand for labor is L}, + Ljg +... + Liy = Liy,,
shown as point Z in graph (c). Connecting points X and Z yields the true market
demand curve Dy.
This analysis shows that wages and employment are inversely related at the
level of both the individual firm and the market. An important insight, how-
ever, is that a wage change isolated to only one firm will have a proportionately
bigger impact on employment than if the wage change occurs in all the firms in
the labor market.
Criticisms of the The marginal productivity theory of labor demand outlined above can predict
Theory of Labor both the optimal level of employment for a firm and how this employment level
Demand will change in reaction to various economic events. The theory has, neverthe-
less, been the subject of much controversy and criticism through the years, espe-
cially by more institutionally oriented economists. The critics contend that it
Chapter 4. The Demand for Labor in the Short Run tose
rests on a number of assumptions that are unrealistic or incorrect. Five of these

criticisms are briefly discussed here.
Limits to Human Cognition One objection is that the information and com-
putational requirements that are necessary to operationalize the theory generally
exceed the mental or cognitive ability of most employers and managers.° For ex-
ample, while the concept of marginal productivity is clear in theory, can the
typical manager of a firm actually measure or even approximate the likely in-
crease in production from hiring a new employee? For small businesses or firms
with fairly simple production processes, the answer may be yes, but the dif-
ficulties, in the critics view, of measuring the marginal product in large-scale
organizations with highly interdependent production processes preclude the use
of marginal calculations. Even if a marginal product can be calculated, the critics
would then ask how the firm can meaningfully compare a worker's marginal reve-
nue product with the wage, when product prices and output levels (and thus the
MRP,) are constantly changing. Finally, the critics argue that case studies and
interviews reveal that the actual process of business decision making does not
follow the marginal calculations assumed in the theory.’
Proponents of the theory respond to these criticisms in several ways. First,
although managers may not consciously use marginal calculations, the decisions
they reach must approximate those predicted by the theory if the firm is to sur-
vive in a competitive business world. Second, proponents argue that while man-
agers may not be able to identify the marginal contribution of an individual
worker or adjust employment one employee at a time as the theory presumes,
they can identify the revenues and costs associated with particular lines of ac-
tivity (e.g., the baggage handling “department or the night shift) and they do
adjust employment in these activities in light of their contributions to profit.®
Nonmaximizing Behavior A second criticism of the marginal productivity

theory deals with the assumption of profit maximization. The impetus driving
the firm to make the calculations of marginal revenue product versus marginal
cost of labor is the goal of maximizing profits. The critics of the theory argue,
however, that business firms, particularly in oligopolistic markets or in corpora-
tions where ownership and control are separated, are more accurately character-
ized as satisficing with respect to profit.’ In this view, the managers of business
firms strive to achieve a minimum level of profit in order to protect their survival
and that of the firm; once this minimum goal is satisfied, the managers tolerate
some “slack” in the operation of the firm and no longer strive to fully minimize
Richard Lester, “Shortcomings of Marginal Analysis for Wage-Employment Problems,” Ameri-

can Economic Review 36, no. 1 (March 1946): 63-82.
7Herbert Simon, “Rational Decision Making in Business Organizations,” American Economic
Review 69, no. 4 (September 1979): 493-513.
8Fritz Machlup, “Marginal Analysis and Empirical Research,” American Economic Review 36,
no. 4 (September 1946): 519-541.
°Simon, “Rational Decision Making in Business Organizations.”
production cost. One important way firms exhibit satisficing behavior, the critics
say, is by employi people than are really needed, as evidenced by the
more ng
bloated ranks of middle management and the inefficient organization of produc-
tion on the factory floor in some large American companies. (See the accom-
panying In The News section.)
The defenders of marginal productivity theory would probably admit that as
a matter of realism, other goals besides maximizing profits enter into business
decision making. They would argue, however, that the ultimate goal of each
firm is survival and that survival in a competitive economy requires keeping
costs and employment down and profits up. Thus, while it may not be literally
true that management squeezes out of the business every last dollar of profit pos-
sible, the pressure on the firm to survive is sufficiently strong that all other goals,
in the long run, are subordinated to the single goal of maximizing profits. The
necessity of making a profit, in turn, assures that business firms will be induced
to economize on labor and make hiring decisions that broadly accord with the
predictions of the marginal productivity theory. The defenders would argue that
this is true even for regulated firms or not-for-profit organizations, since their
survival and growth requires conscious efforts to minimize cost.
Fixed Capital/Labor Proportions A third criticism of the theory asserts that
the nature of technology makes it impossible to derive a continuous marginal
product (MP, ) schedule as drawn in Figure 4.2.'° The derivation of the marginal
product schedule assumes that the fixed stock of capital is divisible in the sense
that it can be “spread” among greater and greater numbers of workers as employ-
ment is increased. According to critics of the theory, however, many types of
production processes require labor and capital in relatively fixed proportions.
One hypothetical example is a small commuter airline with, say, three planes,
each of which requires two pilots. Given the stock of capital of three planes, is it
possible to calculate the marginal product of each pilot? The marginal product of
the first pilot would be zero, since with only one pilot no plane could fly. With a
second pilot, one plane could fly, and both pilots together would yield a positive
increment in production. The addition of yet a third pilot, however, would not
lead to any further increase in production since the second plane could not fly.
The result is that it is impossible to either attribute a unique marginal product to
each individual worker or to derive a continuous marginal product schedule.
The proponents of the marginal productivity theory discount the seriousness
of these objections. They contend that in nearly all real world production situa-
tions there is no fixed, necessary relationship between capital and labor. In the
example above, one pilot could, it would be argued, fly the plane if she or he had
to. The first pilot would have a nonzero marginal revenue product, the second
pilot would also have an identifiable marginal revenue product—the increase in
safety and efficiency with which the flight is operated. The proponents of the
theory would also argue that more often than not, fixed capital/labor require-
Alan M. Cartter, Theory of Wages and Employment (Homewood, IIl.: Irwin, 1959): 45-47.
ments stem from union work rules or regulatory constraints (such as FAA re-
quirements concerning flight crew size), rather than from technology itself. In
any case, the proponents of the theory argue that employers usually are able to
attribute to each individual worker the net contribution to production.
Increasing Returns to Labor A fourth criticism of the marginal productivity

theory is that labor may be subject to increasing returns in the short run, not
diminishing returns as the theory assumes. According to the law of diminishing
returns, as a firm expands employment, the marginal product of labor should
decline as the fixed stock of capital is spread over more workers. It is precisely
this fact that causes the MRP, schedule and the labor demand curve to slope
downward for a competitive firm. Numerous empirical studies have found, how-
ever, that labor productivity (output per hour) actually varies directly with the
level of employment in the firm—when employment rises on a business-cycle
upswing, output rises more than proportionately.'! Since labor productivity is
nothing but the average product of labor, Q/L, some economists have reasoned
that if the average product rises with increased employment this suggests that
the marginal product of labor, AQ/AL, also increases, implying that the mar-
ginal revenue product schedule is actually upward sloping.'? (Remember from
principles of economics that when the marginal rises it pulls up the average.) If
this result is true, the critics ask, doesn’t it invalidate one of the marginal pro-
ductivity theory’s basic assumptions?
Proponents of the theory have attempted to account for the paradox of in-
creasing returns in several ways. One is to argue that it is a statistical illusion
caused by faulty data on the level of production. During a recession, for ex-
ample, firms may have employees work on deferred maintenance projects or the
completion of unfinished assemblies. Since these activities are not counted as
part of output, their omission causes measured labor productivity to vary pro-
cyclically, in apparent conflict with the predictions of marginal productivity the-
ory. Second, proponents also note that the marginal productivity theory assumes
workers put in a constant level of effort, while in reality effort is likely to slacken
in bad times and increase in good times. This fact also makes labor productivity
behave procyclically. Finally, defenders of the theory argue that the paradox of
increasing returns is due in part to the tendency of firms to “hoard” labor during
recessions by consciously keeping more workers than they really need given the
current level of production. As explained later in this chapter, firms are moti-
vated to do so by their desire to avoid additional hiring and training costs, fac-
tors the simple version of the marginal productivity theory overlooks.
Interdependence between the Wage and Worker Productivity A final criti-
cism of the marginal productivity theory concerns the relationship between the
wage rate the firm pays and the level of productivity of its workers. The theory
See Jon A. Fay and James L. Medoff, “Labor and Output over the Business Cycle: Some Direct
Evidence,” American Economic Review 75 (September 1985): 638-655.
2See Arthur M. Okun, “Inflation: Its Mechanics and Welfare Costs,” Brookings Papers on Eco-
nomic Activity (Washington: Brookings Institution, 1975): 378.
IN « THE « NEWS
Was General Motors Off Its Labor Demand Curve?
As noted above, critics of the marginal productiv- or so grievances were outstanding—about one per
ity theory argue that the level of employment in employee. Wildcat strikes interrupted production,
many firms is greater than what would be predicted and labor and management feuded like old-time
by the theory because these firms do not maximize enemies.
profits. Supporters of the theory claim, however, Due to the sharp drop-off in auto sales that ac-
that this objection is largely misplaced since the companied both the recession of 1981-1982 and
pressures of competition will force firms to main- the onslaught of foreign competition, General
tain the profit-maximizing level of employment Motors decided to close the Fremont plant and
if they are to survive. Which side is right? The place all of the employees on indefinite layoff. In
following example from an article in Business early 1983, GM turned the idle plant over to
Week suggests that both arguments may have some Toyota Motor Corp., as part of a joint venture
validity. called New United Motor Manufacturing Inc.
Until 1982, General Motors operated an auto (NUMMI). The new Japanese management added
assembly plant in Fremont, California. In a typical little new technology or equipment to the plant,
year, the plant produced approximately 250,000 but did substantially reorganize the production
automobiles with an employment level of about process and personnel practices. The company
5,000 workers. According to the article, the plant also recognized the United Automobile Workers
was a case study of American industry in decline. union and hired back many of the union’s mem-
The absentee rate hovered at 20 percent and 5,000 bers, including the local’s militant leaders.
assumes that the marginal product schedule is determined solely by the technol-
ogy of production, as represented by the production function in Equation 4.1,
and the size of the firm’s capital stock. One important implication of this as-
sumption is that, regardless of whether the firm pays $2 or $20 per hour for labor,
the MRP, schedule will maintain its position and shape; that is, a change in the
wage will cause the firm to move up or down the MRP, schedule, but the sched-
ule itself will not change. Critics of the theory argue, however, that in real life
the level of worker productivity is likely to be directly related to the wage the
firm pays.'’ One reason is that higher pay allows workers to improve their physi- ‘
cal ability to work through improved nutrition and health, a second is that a
wage increase is likely to stimulate greater work effort and higher morale among
employees. The implication, then, is that there is a separate MRP, schedule for
each level of the wage—the higher the wage, the greater will be each employee’s
work effort and marginal product and the further to the right will lie the entire
MRP, schedule. Because of this, whether employment will increase or decline in
response to a wage change can no longer be unambiguously predicted. A higher
"See George Akerloff and Janet Yellen, Efficiency Wage Models of the Labor Market (Cambridge,
England: Cambridge University Press, 1986).
rc
a a
The result? According to the article, “At Fre- have continued to run the plant in such an inef-
mont today, 18 months after the first Chevrolet ficient manner, in effect sacrificing profits for
Nova rolled off the line, it looks like the frog has the stability and psychological comfort of doing
turned into a prince.” The new Japanese manage- things the “old way.” As the supporters of mar-
ment succeeded in raising productivity dramati- ginal productivity theory point out, however, in a
cally as 2,500 workers produced almost the same competitive marketplace firms cannot sacrifice
annual output that 5,000 workers had under GM. profits forever; eventually they will be driven out
Similarly, at the time the article was written only of business. In essence, this process was at work in
two grievances were outstanding and the absentee- the American automobile industry, as reflected by
ism rate was under 2 percent. In sum, NUMMI GM's decision to close the Fremont plant. The
took the same plant, paid nearly the same wages, Japanese, using better management practices,
and produced about the same number of auto- were able to significantly reduce employment,
mobiles with only half as many workers. thereby moving back to the plant’s marginal reve-
The question, then, is how this set of events nue product schedule. In doing so, they also made
can be explained by the marginal productivity a lot more profit than GM.
theory? One explanation is that GM had placed
itself far to the right of the plant’s marginal reve-
nue product schedule; at the prevailing wage the
number of workers hired by GM greatly exceeded
the profit-maximizing level. Without competition Source: “The Difference Japanese Management Makes,” Busi-
from lower-cost rivals, GM could quite possibly ness Week (July 14, 1986): 47-50.
wage will cause the firm to reduce employment due to the higher cost per unit of
labor, counterbalanced by an increase in employment because each worker be-
comes more productive.'* It is conceivable, that a wage increase could actually
lead the firm to hire more labor, not less, if the higher wage stimulates a suff-
ciently large increase in worker productivity.
Proponents of the marginal productivity theory respond to this criticism in
several ways. First, they admit that, realistically, it is probably true that the level
of worker productivity depends to some degree on the rate of pay. This is likely
to be most true in underdeveloped countries, however, where higher pay could
significantly improve nutrition and health conditions of workers. In developed
economies, on the other hand, modest adjustments in wage rates will probably
have small to negligible impacts on employee work performance, particularly at
the industry or national level. Proponents of the theory also note that there is
considerable disagreement among specialists in organizational behavior over the
extent to which pay really influences performance at the firm level. Some studies
find a positive relationship between wages and employee performance, others
‘For a diagrammatic analysis, see Richard Perlman, Labor Theory (New York: John Wiley and
Sons, 1969): 50-56.
find no relationship, and yet others find a negative relationship.'’ These results
suggest, in the view of the theory’s supporters, that it is safe to continue assum-
ing that labor demand curves slope downward, just as marginal productivity the-
ory predicts.
Conclusion Despite these criticisms, the great majority of economists, at
least in the United States, continue to subscribe to the marginal productivity
theory for several reasons. First, most economists remain convinced that the ob-
jections outlined above do not invalidate marginal productivity theory. They ar-
gue that some objections (e.g., limited cognition) have little substantive merit,
while others (e.g., increasing returns) can be adequately handled by expanding
the theory to take into account additional real-world complications that the
simple version of the theory neglects. Also, most economists believe the theory’s
basic predictions are supported by the available evidence. It does seem to be
true, as the theory predicts, that as the cost of labor rises, firms react by cutting
employment. Finally, most economists are reluctant to throw out an existing
theory until a better one is developed to take its place. So far the critics of mar-
ginal productivity theory have not succeeded in this task.
THE ELASTICITY The downward slope of the labor demand curve shows that the wage rate and
OF DEMAND level ofemployment are inversely related. For many issues, however, it is impor-
FOR LABOR tant to know more than this. In particular, economists and policymakers need to
know how sensitive employment is to changes in the cost of labor. This involves
the concept of the elasticity of labor demand.
The elasticity of labor demand (E,,) is defined as:
VAL SnPogmead
Ep = AW’ (4.4)
where %AL is the percentage change in employment and %AW is the percent-
age change in the wage rate. Since the wage rate and the firm’s desired level of
employment are inversely related, the elasticity of labor demand must be a nega-
tive (or at least a nonpositive) number. For convenience, however, economists
usually neglect the minus sign in discussing the elasticity of demand, a practice
followed here.
The more responsive is the firm’s demand for labor to changes in the wage °
rate, the greater will be the numerical value of the elasticity of demand. It is
possible to distinguish five different cases. Each case is defined below and is illus-
trated in Figure 4.6, graphs (a) and (b).
1. E, =O. Ifan increase in the market wage rate of, say, 10 percent (%AW =
10%) causes no change in the demand for labor (%AL = 0), the elasticity
of demand will equal zero. The demand for labor is known in this case as
perfectly inelastic and is represented by the vertical demand curve in Figure
4.6, graph (a).
See Andrew Szilagy and Marc Wallace, Organizational Behavior and Performance, 3d ed. (Glen-
view, IIl.: Scott, Foresman, 1983): 415.
FIGURE 4.6 The Five Categories of Demand Elasticity

There are five different classifications of demand elasticity. Graph (a) illustrates labor demand curves that are perfectly
inelastic
(Ep = 0), perfectly elastic (E, = cc) and unit elastic (Ep = 1). Graph (b) illustrates labor demand curves that are inelastic (Ep < 1)
and elastic (Ep > 1).
oe
Labor (L) oo Labor (L)
Ep < 1. If the elasticity of demand is less than one but greater than zero,
demand is called inelastic. This would be the case, for example, ifa10 per-
cent reduction in the wage rate led to only a 5 percent increase in labor
demand (Ep, = .5). An inelastic demand curve implies that labor demand is
relatively insensitive to labor cost, a notion that is represented graphically
by the steep demand curve drawn in graph (b).
Ep = 1. If the percentage change in labor demand is just equal to the per-
centage change in the wage rate, then Ep = 1, and demand is known as unit
elastic. A unit elastic demand curve is shown in graph (a) of Figure 4.6. It has
the shape of a rectangular hyperbola, for reasons that are explained below.
Ep > 1. If the percentage change in labor demand exceeds the percentage
change in the wage rate, the elasticity of demand will be greater than one,
and demand is elastic. An elastic demand curve implies that labor demand is
highly responsive to changes in the wage rate, indicated in graph (b) by the
flat demand curve.
Ep = &. If the firm is willing to hire all the additional employees it can at
the prevailing wage, but will hire no employees at any higher wage, demand
is known as perfectly elastic. A perfectly elastic demand curve is illustrated
by the horizontal line in graph (a).
Additional Considerations There are several important points to note about
the elasticity of demand. The first is rather technical in nature and concerns the
relationship between elasticity and the slope of a linear demand curve. While a
flat demand curve such as the one depicted in graph (b) is said to be elastic,
strictly speaking, this is incorrect—any straight-line, downward-sloping demand
curve will have both an elastic and an inelastic portion, as shown in Figure 4.7.
FIGURE 4.7. The Change in Elasticity over a Linear Downward Sloping Demand Curve
Although steep demand
curves are usually labeled
inelastic and flat demand
curves are labeled elastic,
technically this is incorrect.
Along any straight-line,
downward-sloping demand Elastic
curve, the top half is
elastic, the midpoint is unit
elastic, and the bottom half
is inelastic.
Unit elastic
Inelastic Oe
$2
$1
20 30 80 90 Labor (L)
If the wage should fall from $8 to $7, for example, employment would rise from
20 to 30 workers. What is the elasticity of demand? Ep = 50%/—12.5% = 4
(neglecting the minus sign); demand is very elastic. If the wage begins at $2, and
falls by a dollar to $1, employment will rise from 80 to 90 workers. The elasticity
of demand is E, = 12.5%/—100% = .125; demand is quite inelastic.
For the same $1 change in the wage, demand on the top half of the curve is
elastic, on the bottom half it is inelastic, and at the midpoint it is unit elastic.
The reason is that elasticity is measured in terms of percentage changes, while a
linear curve shows absolute changes in wages and employment." The same dollar
decrease in the wage is a small percentage change at $8 but a large percentage
change at $2, thus leading to different elasticity estimates along one straight-line
demand curve. For a demand curve to have a constant elasticity along every
point, it has to be a curved line such as the unit elastic demand curve in graph
(b) of Figure 4.6. Given this caveat, for ease of exposition flat demand curves are —
nee to as elastic and steep demand curves as inelastic in the remainder of
this book. This convention is extremely convenient and does not affect any
major conclusion concerning the theory of labor demand.
'°To avoid this problem, in empirical work labor demand curves are usually estimated in double
logarithmic form: In L = In a + b In W. The coefficient b yields the elasticity of demand since
nal
iG, Jae AL
d(InW) dW %AW’
Ww
A second important consideration with respect to the elasticity of demand

concerns the firm’s wage bill. The wage bill is defined as the firm’s total money
outlay for labor. It is calculated by multiplying the average wage rate per hour
times the total number of workers Eocene ont Seana
demand is inelastic, a rise in the wage rate will result in a net increase in the
total wage bill since the percentage increase in the wage W will exceed the per-
centage decline in employment L, making the product W - L greater than be-
fore. The opposite holds true for a wage decrease. Conversely, if labor demand is
elastic, a rise in the wage rate will cause the wage bill paid for labor to decline
since the percentage decrease in employment will be greater than the percentage
increase in the wage rate.
A third consideration is what determines the elasticity of demand. Why is it
that in some cases labor demand is elastic while in other situations it is inelastic?
The key consideration is the ease of substitution available both to consumers in
their choice of what firm to buy from and to business firms in their choice of
labor, capital, and other inputs to use in producing the product. A rise in the
wage rate in the market raises the costs of production for firms, leading to a re-
duction in employment for two reasons. First, firms will be forced to raise their
product price, resulting in fewer sales, less productio , and a smaller demand for
labor. Second, higher costs of labor will motivate firms to substitute capital and
other inputs for labor in the long run, leading to a further reduction in labor
demand. Thus, the easier it is for consumers to find a substitute good (such as an
imported car) to replace the one whose price has gone up (such as a domestic
car), or alternatively, the easier it is for the firm to substitute capital for labor,
the greater will be the reduction in employment in those firms where wages have
increased—that is, the more elastic will be the demand for labor. This subject is
treated in much more detail in the next chapter.
Estimates of The marginal productivity theory predicts that the demand curve for labor
the Elasticity slopes downward to the right. The most fundamental implication of this predic-
of Labor Demand tion is that wage rates and levels of employment are inversely related—the
higher the wage rate, other things equal, the lower will be the level of employ-
ment in the firm or industry. >
Economists have conducted numerous empirical studies to test this prediction
of the theory. Daniel Hamermesh surveyed much of this literature and arrived at
several conclusions. '’ Every study finds that, as predicted, the demand curve for
labor slopes downward to the right. Whether measured across occupations, in-
dustries or demographic groups, higher wages lead to less employment, holding
all other things constant. A second conclusion from Hamermesh’s survey is that,
in general, demand curves. in the economy are inelastic. Based on the estimates
in individual studies, Hamermesh calculated a consensus estimate for the elas-
17Daniel Hamermesh, “Econometric Studies of Labor Demand and Their Application to Policy
Analysis,” Journal of Human Resources 11, no. 4 (Fall 1976): 507-525. For an update of this study,
see his article “The Demand for Labor in the Long Run,” in Orley Ashenfelter and Richard Layard,
eds., Handbook of Labor Economics (Amsterdam: North-Holland, 1986): 429-471.
ticity of demand of —.32, meaning that for every 1 percent increase in the wage
rate, labor demand would decline by three-tenths of 1 percent. Hamermesh then
decomposed this estimated reduction in labor demand into that part due to the
decline in the level of production and that part due to the substitution of capital
for labor. It was found both effects contributed nearly equally to the total decline
in labor demand.
A related conclusion concerning the elasticity of demand was that it increases
on average as the particular type of employment becomes narrower or more dis-
aggregated. Thus, the elasticity of labor demand for manufacturing was generally
estimated in studies as larger than for the entire economy but still inelastic,
while for a demographic group such as teenagers it was much higher, possibly
greater than 1.0 and, thus, elastic. This result reflects the fact that the narrower
the type of employment and product considered, the greater the substitution
possibilities both in production and consumption.
Finally, evidence suggests that the elasticity of demand is greater for unskilled_. _
and blue-collar workers than for skilled and white-collar workers. '* The reason is
that capital and blue-collar or unskilled workers are.generally substitutes for each
other in production, while capital and skilled or white-collar workers tend to be
complements. A good example is television manufacturing. In response to rising
wages and greater import competition, television manufacturers automated the
production process in order to reduce labor cost. The workers most adversely
affected were unskilled assemblers whose jobs were largely displaced by computer-
controlled automatic inserting equipment. Employment among more skilled ma-
chine operators declined only slightly and actually increased for other skilled
workers such as computer programmers and electronic technicians.”
Empirical Evidence
Union Wage Concessions in the Auto Industry
One of the clearest examples in recent years that labor demand curves have a
negative slope is the series of wage concessions agreed to by unions in the auto,
rubber, steel, trucking, and airline industries, among others. In each case, wage
concessions were made in the face of events or developments that caused the
demand curve for union, labor to become much more.elastic, threatening large ‘
losses in employment and union membership if there was not wage relief.
Consider, for example, the auto industry.’*? Wage concessions were made first
by the United Automobile Workers (UAW) with Chrysler in 1979, followed by
‘Daniel Hamermesh and James Grant, “Econometric Studies of Labor-Labor Substitution and
Their Implications for Policy,” Journal of Human Resources 14, no. 4 (Fall 1979): 518-542.
"See Bureau of Labor Statistics, Technology and Labor in Four Industries, Bulletin 2104 (Washing-
ton: G.P.O., 1982), 34-44.
This discussion is drawn from Bruce E. Kaufman and Jorge Martinez-Vasquez, “Voting for Wage
Concessions: The Case of the 1982 GM—UAW Negotiations,” Industrial and Labor Relations Review
41 (January 1988): 183-194. Also see Harry Katz, Shifting Gears: Changing Labor Relations in the
U.S. Automobile Industry (Cambridge, Mass.: MIT Press, 1985).
cee
concessions with Ford and General Motors (GM) in 1982. During the 1970s
wage gains in the auto industry had substantially exceeded the average rate of
wage increase in the economy in general, rising 114 percent between 1970 and
1979, compared to 91 percent for all nonagricultural industries. Why all of a
sudden did the UAW have to make record-breaking “give-backs” after a decade
of above average wage gains? For a variety of reasons the demand curve for labor
in the auto industry during the 1970s was fairly inelastic. One factor that con-
tributed to this situation was that theUAW had organized all the major pro-
ducers in the industry. Thus as wages and auto prices went up, there were no
domestically produced, lower-priced, nonunion-made automobiles that consum-
ers could substitute toward. Imported cars were, of course, a threat to the UAW,
but until 1979 the dominant component of auto sales were the “gas guzzlers,”
which only Detroit produced and whose sales, therefore, were not as price sen-
sitive as smaller compact cars.
As a result of these factors, the UAW was able to raise wages in the auto
industry by moving up the relatively inelastic industry labor demand curve.
Had demand been elastic, the resulting employment losses from raising wages
114 percent would have been severe enough to cause the union to moderate its
wage gains. As it turned out, employment actually grew 23 percentin the auto
industry between 1970 and 1979, reflecting the net effect of some job loss from
movements up an inelastic labor demand curve coupled with job gains as the
demand curve shifted to the right due to greater auto sales.”
Several developments occurred in the auto market in 1979 through 1982
however, that fundamentally altered the economic environment confronting
both the companies and the union. The most important were increased gas
prices and outright gas shortages that occurred in the aftermath of the revolution
in
Iran. The result was that consumer_demand disappeared for the large cars in
which Detroit specialized and boomed for small, fuel-efficient cars. In response,
the domestic auto companies switched the majority of their production to a seg-
ment of the market where they faced much more competition from imports,
making labor demand much more sensitive to wage costs and thus much more
elastic. Compounding the woes of the UAW, the entire labor demand curve
shifted sharply to the left due to the sharp drop in auto sales brought on by
record high interest rates and two recessions.
A Graphic Analysis The impact of these developments on the demand curve
for union labor in the auto industry is illustrated in Figure 4.8. The steep de-
mand curve D, illustrates the inelastic demand curve facing the UAW up to
1979. At the average wage rate of $9.07 per hour in the auto industry in 1979,
employment was 759,000, shown as point A. As of April 1982 (the date of the
contract concession vote), wages in the auto industry had risen to $11.88. Had
the demand curve remained D,, employment would have declined moderately
1Sales of domestic cars rose from 7.2 million units in 1970 to 9.0 million in 1978 and then
plummeted to 5.5 million in 1982. Imports were 14.7 percent of total sales in 1970, 17.8 percent in
1978, and 29.3 percent in 1982.
FIGURE 4.8 The Change in Labor Demand in the Auto Industry, 1979-1982
Between 1979 and April
1982, average hourly Auto industry
earnings in the auto wage
industry rose from $9.07 to
$11.88. Had the demand
curve remained inelastic,
such as D,, employment
would have declined
modestly from point A to B.
After 1979, however, the $11.88
labor demand curve
became much more elastic
and also shifted to the left,
as to D3. The result was a
substantial decline in
$ 9.07
employment, from point
A to E.
525 759 Auto employment

(in thousands)
to point B. Because of the developments outlined above, however, the labor

demand curve in the auto industry became much more elastic, represented by
the rotation of D, to the left to D;. Now at the wage of $11.88 the decline in
employment is significantly greater, from point A to point C. Coupled with the
greater elasticity of demand was also the leftward shift in the demand curve from
D, to D,; resulting from the sharp drop in unit sales. The combination of the
greater elasticity and leftward shift resulted in quite a sizable drop in employ-
ment from 759,000 in 1979 (point A) to 525,000 in April 1982 (point E).
Do labor demand curves slope downward? In April 1982 the membership of
the UAW voted by a slim majority to forgo scheduled cost of living increases,
the “annual improvement factor” (a deferred wage increase agreed to take effect
in future years), and nine personal paid holidays. The total savings in labor cost
to the company was estimated at 10 percent. Without these concessions, wages
would have continued to rise, moving the union further up the elastic demand
curve D, above point E. The result would have been even more layoffs and plant
closings. In agreeing to give up several billion dollars of wages and other compen-
sation, the UAW was testifying to the inverse relationship Ressen Wangsand
employment. As the headline of the UAW newspaper announced in explaining
ioe package to the membership, “Contract to Save Thousands of
Jobs.””
* United Automobile Workers, UAW—GM Report (March 1982): 1.

esse
Policy Application
Wage Subsidy Programs
Even in the best of times a significant number of persons face great difficulty in
finding employment, persons sometimes identified as the structural or “hard-
core” unemployed. Prominent among the hard-core unemployed are people in
three problem groups: low-skilled adults, disadvantaged youth, and residents of
economically depressed areas. The most glaring example of hard-core unemploy-
ment in today’s labor market is black teenagers, a group whose unemployment rate
in 1987 (33.2 percent) was nearly six times the national average (6.1 percent).
Why can’t the hard-core unemployed find work? One answer is a combination
of discrimination and a shortage of jobs in the economy. A number of econo-
mists argue, however, that these factors are of secondary importance; the more
important explanation is that firms simply do not find it profitable to hire these ¢
people. The fundamental proposition of the marginal productivity theory is that
a profit maximizing firm will hire a worker only if the extra revenue he or she
produces exceeds the wage the firm has to pay. Why would a firm not find it
profitable to hire the hard-core unemployed? One reason is that these workers
have a very low level of productivity. Many disadvantaged youths, for example,
are high school dropouts who have neither the basic educational skills nor the
good work habits that are necessary in today’s job market. The second reason, in
this view, is that the minimum wage law prevents employers from lowering their
rate of pay enough to make it profitable to hire the hard-core unemployed.
From the perspective of marginal productivity theory, therefore, policymakers
have two alternative ways to attack the problem of hard-core unemployment.
One is manpower training programs such as the Job Corps (to be discussed in
Chapter 13) that provide the hard-to-employ with specific job skills. A second
route is to lower the cost to the firm of hiring the hard-core unemployed. One
option advocated by a number of economists to reduce black teenage unemploy-
ment, for example, is a teenage. subminimum wage (to be discussed in Chapter
6). A second option is a wage subsidy program. The next section analyzes this
type of program in more detail.
~
The Employment A wage subsidy program aims to create additional jobs in the economy by subsidiz-
Effect of a Wage ing the employer's cost of labor. Unlike public-service—type programs that create
Subsidy Program new jobs through expanded government employment, wage subsidy programs
focus on employment in the private sector. A variety of wage subsidy programs
have either been adopted or proposed. In recent years, income tax credits have
been used to subsidize the wages of the handicapped, welfare recipients, ex-
convicts, disadvantaged youths, and Vietnam veterans. As a specific example,
Congress passed in 1971 and extended in 1975 a “work incentive tax credit”
(WIN) program that provided the employer a 20-percent tax credit up5to a limit
of $1,000 on the annual earnings of each recipient of AFDC who was hired.”
3 A discussion ofspecific wage subsidy programs and alternative program features is given in Daniel
Hamermesh, “Subsidies for Jobs in the Private Sector,” in John L. Palmer, ed., Creating Jobs: Public
Employment Programs and Wage Subsidies (Washington: The Brookings Institution, 1978): 87— We,
FIGURE 4.9 The Effect of a Wage Subsidy Program on Labor Demand

A wage subsidy program gives rise to the kinked labor demand curve D,7D. At the wage W,, employment increases from L, to L;
in graph (a). The cost of the wage subsidy program is shown by the rectangle in graph (b). The more inelastic the presubsidy
demand curve, the larger the size of the subsidy payment S must be to stimulate an increase In employment of L, — Ly.
(a) (b)
Wage Wage
(W) (W)
W,
Ww,
We
L, Lo Ly L» Labor (L)
The employment effect of a wage subsidy program is illustrated in Figure 4.9.

Shown in graph (a) as D, D, is the “before-subsidy” demand curve.** This curve
shows that at a market wage of W,, in lieu of any subsidy, a typical firm would be
willing to hire L, workers (point A); at a wage of W,, employment would be L,
(point C). Now assume that Congress attempts to stimulate employment by pro-
viding a wage subsidy to employers for every additional worker hired.
In designing a wage subsidy program, policymakers are faced with the same
conflict in goals as the income transfer programs analyzed in Chapter 2. In par-
ticular, they must weigh the trade-off between program costs and the size of the
financial incentive offered to firms to increase employment. One approach to
solving this dilemma involves the following type of wage subsidy program. First,
the program is “categorical” in that the wage subsidy applies only to workers of .
some identifiable low-skill group, such as teenagers or welfare recipients. This
provision serves to target the program at the hard-core unemployed. Second,
the firm receives a wage subsidy only for people newly hired, not for existing
employees. This provision limits the program costs by excluding payments to
workers the firm has hired anyway. This type of program is known as a “mar-
ginal” or “incremental” wage subsidy. A third feature is that only those new
workers who earn a wage less than a target wage rate of W, qualify for a subsidy.
“This graphic analysis is adapted from Robert I. Lerman, “A Comparison of Employer and
Worker Wage Subsidies,” in Robert Haveman and John L. Palmer, eds., Jobs for Disadvantaged Work-
ers: The Economics of Employment Subsidies (Washington: The Brookings Institution, 1982): 159-179.
Chapter 4 The Demand for Labor in the Short Run Lat
This restriction excludes all but the lowest wage workers, targeting the benefits
to those who are likely to have the most difficulty obtaining employment due to
a lack of experience or training, physical impairment, or other such factor.
Fourth, for every worker hired the employer receives a subsidy payment per hour
S equal to some.percent_r of the difference between the target wage Wand the
presubsidy, hiring wage ofW,. The subsidy payment is given by the formula
S =r(W, — W,). (4.5)

The subsidy payment received by the firm will be greater for higher values of the
subsidy rate r or the greater the difference between W, and W,-
There are both positive and negative aspects of this type of wage subsidy pro-
gram. One advantage is that the program provides the largest subsidy and stimu-
lus to employment for those workers who are the lowest paid or most marginal in
the labor market. A second advantage is that the subsidy payment is reduced as
the worker's wage increases with experience and training, gradually weaning the
worker from the program. One negative feature is that the program encourages
employers to replace full-time workers with part-time workers. Although the ac-
tual hours of work performed might not change, the firm’s employment would
significantly increase, qualifying it for subsidy payments. tecond drawback is
that employers may “churn” their labor force by replacing experienced, higher
wage employees with inexperienced, low wage workers from the target group in
order to maximize the amount of subsidy per worker. To the extent that the sub-
ae et
sidy program causes the employer to substitute one type of worker for another,
the net increase in jobs will be diminished. Finally, a third negative outcome of
the wage subsidy program is that it may reduce employment prospects for the
very group of people it is meant to help. Evidence indicates that eligibility for a
subsidy has a stigmatizing effect on disadvantaged workers because employers re-
gard it as a clear|signal a they are “problem cases” and, thus, too big a risk to
employ even at the reduced wage per hour.”
The effect of this wage subsidy program is to create the kinked “after-subsidy”
demand curve D, TD, in graph (a). The before-subsidy demand curve D,D,
shows that at a wage of W, the firm would hire L, workers. After enactment of
the wage subsidy program, however~>the employer's demand for labor at W,
would no longer be L, (point A) but —
rather L/, shown as point B on the kinked
demand curve D, TD. Had the employer’ before-subsidy hiring wage been even
lower, say at W,, a wage subsidy program would lead to a correspondingly larger
increase in employment from L, to L} (point C to E) due to the larger subsidy
payment per worker hired.
Derivation of the After-Subsidy Demand Curve The derivation of the kinked
demand curve D, TD, is illustrated in graph (b) of Figure 4.9. If the employer
pays a wage equal to or greater than the target wage W,, then by Equation 4.5
the subsidy payment S is zero. In this case the employer's demand for labor is
25See Gary Burtless, “Are Targeted Wage Subsidies Harmful? Evidence from a Wage Voucher
Experiment,” Industrial and Labor Relations Review 39 (October 1985): 105-114.
unaffected by the subsidy program and the after-subsidy demand curve remains
D, above point T. If the employer is paying a wage below W,, say W,, the firm
qualifies for a subsidy S$= 1(W,— W,) for each additional worker hired. The
subsidy lowers the employer'snetcostof laborfor each additional worker from W,
to (W, — S) per hour. At W, the “firm’s demand for labor before the subsidy was
L, (point A); what will it now be with the subsidy?
While the actual wage paid by the firm continues to be W,, with the subsidy it
acts as if the before-subsidy wage were only W, — S, leading it to increase em-
ployment from L, (point A) to L, (point B). Plotting the combination of the
wage W, and new level of employment of L, yields point C, which is one point
on the kinked, after-subsidy demand curve. Other points could be obtained in a
similar manner for different before-subsidy wage rates, tracing-out, in the pro-
cess, the demand curve D, TD,. This kinked demand curve is derived for a par-
ticular value of the subsidy rate r. A higher value of r gives rise to an entirely
new after-subsidy demand curve that lies to the right of D, TD).
As hoped for by the program’s designers, the wage subsidy resulted in an in-
crease in employment, represented in graph (b) by the increase in employment
at W, from L, to L,. The benefit of the program is the extra_L, — L, workers
hired by the inline The cost of the program can also be shown in graph (b);
it is the rectangle determined by the amount of the subsidy payment per worker
S (the difference between W, and W, — S) times the L, — L, additional workers
hired by the employer.
The benefits and costs of a wage subsidy program depend critically on the elas-
ticity of demand for labor. The more inelastic is labor demand, the smaller will
be the net increase in-employment for any given amount of subsidy per worker.
For example, if the before-subsidy demand curve D,D, in graph (b) were re-
drawn to be more inelastic, the decline in the employer’s net cost of labor from
W, to W, — S would result in an increase in employment not to L, but to som
lesser amount. This would be reflected, in turn, by a smaller
sm gap best DD;
and the kinked demand curve D, TD). If labor demand is relatively eis a
wage subsidy program will yield less “bang for the buck” in terms of increased
employment, resulting in a high ratio of program cost to jobs created. (If the
before-subsidy demand curve D,D, were more inelastic in graph (b), what
would have to happen to the size of the subsidy payment S to generate the same
L, — L, increase in jobs?)
Estimating the In deciding whether or not to implement a wage subsidy program, considerations
Benefits and Costs of jobs created versus dollar cost to the government are quite important to policy-
of a Wage Subsidy makers. To obtain information on the magnitudes involved, economists build
Program simulation models of the intended program to predict the actual benefits and
costs. A simulation model is a series of mathematical equations that attempt to
represent the responses of firms, workers, and economic variables such as wages
and prices to some specific event, such as a wage subsidy program. After specify-
ing the values of the program “parameters” (the subsidy rate r, the target wage
W,, and so on), the equations can be solved using a computer to predict the
outcomes of the program.
CC ‘
In one such study, John Bishop and Robert Lerman constructed a simulation
model to predict the benefits and costs of a wage subsidy program enacted in
1977 under the “New Jobs Tax Credit” (NJTC).’¢ One of the key parameters in
their model was the elasticity of demand for labor. Based on empirical estimates
from previous studies, Bishop and Lerman set the short-run elasticity of demand
(E>) in their model equal to —.15. Given this and other parameter values, their
model predicted that the NJTC subsidy program might generate as many as
3.7 million new jobs at a relatively low cost of $2,300 per job.
Simulation results such as these represent economists’ “best guesses” as to the
likely labor market effects of particular social programs. Their accuracy, how-
ever, depends critically on two factors: whether or not the structural equations
of the model correctly describe the behavior of workers, firms, and markets, and
whether or not the parameter values for such factors as the elasticity of demand
are correct. How well did Bishop and Lerman’s predictions match the actual re-
sults of the NJTC program? There is some dispute about this. In a later study,
Bishop concluded from a detailed statistical analysis that perhaps as many as one-
third of all the new jobs generated in the construction and retailing industries in
1977 and 1978 were due to the NJTC program.*’ A much more pessimistic as-
sessment was reached in a study by Robert Tannenwald.** Based on telephone
and personal interviews with employers, he found that the NJTC stimulated
relatively few jobs—for a 10-percent reduction in wages due to the wage subsidy,
firms increased their hiring of new workers by only 0.4 percent. This implies that
the elasticity of demand was quite low: Ey = —.04. When asked why they added
so few jobs in response to the wage subsidy, employers cited several reasons. The
most important was that they were reluctant to hire additional workers without a
prior increase in sales. Other deterrents were the complexity of the program
regulations and its temporary 2-year lifetime. Because the elasticity of demand
was so low, the cost per job created by the wage subsidy program was also much
higher than Bishop and Lerman anticipated—estimated at $19,000 to $23,000
per job.
THE RELATIONSHIP Up to this point the chapter has focused on the relationship between the wage
BETWEEN rate and the firm’s short-run demand for labor, a relationship represented dia-
PRODUCT DEMAND grammatically as a movement along a| given labor demand-curve. However, a
AND LABOR number of factors besides the wage rate influence a firm’s desired level of employ-
DEMAND ment. One of the most important is the strength of the demand for its product.
As emphasized in Chapter 1, the demand for labor by a firm is a derived demand.
26John Bishop and Robert Lerman, “Wage Subsidies for Income Maintenance and Job Creation,”
in Robert Taggart, ed., Job Creation: What Works? (Salt Lake City: Olympus, 1977): 39-70.
John Bishop, “Employment in Construction and Distribution Industries: The Impact of the
New Jobs Tax Credit,” in Sherwin Rosen, ed., Studies in Labor Markets (Chicago: University of Chi-
cago Press, 1981): 209-246.
28Robert Tannenwald, “Are Wage and Training Subsidies Cost Effective? Some Evidence from
the New Jobs Tax Credit,” New England Economic Review (September/October 1982): 25-34.
The business firm has as its primary goal the maximization of profit. Its demand_
for labor rises only to the extent that labor is necessary to produce the level of—
output desired by the firm’s customers. The fact that labor isderived from the
demand for the product implies that changes in the level of sales and production
of the firm will necessarily cause a concomitant change in the firm’s desired level
of employment, represented graphically by a shift in the labor demand curve.
This section examines more closely three aspects of this relationship: the cyclical
fluctuation in employment over the business cycle, the pattern of interindustry
employment growth due to long-run shifts in consumer expenditure patterns, f
}
and the impact of imports on domestic employment. A
The Demand for An enduring feature of a capitalist economy is the business cycle. The business
Labor over the cycle is the wave-like pattern that occurs in general business activity as spending
Business Cycle and production increase during the expansion phase of the cycle and then con-
tract during periods of recession. From the end of World War II through 1987,
there were eightrecessions in the U.S. economy. These are illustrated in Figure
4.10 by the bars denoting the respective peak (P) of each cyclical upswing and
the trough (T) of the ensuing recession. Given the derived nature of labor de-
mand, it would be expected that employment in the economy would also rise
FIGURE 4.10 The Growth in Nonagricultural Employment over the Business Cycle, 1948-1987
source: Department of Commerce, Business Conditions Digest (Washington: G.P.0., March 1988).
aa rrr
TABLE 4.2
Percent Change in Payroll Employment
Percent Decline in
Percent Total Durable Total
Industry Employment Change in __—Total Goods Manufac- Service Govern-
during Recessions Recessions Real GNP Nonfarm Sector turing Sector ment
Nov. 1948-Oct. 1949 —1.4 —5.0 —10.7 =(87 S09 1.7
July 1953-May 1954 =a9) —3.0 =) —10.5 0.0 2.0
Aug. 1957-Apr. 1958 =) 3 —4.0 —8.0 —11.8 14 1.6
Apr. 1960-Feb. 1961 —0.6 =p) =—o0 =| —0.1 07
Dec. 1969-Nov. 1970 —0.6 =e? —6.8 =12 (led 2.8
Nov. 1973—Mar. 1975 =a —18 —10.9 =A 2.5 5.2
Jan. 1980—July 1980 = 28 =e —5.8 —6.8 —0.6 0.8
July 1981—Nov. 1982 3.0 =a Salil = 728) 0.0 ={lk5
source: Norman Bowers, “Have Employment Patterns in Recessions Changed?” Monthly Labor Review 104, no. 2
(February 1981): 15-28. Data for the 1981 to 1982 recession calculated by author.
and fall with the business cycle. This is clearly confirmed by the line showing
nonagricultural employment in Figure 4.10. From 1947 to 1987, nonagricultural
employment increased from 48 million in 1947 to 102 million in 1987. This
secular increase in employment was temporarily interrupted eight times, how-
ever, by the sharp contraction in employment that occurred during each re-
cession. In the 1981 to 1982 recession, for example, employment fell from
91.3 million in August 1981 to 88.5 million in December 1982.
The cyclical expansion and contraction of total spending in the economy
gives rise, therefore, to a similar cyclical fluctuation in employment. Two addi-
tional facets of this relationship are important to consider.
Industry Sensitivity to the Business Cycle The first issue is the much greater
impact that recession has on employment in certain cyclically vulnerable indus-
tries in the economy, particularly in durable manufacturing. This is illustrated in
Table 4.2. Table 4.2 shows the percentage decline in real gross national product
and nonfarm payroll employment, respectively, in the goods-producing and
service-producing sectors of the economy in each of the eight postwar reces-
sions. Employment in the goods-producing sector of the economy is subject to
much greater cyclical volatility than is employment in the service sector. Over
all eight recessions, the average percentage decline in employment in the goods-
producing sector was 8.3 percent, while employment in the service-producing
sector actually increased by 0.3 percent. Within each sector, the extremes in the
cyclical responsiveness of employment are illustrated by the durable manufactur-
ing industries and government, respectively. Government employment con-
tinued to expand in all but the 1981 to 1982 recession, while employment in
durable manufacturing declined an average of 11.1 percent in each recession.
ihe markedly different ent cyclical behavior of employment in the goods-
producing and service-producing sectors reflects the difference in expenditure
patterns for the products of each sector. Expenditures for durable goods (for ex-
ample, new homes and factories, automobiles, and capital equipment) tend to
be quite volatile because often the purchase can be postponed, involves a very
large lump-sum payment, or is financed by borrowing. Expenditures on service-
type goods (such as purchases from department stores, grocery stores, and insur-
FIGURE 4.11 The Length of the Average Workweek over the Business Cycle, 1948-1987
Average >> Oe ae o> >} c>>2

(e} 53)
workweek
(in hours)
ze © 33. 22r2e a2 2 $332
Pai Pei eT PT PIP.T
41
40
39
38
194850 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 1987
source: Department of Commerce, Business Conditions Digest (Washington: G.P.0., March 1988).
ance companies) tend, on the other hand, both to be more divisible and to
involve smaller outlays on the part of the purchaser.
The relative stability of employment in the service-producing sector helps to
explain another important trend that is evident in columns | and 2 of Table 4.2.
In comparing the percentage decline in real gross national product in each reces-
sion with the resulting decline in employment, it is evident that (with the possible
exception of the 1981 to 1982 recession) sensitivity of employment to cyclical
swingsin the economy has gradually diminished. One reason for this is ae
cline in the relative size of the goods-producing
producing industries in the economy—
from 41.7 percent of all nonfarm jobs in 1947 to only 24.4 percent of nonfarm
jobs in 1987.
Layoffs versus Cuts in Hours So far the demand for labor has been discussed
solely in terms of persons employed as if labor input were available only in dis-
crete “lumps” of 40 hours per week per person. In actuality, firms are able to
adjust either the number of persons employed or the hours per week each person
works. During an unanticipated recession, which dimension of labor demand are
firms more likely to change—hours or jobs?
A partial answer is provided in Figure 4.11. It shows that the average length
of the workweek for production workers in manufacturing over the 1947 to 1987
period. Hours of work per week scheduled by firms exhibit a clear cyclical pat-
tern as does employment, rising during cyclical upswings and falling durin re-
cessions. A second and Lose ane? point concerns the relative Re of
changes in hours and employment. A comparison of the cyclical turning points
of employment in Figure 4.10 and the average workweek in Figure 4.11 reveals
that changes in hours of work precede changes in employment at both peaks and
troughs of the business cycle. One study found that, on average, the length of
the workweek began to decline 5.4 months before the decline in employment
during cyclical downturns.”
Philip L. Rones, “Response to Recession: Reduce Hours or Jobs?” Monthly Labor Review 104,
no. 10 (October 1981): 3-11.
Chapter 4 The Demand for Labor in the Short Run L77
When faced with the prospect of declining sales and production, why do firms
choose to cut hours before jobs? An important reason has to do with various
fixed-costs of employment.” Firms make substantial investments in workers in
the form of recruitment and hiring costs and in the form of substantial expen-
ditures on training. According to one estimate, for example, a firm’s cost of
hiring or firing a white-collar worker is between two weeks’ and two months’ pay,
and between two days’ and ae eds pay for a blue- collar worker.*! Both hiring
ae
proportionately with ae of hours the employee works. Havinging made

this investment in the worker, the firm obviously has a financial incentive to
avoid worker turnover, since these fixed costs would be lost should the worker
leave the firm. For this reason, firms react to a decline in labor demand by first
chopping hours and only later, if the recession persists or deepens significantly,
by instituting layoffs. One implication of this theory is that the greater a firm’s
investment in fixed hiring and training costs in a worker, the lower is the proba-
bility that he or she will be laid off. This helps explain the tendency for firms to
“hoard”
siatesat |their skilled
ASL EES cae workers
soba during cyclical
ee Eeedownswings
aa while permitting
Selayoffs NE
for the less skilled.*
Overtime Hours The data in Figure 4.11 also show that during cyclical up-
swings of the economy the average, workweek begins to increase above 40 hours
per week as employers schedule overtime hours. If all lapomeoscamere variable,
employers would find it profitable to hire a new part-time worker rather than pay
an existing employee time-and-one-half for overtime. The greater the propor-
tion of fixed costs of employment in a firm’s compensation package, however,
the more likely it is that paying an employee even at a rate of time-and-one-half
will be cheaper to the firm than hiring a new worker and paying the initial hiring
and training costs.”
An important consideration in this regard concerns the effect of fringe benefits
on the demand for labor. Fringe benefits include a variety of nonwage forms of
compensation such as paid vacations, life and health insurance -and government-
required contributions to Social Security, workers’ compensation, and other
such programs. Fringe benefits have become an increasingly large share of a firm’s
total cost of labor, averaging percent
38 of payroll in 1985.
Fringe benefits frequently represent a fixed cost in the sense that once a
worker is hired the expense of health insurance or a paid vacation, for example,
does not vary with the number of hours worked. The larger the ratio of fringe
See Walter Oi, “Labor as a Quasi-Fixed Factor,” Journal of Political Economy 70, no. 6 (De-
cember 1962): 538-555; and Robert A. Hart, The Economics of Non-Wage Labour Costs (Boston:
Allen and Unwin, 1984).
31See S. J. Nickell, “Dynamic Models of Labour Demand,” in Ashenfelter and Layard, eds.,
Handbook of Labor Economics: 517.
32See, for example, Farrell E. Bloch, “Labor Turnover in U.S. Manufacturing Industries,” Journal
of Human Resources 14, no. 2 (Spring 1979): 236-246.
3 See Ronald G. Ehrenberg, “The Impact of the Overtime Premium on Employment and Hours
in U.S. Industry,” Western Economic Journal 19 (June 1971): 199-207.
IN « THE « NEWS
The High Cost to Employers of an Older Work Force
The cost to a company of an hour of labor depends According to the article, the average age of
significantly on the age of the worker. The reason hourly employees at Ford Motor Co., for example,
is that fringe benefits and other nonwage labor jumped from 37 in 1978 to 44 in 1987. Compared
costs tend to rise as an employee ages. Normally, to the new Japanese auto plants in the United
this is not a significant problem for firms. One rea- States with their young, healthy workers, Ford es-
son is that older workers tend to be more produc- timates its costs for fringe benefits are fully $6 an
tive due to greater experience and a lower rate of hour higher.
absenteeism, a second is that in a period of ex- An important cause of this higher fringe benefit
panding sales firms are continually adding new, cost is health insurance. Men aged 50 to 64 submit
younger workers to the payroll. Since 1980, how- an average of $1,897 a year in medical expenses,
ever, the situation has changed dramatically for more than triple the amount for men aged 20 to
many of America’s manufacturing firms because 35. Because Ford’s work force has become increas-
so many of their younger workers have lost their ingly older, the company estimates its health care
jobs due to layoffs. As described in a recent Wall costs for hourly retirees and their spouses will jump
Street Journal article, the result has been a dra- from $160 million in 1984 to $300 million in
matic increase in the average age of the factory 1990.
work force, and with it a substantial increase in A second way an older work force leads to
labor cost for firms and yet further pressure to higher fringe benefit costs is through additional
reduce employment. pension payments. At General Motors, for ex-
benefits to total compensation, therefore, the greater is the inducement for the ©
firm to use Overtime as a means of meeting an increased demand for labor rather
than hiring a new worker.
Fringe benefit costs may also have an impact on the firm’s layoff decision dur-
ing a recession. If a worker’s marginal revenue product is likely to fall perma-
nently below the variable cost_of labor, (the wage) because of a slowdown in
sales, the firm would find it ee oe im or her go even if it means losing
the sunk costs of hiring and training. Given this, which worker would the firm
be most_apt to layoff? In a number of instances the answer is the older worker, ‘
since the firm thereby avoids or minimizes the future expense of pension bene-
fits.** Prevention of this practice is a major factor in the insistence of labor
unions for strong seniority systems that require layoffs to begin with the least
senior worker. This practice has also been circumscribed by the provisions of the
Employee Retirement Income Security Act of 1974 (ERISA), which mandate
that a worker's right to pension benefits must be guaranteed or “vested” after a
certain length of service with the firm, regardless of whether he or she later quits
or is laid off.
4One ex: ample of this is described in “G.M.'s Salaried Employees, Fearful of Layoffs, Flirt with
Unionization,’ ’ The Wall Street Journal (May 17, 1982): 29.
rr
OO————
a
eea
ample, the ratio of active to retired hourly workers ably. In the state of Michigan, for example, GM
dropped from 4.8 to 1 in 1970 to 2.3 to 1 in 1985. paid $954 per employee in unemployment insur-
Because of this trend, the company suddenly finds ance contributions in 1985, while a new company
that its pension liabilities are looming as a larger without GM's history of layoffs paid only $243 per
and larger drain on its available cash. That con- worker.
trasts with Nissan Motor Co.'s plant in Smyra, The end result for companies such as Ford and
Tennessee where the average age of the 3,300 em- GM is something of a vicious circle—because
ployees is in the mid-30s and only one person is their labor costs were higher than overseas auto-
eligible for retirement. makers they lost market share and were forced to
A third cost from an older work force has to do lay off thousands of their workers, which further
with paid time off, such as vacations. raised their average cost of labor because of the
In most
companies, a worker's annual days of vacation in- higher fringe benefit costs associated with older
crease with years of service. As the factory work | work forces.
force ages, therefore, companies can expect an in-
crease in costs for vacations.
The article notes a final way in which com-
panies such as GM and Ford have been penalized
by the aging of the factory work force. Because
these companies have had substantial numbers of
Source: “An Older Work Force Burdens Big Producers in the
layoffs, their required payments to state unemploy- Basic Industries,” The Wall Street Journal (March 5, 1987):
ment insurance funds have also gone up consider- il, Pil.
Consumer Table 4.3 shows data on the percentage change in employment in 16 consumer
Expenditure goods industries over the period 1960 to 1986. The growth rates range from a
Patterns and the 58 percent decline in employment in the leather goods industry to a 258 per-
Demand for Labor cent increase in employment in eating and drinking establishments (restaurants
and bars). What factor or factors account for this disparate behavior of labor
demand among these industries?
One important explanation has to-do with the pattern of camsumer expen-
ditures, Between 1960 and 1986 real per capita disposable income in the United
States grew by 79 percent. As consumers’ income grows, a relatively small frac-
tion of the additional income is devoted to savings, and the remainder of the
increase is spent on additional purchases of goods and services. Consumer ex-
penditures do not increase equally for all goods and services however; the exact
proportion depends on the product's income elasticity of demand.” If the income
elasticity of demand is negative, the good in question is an inferior good, mean-
ing that the quantity demanded of it declines as consumers’ income increases. A
product that has a positive income elasticity is a normal good. Goods that are
staples_gr necessities, such as potatoes, bread, household supplies, and leather
The income elasticity of demand is measured as the percentage change in quantity demanded of
the good (Q) divided by the percentage change in income (Y), i-e., %AQ/%AY.
TABLE 4.3 Percent Percent

Employment Growth Change in Change in
Employment, Employment,
in 16 Consumer
Industry 1960-1986 Industry 1960-1986
Goods Industries,
Leather goods —58% Sporting goods 43%
1960-1986 Dairy products —48 Photographic equipment 70
Tobacco —35 Educational services 87
Bakery products 29 Food stores 116
Tires 5 Hotels and motels 160
Household furniture 9 Airline transportation 186
Soft drinks 21 Hospitals 195
Newspapers 42 Eating and drinking establishments 258
source: Bureau of Labor Statistics, Employment and Earnings (March 1961 and March 1987), Table B—2.
products, generally have low or even negative income elasticities, as do certain

services such as shoe repair. Higher priced durable goods such as cameras and
jewelry, services such as education and health care, and retail products such as
restaurant meals, on the other hand, tend to have high income elasticities.
As income grows over time, the product demand curve in those industries or
occupations with a high income elasticity of demand should shift to the right.
quite rapidly, causing a similar rightward shift in the labor demand curve. This
prediction is borne out in the data shown in Table 4.3. Employment in indus-
tries such as air transportation, hospital services, education, and photographic
equipment experienced quite sizable increases in employment between 1960 and
1986. Conversely, industries with low income elasticities should experience slow
or even declining growth in product demand and, thus, in labor demand. This is
also evident in Table 4.3 by the decline in employment that took place in the
leather goods, dairy products, and tire industries. *°
Policy Application
Imports and the Demand for Labor
One of the most hotly debated subjects facing economic policymakers is the
issue of imports and, in particular, the threat of foreign competition to Ameri-
can jobs. Economic theory demonstrates that both sides gain from international
trade because of the specialization and division of labor that trade allows. As‘
Table 4.4 illustrates, however, the gains from trade are not evenly shared by all
segments of the economy. The data in Table 4.4 show the 8 industries that
gained the largest percentage increases in employment from exports over the
11-year period from 1970 to 1980 and the 8 industries that lost the largest per-
centage of jobs to imports. In general, the United States gained employment from
foreign trade in industries that were technologically advanced (such as aircraft
and computers) or that were capital intensive (such as engines and turbines).
Part of the decline in employment in these industries is also due to productivity growth. This
topic is discussed in the next chapter.
——$—
TABLE 4.4
Percent of
Eight Industries Most Percent of Jobs
Adversely and Jobs Lost Gained
Industry to Trade? _— Industry from Trade?
Favorably Affected by
Foreign Trade, Footwear —15.9% Construction and mining machinery 19.9%
Motor vehicles = (Hilal Engines and turbines 17.8
1970-1980 Electrical components —718 Office and computing machines 16.1
Apparel —6.3 Aircraft and parts 12.8
Leather products —6.3 Miscellaneous machinery 8.0
Radio and TV sets = iy) Electrical equipment Jal
Miscellaneous manufacturing —5.0 Service industrial machinery 5y/
lron and steel =D) Plastics and synthetics 5.4
a a
“Sum of direct employment change in industry plus indirect employment change in supplying industries relative to
1970 employment level.
source: Robert Z. Lawrence, “Is Trade Deindustrializing America? A Medium-Term Perspective,” Brookings Papers
on Economic Activity, | (Washington: The Brookings Institution, 1983): 159.
Job losses from imports were centered in the textile, apparel, and footwear in-
dustries, as well as in steel and certain assembly oriented manufacturing indus-
tries such as automobiles and radios and televisions. These industries were vul-
nerable to import competition because of higher labor costs and, in a number of
cases, because imported goods were of a better quality than similar goods made
in the United States. If data were available for the 1980 to 1986 period the job
losses due to imports would no doubt be much larger than reported in Table 4.4
because of the substantial appreciation of the dollar over that period.
Despite the lesson of economic theory, the sizable inroads that imports have
made in the sales and employment of these industries has led to growing de-
mands on the part of employer and labor groups for increased protection in the
form of higher tariffs or import quotas. Policymakers are thus torn between a
philosophical commitment to free trade and political pressure to protect Ameri-
can jobs. The issues involved in this debate, and particularly who the winners
and losers are from protectionist legislation, can be seen with the economic
theory just developed.
Import Quotas and A case study by Robert Crandall on the economic effects of the voluntary import
Auto Employment quotas on Japanese cars negotiated by. the Reagan administration illustrates the
benefits and costs of using protectionist legislation to insulate domestic jobs from
import competition.*’ As described earlier in this chapter, the bottom fell out of
the market for domestically produced autos in 1980 as consumer demand shifted
from the “gas guzzlers” produced by Detroit to fuel-efficient imports, particularly
those made in Japan. Between 1979 and 1980 sales of U.S.-made cars fell from
8.0 million to 6.3 million units, while imports increased from 2.4 million to
2.5 million units. In 1980, the domestic auto companies suffered a combined
financial loss of $4 billion. One response of the auto producers was to demand
7Robert W. Crandall, “Import Quotas and the Automobile Industry: The Costs of Protection,”
The Brookings Review 2, no. 4 (Summer 1984): 8-16. He provides an updated analysis in “The
Effects of U.S. Trade Protection for Autos and Steel,” Brookings Papers on Economic Activity, 1
(Washington: The Brookings Institution, 1987): 271-288.
FIGURE 4.12 The Effect of Import Quotas on Auto Employment

The import quota on Japanese cars shifts the demand curve for American autos from D, to D, in graph (a), leading toa rise in
price to P, and in level of sales to Q). The increase in production in the domestic auto industry results in a rightward shift of the
industry's labor demand curve from D(Q,) to D(Q,) in graph (b). At the prevailing wage of W,, auto employment increases from
L, to L>.
(a) Product Market (b) Labor Market

Auto
industry
wage
(W)
Ww,
Qy > Quantity of L, Lo Auto .

domestic-built employment
autos (Q)
wage concessions from their unions. A second response was to appeal to the fed-
eral government for temporary import quotas on Japanese automobiles. The ra-
tionale for import quotas was that they would allow the domestic auto industry a
“breathing spell” in which to redesign their product line, modernize their plants,
and get their costs under control in order to compete effectively against the Japa-
nese. Perhaps the major selling point used by the industry to justify import
quotas was that they would save American jobs. Since Chrysler Corporation was
on the verge of bankruptcy with a threatened loss of over 100,000 jobs, this was
a potent argument in favor of quotas.
In early 1981, President Reagan announced that an agreement had been
reached with Japan on a voluntary export restraint that would limit Japanese -
automobile exports to the United States to 1.68 million cars per year. year. The
voluntary quotas were renewed in 1983 and 1985, but with slightly higher limits
of 1.85 million and 2.3 million cars per year, respectively. After the United
States let the quotas expire in 1985, the Japanese government decided to main-
tain them at the 1985 level. What was the impact of these quotas on auto em-
ployment? Did they save American jobs? Who were the winners and losers from
import quotas? The answers to these questions are illustrated in Figure 4.12,
graphs (a) and (b).
Shown in graph (a) is the market for American-made automobiles. Without
import quotas, the demand curve for domestically produced cars is assumed to be
D,, and the industry's supply curve is $,. The equilibrium price for American-
made cars is P,, and the level of production is Q, (point A). Graph (b) shows
the industry's demand curve for labor D(Q,). Given the prevailing wage of W,
and the level of production of Q,, the level of employment in the auto industry
is L, (point X). The imposition of import quotas on ee: cars would be pre-
dicted to have several effects on the HOHESIG auto te
rightward shift ofthe demand curve from D, toD>. (The oe

increase hinges critically on the degree to which Japanese and American cars are
5) 3 the new eae om B), output ofAmeri-
can cars increases fro ©
69 |¢ i 0 Q,, the strv’s dema - labor
crease, Tassos in = (b) by the sneeeemimieianialitimmmeminiian ss

from D(Q,) to D(Q;). At the prevailing union wage of W,, employment in the
auto industry would increase from L, to L, kpaine s tou);
This simple model predicts that im ) to hi ut,
<M mena ee to his estimates, in 1983 ne quotas were responsible

. the pais ofhear etetr eer —— Fe a rise in prices
Who were the winners and losers from the import quotas? The winners are
obvious, with one Sreepuion aOne group made better off . the import quotas
were the domes uto c nies 2 ;
that the quotas increased-compa
The quotas also gave the auto industry pdicnel time to design and build
smaller cars that could compete more effectively with imports. A second group
made better off by the quotas were auto workers and their union. Not only did
the quotas create 21,000 to 46,000 additional jobs, but by restraining competi-
tion from lower cost Japanese producers they also
int 83, total hourly compensation in U.S.
motor vehicle production was $19.02 an hour, compared to $7.91 in Japan.)
Unexpectedly, a third group apparently made better off by thequotas were Japa-
auto
companies!
‘nese demand fortheir cars,the shortage
Given the inelastic
caused by the quotas allowed the Japanese car makers to raise prices by over
$US SE PE InGoce ERODESOUS Me Penal
The losers from the import quotas are also easily identified. The first are Ameri-
but auto buyers faced less choice in the cars available for purchase and substan-
tially longer waiting periods for cars. Although it cannot be shown in Figure
dlve a second group of [oer from the GUSTER &
was the citizens
of the United
tates ref whether the} hase or not. cars
If Japan can produce
ataTees priceethan the Onied sa ie ie a comparative advantage shows
that the average real income of Americans can be increased by purchasing im-
ported cars and specializing our production in the computers, aircraft, and agri-
cultural goods in which we have a relative advantage. Through specialization it
is possible that Americans can enjoy more of both cars and other goods.
Perhaps the most important lesson from this discussion is to illustrate the
interaction of normative and positive economics. Positive economics is able to
clearly show that in the long run American citizens benefit from free trade and
the higher standard of living it makes possible. :
tain 2Tro ry a O worke O
Jiven both the benefits and co icymakers to

decide on the ociety will, }
i te likely to be
his decision involves a consideration of equity,
fairness, and political power, or what is called normative economics. The most
that economists can do in resolving this question is to quantify the benefits and
costs of protectionist legislation so that an informed, carefully balanced decision
can be made. In his study, for example, Crandall estimated that every job saved
by the auto import quotas cost the American public $160,000. Is this price
worth the benefits received? This is a decision that the voters and their elected
representatives have to make.
SCREENING AND The decision rule W = MRP, given in Equation 4.2 provides a clearcut answer
THE DEMAND for the firm to the question of how many workers to hire.** It does not answer,
FOR LABOR however, the question of whom to hire. If all workers were completely identical,
the firm would presumably be indifferent as to which worker was taken on, since
the productivity of each would be the same. Far from being identical, however,
workers are differentiated one from another on the basis of numerous and some-
times hard-to-measure characteristics such as age, race, sex, education, motiva-
tion, skill, and so on. Given the decision to hire an additional employee, the
firm is faced with the difficult task of screening the available group of job appli-
cants to find the person who will be the most productive worker.
What is the process by which firms perform this screening function, and what
are its implications for employment and hiring in the labor market? To begin,
assume that for each job opening in the firm there exists a queue of N persons in
the applicant pool. The object of the employer is to choose from the applicant —
pool the person with the highest marginal revenue product; the problem is that
each person's productivity is uncertain prior to hiring. The hiring decision,
therefore, resembles a contest where the firm attempts to pick the prize (worker)
with the highest expected return.
While the employer cannot observe each person's productivity prior to hiring,
there is a plethora of personal data in the form of observable characteristics and
attributes of the individual that collectively define the “image” of each appli-
Some instructors may wish to postpone this section until Chapter 7.

cant. Of these observable characteristics, some are immutably fixed, such as age,
sex, and race. Others, such as years of education, military service, scores on job
placement tests, and to some degree personal appearance, are capable of being
changed. The immutable characteristics are known as indices, and the alterable
characteristics are known as signals.
The importance of indices and signals for the screening process rests on their
usefulness as predictors of each applicant’s actual productivity on the job. Some
time after hiring an individual, the employer will know his or her productivity
and, on the basis of repeated sampling of this type, will gradually form a subjec-
tive estimate of the indices and signals that are most highly correlated with suc-
cess on the job. When selecting a new employee from the applicant pool, the
employer will use these conditional expectations based on indices and signals to
rank each individual from highest expected success to lowest.
The use of indices such as race or sex for the screening process can lead to
what is known as statistical discrimination in the labor market.” In screening job
applicants, employers may use not only individual characteristics but also group
characteristics such as race or sex that they have come to believe are correlated
with worker productivity. The problem with this practice is that it results in
people who are equal in all other respects in terms of productive ability being
treated differently because of affiliation with a particular group. This leads to
hiring patterns that appear blatantly discriminatory and that have, therefore,
been heavily circumscribed by equal opportunity and affirmative action laws.
This is a subject examined in much more detail in Chapter 9.
In this section attention is focused on the use of signals as a screening device
in the hiring decision. Two issues are given particular attention. The first is why
some worker characteristic, such as years of education or veteran's status, might
be a useful signal in the screening process. The second is how an employer deter-
mines the optimal level of a signal to use as a hiring standard in screening job
applicants. A simple model drawn from the path-breaking work of Michael
Spence elucidates both points.*
Job Market To begin, assume there are two distinct groups of people of equal size in the firm’s
Signaling applicant pool, low-ability group A people and high-ability group B people. Be-
cause individuals in group A have a lower innate ability, their marginal revenue
product for the firm will be, say, $5 per hour, while the marginal revenue prod-
uct of higher-ability group B people is $10 per hour. If the firm possessed no data
on the personal characteristics of each applicant, it would randomly select a
worker from the applicant pool with a probability of .5 of getting a person from
group A and group B, respectively. The expected value or average level of each
worker's MRP, ,,therefore, would be .5($5) + .5($10) = $7.50 per hour. Given
that competition forces a firm to pay workers a wage commensurate with their
See Dennis J. Aigner and Glen G. Cain, “Statistical Theories of Discrimination in Labor
Markets,” Industrial and Labor Relations Review 30, no. 2 (January 1977): 175-187.
Michael Spence, “Job Market Signalling,” Quarterly Journal of Economics 87, no. 3 (August
1973): 355-374.
marginal revenue product, the wage paid by the firm would also be $7.50 per
hour.
The reason firms engage in screening is obvious from this example. At the
going market wage of $7.50, any one firm could readily increase its profits if it
could find some characteristic that identified a high-productivity group B person
from a low-productivity group A person. It would reject all group A workers and
hire only group B workers, paying a wage of only $7.50 (since this is what other
firms are paying), yet reaping a MRP, of $10. All firms, of course, have the same
incentive to engage in screening. To the extent they are successful, wages in the
market will move toward $5 for group A workers and $10 for group B workers,
eroding the short-run profits that come from superior screening. As long as any
uncertainty remains about the actual productivity of job applicants, however,
the firm has an incentive to improve its employee selection process in order to
obtain a work force that is more productive relative to its competitors.”
Given that it pays to engage in screening, how does a firm successfully do it?
Each firm will attempt to identify characteristics or signals of workers that are
correlated with employee performance on the job. There is a dynamic inter-
action in the screening process, however, since workers, realizing that a firm is
using some characteristic such as years of education or veteran's status as a signal,
will attempt to acquire more of it in order to improve their chances of being
selected. Without some limit or constraint on the ability of both low- and high-
productivity workers to acquire the signal, the firm will not be able to success-
fully screen workers of one group from another.
To illustrate under what conditions a worker characteristic will be able to
serve as a successful screening device, the following discussion uses years of edu-
cation as a case study. While firms use many other signals in the screening pro-
cess, years of education is undoubtedly one of the most important, as reflected by
hiring standards at most firms that specify minimum years of educational attain-
ment as a prerequisite for consideration for employment.
Education as a Screening Device The conditions under which years of edu-
cation serve as a successful screening device are illustrated in Figure 4.13. The
horizontal axis measures the years of education (E) obtained by the job appli-
cant, beginning at the minimum level E, that each person is required by law to
complete (eight years in the United States). As before, low-ability group A
workers and high-ability group B workers are assumed to be in the applicant -
pool. The goal of each business firm is to use years of education as a screening
device to separate group A workers from group B workers. Assume that based on
past experience, firms believe that all workers with less than E* years of educa-
tion are low-ability group A workers with a MRP, of $5 per hour, while workers
with E* or more years of education are high-ability group B workers with a MRP,
*'Effective employee screening can result in large cost savings for employers. According to one
study, the use of cognitive ability tests by the Philadelphia police department resulted in labor sav-
ings of $18 million. One reason why screening pays is that performance varies considerably among
employees. Based on 39 studies, it is estimated that the top 5 percent of a firm’s employees are twice
as productive as the bottom 5 percent. See John E. Hunter and Frank L. Schmidt, “Ability Tests:
Economic Benefits versus the Issue of Fairness,” Industrial Relations 21, no. 3 (Fall 1982): 293-308.
FIGURE 4.13 Benefits and Costs to Individuals WhenUTA Is Used as a Screening Device
The hiring standard E* Ss ee
successfully screens the
applicant pool into low-
productivity group A
workers and high-
productivity group B
workers. Group A workers
choose Fy years of
education since their net
wage is $5, compared to
only $3 if they choose E*.
Group B workers choose
E* years of education
since their net wage is $7,
compared to only $5 if & is
chosen.
= Et Years of education (E)
of $10 per hour. Since competition will force firms to pay a wage equal to the
MRP, of workers, the hiring standard of E* gives rise to the discontinuous wage
schedule W(E). Workers with less than E* years of education are presumed to
be group A workers and are paid only $5; workers with E* years or more of edu-
cation are assumed to be group B workers and are paid $10.
Will E* effectively screen group A from group B workers? The answer depends
on the costs of acquiring the signal for persons in each group. If there were zero
costs to obtaining E* years of education, every worker would do so in order to
earn a wage of $10. In this case, years of education would fail as a screening
device since the firm would be unable to successfully differentiate between group
A and group B workers. For any hiring standard to successfully sort group A from
group B workers, the essential condition is that the cost of acquiring the signal
be negatively correlated with the worker’s productivity. This condition implies
that it be more costly for the less-able workers in group A to obtain E* than for
group B workers. Why might this be so? The costs of acquiring additional edu-
cation involve direct money costs for tuition and books, indirect money costs
in the form of forgone earnings from work, and “psychic” costs of stress and
anxiety. An argument can be made that all three components of the costs of
education will be greater for low-ability group A persons because of the extra
time, expense, and difficulty they will face in mastering the subject matter. This
may be particularly true at the college and post-graduate levels. Assuming that
costs of education and ability are negatively ccrrelated, this gives rise to the cost
schedules C, and Cy, in Figure 4.13 (drawn as straight lines for convenience).
Each line shows the dollar cost (pro-rated per hour of work) of acquiring various
yeats of education for persons in groups A and B, respectively. It is assumed that
the cost of obtaining the minimum Ep, level of education is zero for persons of
both groups; as years of education increase, so do the costs. The line Cz lies
below C, on the assumption that the more able persons in group B can obtain a
particular level of education at a lower cost than the less able workers in group A.
Given the wage schedule W(E) and the cost schedules C, and Cz, it can
easily be shown that the hiring standard of E* will, in fact, successfully separate
the firm’s applicant pool into group A and group B workers. To see this, deter-
mine the optimal level of education for the persons in each group. The optimal
level of education is the one where the net wage (the wage W minus the cost of
education C) is the greatest. For the low ability workers in group A, the greatest
net wage is at the education level Ej, where the wage is $5 and the cost of edu-
cation is zero. Although group A workers could obtain a wage of $10 by acquir-
ing E* years of education, it would not pay to do so because of the high costs—
the wage is $10 (point Z), but the costs of obtaining E* are $7 (point Y on Cy),
yielding a net wage of $3. The optimal level of education for high-ability group B
people can be determined in a similar fashion. If they were to choose Ep, their
net wage would also be $5. Could they do better by acquiring E* years of educa-
tion? The answer is yes: their wage would be $10 (point Z), and the cost would
be only $3 (point X on Cx), yielding a net wage of $7. While it pays group B
workers to obtain E*, it would not pay to obtain more than E*, since the wage
remains constant but the costs rise. (For similar reasons it does not pay for group A
workers to obtain more than Ey.) The conclusion is that the hiring standard E*
is a successful screening device because it correctly identifies or sorts the people
in the firm’s applicant pool into low-productivity group A workers and high-
productivity group B workers.
Signaling Failures and the Optimal Signal The hiring standard E* is what
Spence called a “signaling equilibrium,” since it successfully identifies workers as
belonging to group A or B. If it did not do so, employers would eventually dis-
cover, for example, that some workers thought to be high-productivity group B
workers were actually low-productivity group A workers, and this information
would cause the hiring standard E* to be raised. It is a relatively easy task to
identify the hiring standards that would not be a signaling equilibrium. This is
shown in graph (a) of Figure 4.14.
Consider first if the hiring standard were lowered from E* to E,. The wage
schedule W(E) would rise from $5 to $10 at E,, shown by the broken line at
point V. Instead of maximizing their net wage at Ej, group A workers would find
it profitable to invest in E, years of education since the distance between wages
and costs (VU) now exceeds that at E)(XO). Group B workers would also select
E, since the net wage VT is the largest that is obtainable.
In this situation, firms would be motivated to raise hiring standards in an at-
tempt to separate group A and group B workers. What if the hiring standard
were raised all the way to E,? The wage schedule W(E) would now become ver-
tical along the broken line at point Z. Given this, it is easily seen that E, also
fails as a signaling equilibrium since neither group A nor group B workers have an
incentive to invest in this much education. At E,, group A workers would suffer
costs greater than the wage and would thus find Ey more profitable; likewise
group B workers would also choose E, since the distance XO exceeds ZY.
The final issue is whether there is an optimal hiring standard in this model.
“Optimal” in this context means the level of the signal that is able to success-
FIGURE 4.14 Signaling Failures and the Optimal Signal

In graph (a),: a relatively low
) hiring standard of F,; failtails to successfully screen between low- and high-producti
i ivi workers, since
vity
ae groups A and B obtain a higher net wage at £, than Ep. Similarly, a relatively high standard of E, also
fails as a screening
evice, since the net wage for both groups A and B at Ej is higher than at E>. In graph (b), the optimal hiring standard
is the
education
\ level just to the right of E*, ' since it successfull y screens between low- and high-productiviivity workers
in terms of resources devoted to education. ae : UN aga
(a) Signaling Failures (b) The Optimal Signal

W W
$10 $10
$5
Bee E
fully differentiate between group A and group B workers at the least cost in terms
of resources used in the screening process. There is, in fact, an optimal hiring
standard and it is an educational level just to the right of E* in graph (b) of Figure
4.14. A hiring standard of E* or less would result in both group A and group B
workers investing in the signal, and its screening value would be lost. If the
hiring standard were increased beyond E* to, say, E,, individuals would still be
accurately sorted into group A and group B workers, but the cost would be more
than necessary since group B workers would be required to invest in E, years of
education when slightly more than E* years would fulfill the same screening
function.
Chapter 7 discusses the empirical evidence on the role of education as a
screening device. One interesting aspect might be noted here, however. This
concerns the impact of “grade inflation” on hiring standards. Beginning in the
late 1960s, the grade point average of college students began to rise rather mark-
edly as a number of professors relaxed their grading scales (a grade C was no
longer considered ‘“‘average”). The effect was to make it easier for low-ability
students to earn college degrees, an event that would be represented in this
model by shifting down the cost line C, in graph (a) to, say, Cg. While E* years
of education (for instance, a bachelor’s degree) was an effective screening device
prior to the advent of grade inflation, with the shift of C4 to Cz it no longer
was.” The result was that firms raised their hiring standards for a variety of entry
The headline of a recent newspaper article “Educators: B.A. Degree Losing Value” (Atlanta
Journal, February 11, 1985: 1) illustrates this point. The article, summarizing a recent report by the
Association of American Colleges, states “College curriculum has been so watered down that almost
anything goes and the bachelor’s degree has lost its intrinsic value.”
level positions from a bachelor’s to a master’s degree (such as an MBA), repre-

sented in graph (a) as the education level E,. At E,, education is again an effec-
tive screening device.
SUMMARY This chapter examines the firm’s demand for labor in the short run. The most
important concept developed here is the demand curve for labor. The demand
curve has a negative slope, illustrating that there is an inverse relationship be-
tween the wage rate and the number of employees a firm will desire to hire.
Based on the marginal productivity theory, the demand curve for labor is shown
to be the firm’s marginal revenue product schedule. For many economic and pub-
lic policy issues, the elasticity of the demand curve is also of vital importance.
The more elastic is demand, the greater will be the decrease in employment in
response to a rise in wage rates. This was illustrated by the large decline in em-
ployment that occurred in the auto industry in the early 1980s.
The demand for labor is affected by more than just wage rates, of course. A
second key influence on a firm’s desired level of employment is its level of sales.
On a year-to-year basis, employment exhibits a distinct cyclical movement asso-
ciated with the business cycle. Over the longer run, employment among particu-
lar industries or sectors of the economy is also heavily influenced by shifting pat-
terns of consumer expenditures. A final consideration is the impact that foreign
trade has on domestic employment. Although most attention is focused on the
displacement of jobs by imports, the United States also gains thousands of jobs
because of exports.
The final topic focused on in this chapter is the screening process. Firms have
to decide not only how many workers to have but also whom to hire. In improv-
ing their selection of employees, firms use hiring standards such as minimum
educational requirements or prior work experience in order to screen the more
able from the less able in the applicant pool. If a characteristic such as years of
education is to be a successful screening device, the costs of acquiring it must be
negatively correlated with worker productivity.
GLOSSARY
Elasticity of labor demand (E,;)—A measure Law of diminishing returns — An economic law -
of the sensitivity of employment to changes in that states holding one factor of production
the wage rate. It is defined as the percentage constant, as additional units of another factor
change in employment divided by the percent- are added, output will increase but eventually at
age change in the wage rate. a diminishing rate.
Fixed-costs of employment — Costs of employ- Marginal product of labor (MP, )— The change
ment that do not vary with hours worked. Ex- in total output resulting from a one unit change
amples are hiring, training, and certain fringe in labor input, holding capital constant.
benefit costs. Marginal revenue product (MRP, )— The addi-
Indices — Characteristics of workers that cannot tional dollars of revenue earned by the firm from
be changed, such as race or sex. the output produced by the last worker hired.
Calculated by multiplying the marginal product Short-run demand curve for labor— A line that
of labor times the marginal revenue. depicts the relationship between the wage rate
Production function— A mathematical equation and the firm’s desired level of employment,
that states the amount of output that can be holding capital and all other factors constant.
produced given the amount of factor inputs and Signals — Characteristics of workers that can be
the state of technology. changed, such as years of education or veteran's
Screening — The use by firms of personal charac- status.
teristics such as years of education or veteran’s Wage bill—The firm’s total money outlay for
status to rank job applicants based on expected labor, calculated by multiplying the average
productivity. wage rate times the number of employees.
Ss
nnmmeeemmmeememeneemt ee eu IfSQ@NI_H_T2OEOEEEOEE
REVIEW QUESTIONS
1. Use the marginal productivity theory of labor by a majority of the membership. Given this, if a
demand to predict the impact on the firm’s em- layoff resulted in 30 percent of the membership
ployment level of the following events. Explain losing their jobs, would the union vote for a wage
why the change in employment occurs and show it concession? How large would the layoff have to be
in a graph. before the union would accept a concession? What
a. A decrease in the wage rate influence does the elasticity of demand have on the
b. A\n increase in the demand for the firm’s size of the wage concession that the union would
product have to agree to in order to preserve the jobs of its
c. A lower tariff on imported goods members? If demand were highly inelastic, how
d. The conversion of the firm from a per- would this affect the union’s attitude towards a
fectly competitive firm to a monopolis- concession?
tically competitive firm 4. Employment in the government sector grew
2. State whether items a—d are true, false, or by 100 percent between 1960 and 1986. Can you
uncertain, and briefly explain why. explain this rapid growth based on what you think
a. Diminishing marginal productivity of la- is the likely elasticity of the labor demand curve
bor begins when the total product curve for government workers, the income elasticity of
reaches a peak and then declines. demand for government services, and the cyclical
b. If the labor demand curve is inelastic, sensitivity of government employment?
lowering the wage rate will result in a 5. The only manufacturer of large motorcycles
decrease in the firm’s wage bill. in the United States is the Harley-Davidson Com-
c. Since skilled workers are paid a higher pany. The company’s market share dropped pre-
wage than the less skilled, the firm has an cipitously in the late 1970s and early 1980s due to
incentive to lay off the skilled workers greater competition from Japanese producers. After
first during a recession. several years of losses, the company announced it
d. Whether or not a college student had would soon close down, with the loss of several
been elected to office in the student gov- hundred jobs. In response, Congress passed a bill
ernment is a useful screening device for that significantly raised the tariff on large motor-
employers. cycles imported from other countries. What effect
3. Employers urge labor unions to accept wage would this bill have on employment at Harley-
concessions as a means to increase jobs. For a Davidson? Show this in a graph. Who were the
union td do so, however, usually requires approval winners and losers from this legislation?
6. Indeciding whom to admit to graduate school, (GRE). Are both criteria likely to be good signals
many universities in recent years have given less of a person's performance in graduate school? Why
weight to a student’s grade point average and more might universities now be giving less weight to a
weight to his or her performance in entrance ex- student’s grade point average?
aminations such as the Graduate Record Exam
ADDITIONAL READINGS
Arrow, Kenneth. “Higher Education as a Filter.” Rosen, Sherwin. “Short-Run Employment Varia-
Journal of Public Economics 2, no. 2 (July 1973): tion in Class I Railroads in the U:S., 1937-
193-216. One of the original contributions on 1964.” Econometrica 36, no. 4 (October 1968):
the subject of screening. Discusses whether edu- 511-529. Examines the relative variability of
cation results in higher earnings because of its employment among various occupational groups
effect on worker productivity or whether it is in the railroad industry. Finds that occupational
simply a filter to separate the more able from groups with high fixed costs have the lowest
the less able. variation in employment.
Berndt, Ernest R. “Modelling the Simultaneous Thurow, Lester. Generating Inequality. New York:
Demand for Factors of Production.” In Econom- Basic Books, 1975: 211—230. A nontechnical
ics of the Labor Market. London: Her Majesty’s discussion and critique of the key assumptions
Stationery Office, 1979: 125-142. A survey underlying the marginal productivity theory.
of the theory of factor demand with emphasis Topel, Robert H. “Inventories, Layoffs, and the
on labor. Reviews empirical work on labor de- Short-Run Demand for Labor.” American Eco-
mand, particularly with respect to labor/capital nomic Review 72, no. 4 (September 1983): 769—
productivity. 787. Examines the short-run variation in hours
Clark, Kim B., and Richard B. Freeman. “How and employment in various industries. Finds
Elastic Is the Demand for Labor?” Review of that hours are changed before employment and
Economics and Statistics 62, no. 4 (November that employment fluctuation is least severe in
1980): 509-520. Argues that earlier studies in- industries where inventories can serve as a
correctly estimated the elasticity of labor de- buffer between production and sales.
mand. The empirical results indicate that labor Zudak, L. S. “Labor Demand and Multi-Product
demand is more elastic than originally thought. Cost in Semi-Continuous and Multi-Process
Office of Technological Assessment, United Facilities.” Journal of Industrial Economics 19,
States Congress. “The Effect of International no. 3 (July 1971): 267-290. Examines labor
Trade on U.S. Skills and Employment.” In Tech- demand in a steel-making plant where pro-
nology and Structural Unemployment: Reemploy- duction involves a number of interdependent '
ing Displaced Adults. Washington: G.P.O., 1986: products and processes. Finds that a continuous
371-408. marginal product curve cannot be derived.
CHAPTER
The Demand for Labor

in the Long Run
ie chapter examines the demand for labor in the long run. The distinctive
feature of the long run is a sufficiently long period of time for the firm to vary the
amounts of both labor and capital in response to changes in factor prices, product
demand, and ttechnology. This aability to substitute between capital and labor
raises a number of important issues with respect to the demand for labor. One
that is focused on extensively concerns the impact that capital/labor substitu-
tion has on the elasticity of labor demand. Not surprisingly, it can be shown that
an increase in the cost of labor will lead to a much larger decline in employment
in the long run than in the short run, when capital is fixed. A second important
issue concerns the impact that technological change and productivity growth
have on the level of employment in the economy. Over time technological
change and productivity growth continuously reduce the amount of labor needed
to produce a unit of output. Does this mean that automation will lead to growing
unemployment as robots and computers replace humans? Fortunately, the an-
swer is no, for reasons that are explained here. The last topic considered in this
chapter is how the theory of labor demand can be used to forecast the future
level of employment in specific occupations and industries.
~
THE PATTERN OF A business firm can use many different combinations of capital and labor in pro-
CAPITAL/LABOR ducing its product. Which combination will be chosen? This is the central issue
SUBSTITUTION that the theory presented in this chapter attempts to answer. Before developing
this theory, it is important first to gain a better perspective of the facts to be
explained.
Figure 5.1 shows data on the capital/labor ratio for 10 industries in 1960 and
1983. The capital/labor ratio measures the dollar amount of capital (plant and
equipment) used in production relative to each employee. hour _of labor. The
capital/labor ratio varies tremendously among industries. Petroleum refining and
communications, for example, are highly capital intensive, while apparel and
construction are quite labor intensive. One important reason for this difference
is the technology of production. The current state of technology in petroleum
193
194 Chapter 5 The Demand for Labor in the Long Run
FIGURE 5.1
The Capital/Labor
Ratio in Ten
Industries,
1960-1983
source: Bureau of the Census, Statistical Abstract of the United States, 1985 (Washington: G.P.0., 1986), Table 900.
refining requires gasoline to be produced with large amounts of capital in the

form of oil crackers and distillers. The production of a woman's dress, on the
other hand, requires large amounts of hand labor for cutting, stitching, and
fitting.
While technology places important constraints on the proportions of capital
and labor used in production, it would be a mistake to conclude that a firm has
no choice in this matter. In general, quite the opposite is true. This is most
starkly revealed by comparing how particular products are made in the United
States relative to other countries, particularly Third World countries where the
cost of laboris much lower. Whether in a capital intensive industry such as pe-
troleum refining or a labor intensive industry such as apparel, producers in less ‘
developed countries adapt to the high costs of capital and low price of labor by
conserving on the use of machinery and utilizing additional labor. Thus, on a
Chinese construction site, bricks and dirt are carried by hand or pushcart, while
in the United States bulldozers and forklift trucks do the same tasks. Likewise,
cars produced in the United States are welded by robots, while cars produced in
Mexico are welded by humans.
Since technology generally affords a firm a “menu of choice” as to the propor-
tion of capital and labor it uses, the exact proportion chosen will be significantly
influenced by the relative prices of the two inputs. Over time, wages in the
United States have risen faster than the costs of capital, causing firms to sub-
Chapter 5 The Demand for Labor in the Long Run 195
stitute capital for labor.’ This process is revealed in Figure 5.1 by the rise in the
capital/labor ratio between 1960 and 1983. The shift was most pronounced in
mining, where the amount of capital per employee hour more than doubled.
This discussion points out two central issues that the theory of labor demand
must address: first, what determines at a point in time the optimal mix of capital
and labor in production, and second, how does this optimal mix of capital and
labor change over time as wages and the cost of capital change? These questions
are answered in the next several sections of this chapter.
THE THEORY OF The last chapter derived the firm’s short run demand curve for labor under the
LABOR DEMAND assumption that the amount of capital was fixed. In this section the theory is
IN THE LONG RUN extended to cover the long run when both capital and labor are variable.
The Technology The goal of the firm is to produce the level of output that maximizes profits. To
of Production: accomplish this, the firm must decide on the least-cost production method and,
lsoquants in particular, on the appropriate combination of capital and labor. This decision
is influenced by two factors: the first is the constraints that technology places
on the mix of capital and labor, and the second is the relative prices of the fac-
tor inputs. ;
The technology of production is embodied in the firm’s production function.
As discussed in the last chapter, each firm has a production function of the form
Q=f(K, L), (5.1)

which expresses the relationship between the level of capital and labor inputs
and the maximum obtainable level of output, given the current state of technol-
ogy. It is possible to use the production function to illustrate graphically the al-
ternative production processes that current technology makes available to the
firm. This is shown in Figure 5.2.
The production function states, for example, that if K, amount of capital and
L, amount of labor are used by the firm, the resulting level of output will be Q,
(point A). Similarly, if capital and labor are increased to K, and L,, respec-
tively, output would increase to Q, (point B). A further increase to K,; and L,
would yield Q; (point C). The production function thus shows the relationship
between increases in capital and labor and increases in output, represented by
the movement from point A to B to C. It can also be used to answer a different
type of question, however. Returning to point A, if K, units of capital and L,
units of labor can produce Q, units of output, could any other combinations of
capital and labor also produce Q,? In general the answer is yes, as illustrated by
the curve Q, that passes through point A. The curve Q, is called an isoquant; it
' This process also occurs in low-wage, Third World countries. In the summer of 1987 a wave of
strikes hit employers in South Korea, forcing them to substantially raise wages. One result, according
to Business Week [“What Kind of Korea Will the New Labor Movement Build?” (October 12, 1987):
56], is that “Hyundai Motor Co. postponed a $375 million plan to build a new assembly line. Hyun-
dai, in view of the increased labor costs, wants to design a more automated plant.”
FIGURE 5.2 A Set of Isoquants

The curves Q,, Q), and Q;
are isoquants. Each curve Units
capita
shows the alternative
combinations of capital
(kK)
and labor that can produce
a given level of output.
The slope of an isoquant
measures the marginal rate
of technical substitution
(MRTS)—the rate at which
labor can be substituted
for capital while keeping
output constant.
Unitsoflabo(L
r)
shows all the various combinations of capital and labor that are just able to
produce a given level of output.
An important feature of isoquants is that they are convex to the origin. The
reasoning is much the same as for indifference curves. In moving from point B to
point D along isoquant Q,, labor must be increased by L; — L, and capital de-
creased by K, — K,. This process of substitution could be repeated, but the ques-
tion is: to reduce capital usage by the same amount as before but still keep output
constant at Q,, how muchmore
m labor would be needed? The answer is that
more than an additional L; — L, amount (see point
int E) will be needed since
each worker now has less capital to work with. In general, then, isoquants are
convex, giving rise to a diminishing marginal rate of technical substitution
(MRTS). The marginal rate oftechnical substitution measures the additional
amount of capital that is required if labor is changed by one unit and output is
keptconstant. The MRTS.is given by the slope of the isoquant. Moving down
an isoquant, the MRTS becomes progressively smaller.” _
For most goods, the technology of production allows at least some room for
choice in the exact proportion of capital and labor that is used in production. In
some cases, however, the technology is such that capital and labor must be used
in strict proportion to each other, allowing one and only one capital/labor ratio
in production. Examples include airplanes that require a fixed crew size, or
manufacturing equipment (such as a drill press) that requires a fixed number
‘Observe that the convexity of the isoquants also gives rise to diminishing marginal productivity
of labor in the short run. This can be shown in Figure 5.2 by drawing a horizontal line beginning at
K, capital and zero labor; to reach each successive isoquant takes ever larger increases in labor.
Chapter 5 The Demand for Labor in the Long Run 197 i)
FIGURE 5.3 A Production Technology with Zero Substitution Possibilities

The L-shaped isoquants
Q;, Q2, and Q3 represent a "Capital (Ky)
production technology
with zero substitution
possibilities between
capital and labor. In this
example, a unit of output
requires labor and capital
in the proportion of 2:1.
Thus, to produce Q; = 100,
production requires L = 20
and K = 10. Any less
of either L or K means
that production would 20 Q;=200
necessarily fall to a lower
level, no matter how much 10 Q,; =100
of the other input is

provided.
20 40 60 Labor (L)
of operators. This type of technology gives rise to right-angled, L-shaped iso-

quants.’ This is illustrated in Figure 5.3. ia
It is assumed that each unit of output requires .2 units of labor and .1 units of
capital. Thus, to produce Q, = 100 units of output requires K = 10 and L = 20
(point A); to produce Q, = 200 units requires the same proportion of capital
and labor, but twice as much, K = 20 and L = 40 (point B). There are several
important points about these isoquants. First, for a given level of output there
are zero substitution possibilities between capital and labor (MRTS
= 0). Since
capital aand labor are required in fixed proportions, it is impossible to produce a
given level of output with more of one but less of the other. Second, if one factor
is held constant, adding more of the other will not increase output at all, as can
be seen by starting at point A, and keeping K at 10, but increasing L to 40. The
result is no increase in output (point D) since the additional labor has ng capital
to work with. Finally, also drawn in Figure 5.3 is a ray from the origin through
the kink points of the isoquants. The slope of this ray defines the¢. I/labor
ratio necessary for production. In this example, the slope is .5. \Ol20
Factor Prices: The isoquants show all the possible combinations of labor and capital that a firm
Isocost Lines could choose to produce a particular level of output. Before the firm can choose
3A concept known as the elasticity of substitution (ox,) measures the relative curvature of the
isoquants and, thus, the degree of substitution possibilities between capital and labor. It is defined
(where R is the cost of capital) as
Ce
sal 6
%A(-F)
The greater the curvature of the isoquant, the greater will be the change in the ratio of capital to
labor (the numerator) for a given percentage change in relative factor prices (the denominator). If
the isoquant is L-shaped (fixed proportions), oK,, = 0.
FIGURE 5.4 A Set of Isocost Lines

Each isocost line shows all
the combinations of capital Units of
and labor that can be capital (K)
purchased for a fixed dollar
expenditure, given the Kg
prices of the two inputs.
The slope of an isocost line
is equal to the negative of Ke
the ratio of factor prices,
W/R. An increase in the
wage rate will cause an
isocost line to rotate to the
left, such as from the solid
line CD to the dashed
line AD.
Units of labor (L)
the particular combination that would maximize profits, it also has to know what
the relative prices of labor and capital are.
At a particular point in time, a going price or “rental rate” will exist for a unit
of capital of Ry dollars and wage rate for labor of W, dollars per hour that the
firm must pay a its capital and labor inputs.* An isocost line shows, given these
factor prices, all the various combinations of capital and d labor that the firm
could purchase with a particular amount of money expenditure. A series of iso-
cost lines AB, CD, and EF are shown in Figure 5.4. The derivation of the iso-
cost line AB is straightforward. For a given level of expenditure by the firm of
E, dollars, the total amount of labor that could be hired, given the wage W,, is
E,/W, = L, workers, shown as point A. Similarly, if the firm spent all E, dollars
on capital, it could purchase E,/R,; = K, units, shown as point B. Connecting
these two points yields a straight line that shows all the other combinations of
capital and labor that could just be purchased with E, dollars of expenditure.
Several features of isocost lines are important. First, the slope of an isocost
line is constant, and equal to the (negative) ratio of the factor prices_—W,/R,.
Thus, if the wage is $5 and the cost of capital is $10, the slope of the isocost line
s —.5, reflecting the fact that to purchase one more unit of capital, two units of
labor have to be given up. The slope of an isocost line is constant, reflecting the
assumption that the firm can buy as much labor or capital as it desires at the
going market prices of W, and R,.
‘Since the price of labor (the wage) is measured per hour, the price of capital should be similarly
measured. To do so, economists assume the hourly cost of capital (R) to be the amount per hour a
firm would have to pay to rent the plant and equipment.
FIGURE 5.5 The Profit-Maximizing Combination of Capital and Labor

The firm must choose
the level of output that __ Capital (K)
maximizes profits, and then
the combination of capital
and labor that produces
that level of output at least
cost. Assuming that Q, is
the optimal output level,
the least-cost combination
of capital and labor is K,
and L,, where the isoquant
Q, and isocost line CD are
tangent (point X).
Second, for given factor prices, a higher level of expenditure by the firm on
capital and labor will give rise to an entirely new isocost line that is parallel and
to theright of the original one. Thus, if the number of dollars budgeted for labor
and capital were increased from E, to E,, the total amount of labor that could be
purchased would increase to E,/ W, = L, (point C), and total capital purchased
could increase to E,/R, = K, (point D). A still higher expenditure level of E,
would give rise to the isocost line EF. The isocost lines are parallel to each other
since, for a given wage rate and cost of capital, each has the same slope, of
= Wilkie
Finally, if either the priceof capital or the wage rate should change, it will
give rise to an entirely new set of isocost lines. Given the factor prices W,, R,,
and an expenditure level of E,, for example, the isocost line is CD. What would
happen to the isocost line, however, if the wage rate would rise from W4_to_W,?
Given the same cost of capital, R,, the firm could still purchase E,/R, = K,
units of capital (point D). At the higher wage of W,, however, only
a smaller
E,/W,= L, units of labor (point A) could be purchased. Connecting points D
and A by tlau broken line gives a new isocost line that is steeper and to the left of
the original isocost line DC. The slope of the isocost line DA is steeper reflect-
ing the higher ratio of factor prices, —W,/R,. A_rise_in the.wage, therefore,
causes the isocost line to zotate to the left;.a Aone. in the wage would cause it
to rotate to the right. (How would a decline in the cost of capital change the
isocost line?)
The Equilibrium The isoquants Q,, Q,, and Q; from Figure 5.2 are combined in Figure 5.5 with
Level of the isocost lines AB, CD, and EF from Figure 5.4. In order to determine its
Employment optimal level of employment, the firm has two related decisions or calculations
to make. The first decision concerns the optimal level of production for the firm.
To maximize profits, should it produce Q,, Q>, or Q;? Given the factor prices of
W, and R,, microeconomic theory shows that the profit maximizing level of out-
put for a competitive firm is where the product price (P) equals the marginal cost
of production (MC). For the sake of exposition, assume that P = MC at the
output level of Q, in Figure 5.5.
Given the choice of Q, as the rate of output, the second decision facing the
firm is how to produce Q,. This decision involves the following question: what
combination of capital and labor would produce Q, at minimum cost? The an-
swer is given by the tangency between the isocost line CD and the isoquant Q,
at point X. At point X the firm minimizes the cost of production, since no other
combination of capital and labor would allow it to reach a lower isocost line.
Point Z, for example, is an alternative way to produce Q), but it lies on the
higher isocost line EF. Points V and Y represent the respective cost-minimizing
input combinations for producing output levels Q, and Q; although, by assump-
tion, neither would be the profit maximizing point of production.°
At the tangency of the isoquant Q, and isocost line CD, the slopes of the two
lines are just equal. Since the slope of the isocost line is given by the ratio of
factor prices W/R (neglecting the minus sign) and the slope of the isoquant is
given by the marginal rate of technical substitution, MRTS, the minimum cost
combination ofcapital and labor is given by the following equilibrium condition:
W
MRTS = ope : (5.2)
Using this equilibrium condition, the firm in Figure 5.5 would minimize cost by
employing L, workers and K, amount of capital.
A Change in the The next issue concerns how the firm’s desired level of employment will change
Wage Rate if the cost of labor should change. This is illustrated in Figure 5.6, graphs (a) and
(b). To make the graphic presentation simpler, only the conditions for cost
minimization (rather than profit maximization) are analyzed.
To begin, assume that the wage rate is W, and cost of capital is R,, giving rise
to the isocost line AB in graph (a). The tangency point between the isoquant
Q, and the isocost line AB (point X) yields the firm’s least cost level of employ-
ment L,. Graph (b) shows the horizontal product demand curve of a perfectly.
cempetitive firm D, and its marginal cost curve MC,. As stated before, the
competitive firm will produce the output where P = MC; in graph (b), this oc-
curs at the output level Q, (point f3)) given the prevailing price in the market
of Pe
Assume now that the wage rate increases from W, to W,, the cost of capital
remaining constant at R,. What will be the impact on the firm’s demand for
labor? In answering this question, the adjustment process is broken into two dis-
*Connecting points X, Y, and Z yields the firm's expansion path. The expansion path shows all the
combinations of capital and labor that minimize the cost of producing alternative levels of output
such as Q), Q2, and Q;.
FIGURE 5.6 The Change iin the Short-Run and 1a Run Demand for Labor Due to a Wage Increase
A rise in the wage causes
the firm's demand for labor se
to fall from L, to Ly (point X
to Y) in the short run due
Capital
to the scale effect as
(kK)
the firm cuts back its
production level from Q, to
Q,. In the long run, the firm
is able to substitute capital
for labor in the production
of Q,, leading to a further
decline in labor demand
from L, to L3 (point Y to Z)
due to the substitution
effect.
Labor (L)
&
ay eee ews e:
RUL
OgerO, Quantity (Q)
crete steps: the initial change in labor demand when capital is fixed in the short
run, and the further change in labor demand when capital becomes a variable
input in the long run.
A rise in the wage rate from W, to W, rotates the isocost line to the left in
graph (a) from AB to CB. Given the higher cost of labor, the firm has to recon-
sider what its optimal level of output is. Assuming the wage increase is isolated
to this particular firm, the impact on the firm’s level of output is shown in
graph (b). A rise in the wage will increase the furm’s costs of production, shifting
its marginal cost curve to the left from MC, to MC). At the going market price
of P,, this rise in production cost causes the firm to lower its level or “scale” of
production from Q, to Q, (point G to H) where P, = MC).
The initial impact of an increase in the wage rate, therefore, is to reduce the
firm’s level ofoutput,in this case from Q, to Q). As the level of production is
reduced, it also leads to a declinein the firm’s demand for labor. This is illus-
trated in graph (a). The original level of labor demand was L, (point X). By
increasing production costs, the higher wage of W, makes this firm less competi-
tive in the product market, forcing it to reduce output to Q). In graph (a),
therefore, the isoquant Q, becomes the new level of production. Given this, the
firm must then decide on the least-cost_combination of capital and labor to
produce Q,. In the short run, this decision is constrained by the fact that the
amount of capital is fixed at K,. Thus, in the short run, the firm’s optimal level
of employment falls fromL, (point X) to L, (point Y) where the input combina-
tion of K,, L, is just able to produce the new output level Q).
Point Y is not a long-run equilibrium point, however, because it does not pro-
duce the output level Q, at minimum cost (it is ygt on the isocost line CB).
Given the higher wage of W,, the cost minimizing combination of capital and
labor to produce Q, is given by the condition MRTS,= W,/R,. This occurs at
Aree point Z where the isoquant Q, and isocost line CB are tangent.In the short
run the firm is constrained to use K, amount of capital and thus cannot reach
point Z. In the long run, however, the higher price of labor will induce the firm
to substitute capital for labor in producing Q,, increasing capital to K, but re-
ducing labor still further to L;.°
The decrease in labor demand from L, to L; in graph (a) can be decomposed
conceptually into two separate parts. The first impact of a rise in the wage is to
cause the firm to reduce its level orscale of production from , and, thus,
its demand for labor. This decline in labor demand due to the decrease in the
level of production is known as the scale effect. Graphically, the scale effect is
measured as the shift from one isoquant to another. In graph (a), the scale effect
is measured by the decline in labor demand from L, to L, (point X to point Y).
An increase in the wage rate also leads to a second adjustment in the firm’s-
demand for labor. For any given level of output such as Q,, the change in rela-
tive factor prices will cause the firm to substitute capital for labor in the long
run. The resulting change in labor demand due to capital/labor substitution is
known as the substitution effect. Graphically, the substitution effect is mea-
sured by the movement along an isoquant, such as from point Y to point Z on Q).
It shows that even if the scale of production did not change, a wage increase
would cause a decrease in labor demand of L, — L, from the decision of the firm
to utilize more capital and less labor in the production process.
The Long-Run Figure 5.6 can be used to derive the firm’s short-run and long-run demand curves
Demand Curve for labor. At the original equilibrium (point X), the wage rate was W, and the
for Labor level of employment in the firm was L,. This combination of wage rate and level
of employment is plotted as point X in Figure 5.7. By increasing the wage to W),
Figure 5.6 shows that in the short run the scale effect causes the firm to reduce
*Point Z represents a position of long-run cost minimization, but it is not the firm’s ultimate point
of profit maximization. The reason is because the movement from point Y to Z reduces the firm’s
marginal costs of production, making the profit maximizing level of output greater than Q) (but still
less than Q,). After all adjustments are complete, therefore, the scale effect will be smaller than
indicated in Figure 5.6.
FIGURE 5.7 The Long-Run and shorts Run ia Curves for Labor
The long-run labor demand
curve D, is more elastic
than the short-run demand
curve Ds. A rise in the
wage from W, to W leads
to a decline in labor
demand in the short run
from L, to L, (point X to Y)
due to the scale effect. In
the long run, however,
labor demand declines
further from L, to L3
(point Y to Z) due to the
substitution effect.
Labor (L)
employment to L, (point Y). This combination of W, and L, is then plotted in

Figure 5.7 as point Y. Connecting points X and Y yields the firm’s short-run de-
mand curve for labor Ds.
~ In the long run, the rise in the wage to W, provides an incentive for the firm
to reduce its employment even further by substituting capital for labor. This was
shown in Figure 5.6 by the substitution effect as labor demand declined from L,
to L;. In Figure 5.7, at W, the short-run demand for labor is L, (point Y), but in
the long run the demand for labor is only L; (point Z). Connecting points X
and Z yields the long-run demand curve for labor D, . It shows that in response
to the rise in the wage from W, to W,, the demand for labor in the long run
declines from L, (point X) to L; (point Z). This decline in employment is made
up of two components: the short-run decrease from L, to L; because of the scale
effect, and the additional decline from L, to L; because of the substitution
effect.
The most important point to notice from Figure 5.7 is that the long-run de-
mand curve is more elastic than the short-run demand curve. In the short run
the firm’s ability to adjust to a wage increase is limited because it is locked into a
fixed amount of capital. In the long run, however, the ability to change not only
the level of production but also the amount of capital provides the firm with
much more, flexibility in adapting to higher or lower wages, resulting in a de-
mand curve for labor in the long run that is more sensitive to changes in wage
costs than in the short run.
THE The elasticity of labor demand is one of the most important concepts in labor
DETERMINANTS OF economics because it is at the center of so many economic and public policy
THE ELASTICITY OF issues. For example: how much will an increase in the minimum wage cause firms
LABOR DEMAND to cut back on employment? To what extent will a wage subsidy program stimu-
late new jobs? If unions push up wages, how many of their members will lose
their jobs as a result? The answer to all of these questions depends on the sen-
sitivity of employment to wage rates.
Economists have identified four specific factors known as the four laws of de-
rived demand that determine the relative size.of the eae
These four laws state that the demand for labor, other things equal, will be more
elastic:
1. The larger the price elasticity of demand for the product being produced.
2. The greater the share of labor cost as a percent of the total cost of production.
3. The greater the ease of substitution in production between labor and other
factor inputs such as capital.
4. The greater the elasticity of supply of other competing factors of production.
The reasoning behind each of these four laws is briefly explained below.
Demand for the Final Product Will the demand for labor in a particular in-
dustry or occupation be elastic or inelastic? In answering this question, the price
elasticity of demand for the product is of crucial importance. The reason is that a
rise in wages immediately results in higher costs of production, causing firms to
raise the price of their product in order to maintain profit margins. As the price
of the product goes up, consumers buy less of it. If the demand for the product is
very elastic, the result of this price increase will be to cause a large decrease in
sales and hence a relatively large decrease in the demand for labor.
This first law of derived demand affects the elasticity of labor demand through
the scale effect since, in this case, the impact of a wage change is felt through a
change in the leveLof production. The more elastic the demand curve for the
product, the greater will be the decrease in employment due to an increase in
the wage rate. In terms of Figure 5.6, this means that the short-run decrease in
employment from point X to point Y will be relatively large as the firm is forced
by the substantial decline in sales to produce on a much lower isoquant. In Fig-
ure 5.7, in turn, the larger the scale effect, the more elastic will be both the
short-run and long-run demand curves for labor.
This law of derived demand has several important implications. The first is
that the demand curve for labor for the individual firm will be more elastic,
other things equal, the more competitive the product market in which it oper-
ates. The reason is that the greater the number of firms selling a particular prod- '
uct, the greater is the ability of consumers to switch their purchases from one
company to another, making each company’s product demand curve more sen-
sitive to the price it charges. This is why, as shown in Chapter 4, the labor de-
mand curve for an imperfectly competitive firm is more inelastic than that of a
perfectly competitive firm.
This result helps explain, for example, why unions in the 1970s were able to
push up wages more rapidly in oligopolistic industries such as autos or steel and
’See John R. Hicks, The Theory of Wages, 2d ed. (New York: St. Martin's Press, 1963), 241-247.
in regulated industries such as trucking and airlines.* In both situations, either

large economies of scale or government restrictions on entry resulted in only a
relatively small numberof firms in each industry or market, reducing consumers’
substitution possibilities and, as a result, making the demand curve for labor in
‘each industry more inelastic. The more inelastic the labor demand curve, in
turn, the smaller will be the employment loss for a union as it raises wages.
A second implication of this law of derived demand is that, as shown in Chap-
wages at one tire manufacturer, for example, would increase its production costs,
forcing it to raise tire prices. Since one brand of tire is a close substitute for an-
other brand, the tire manufacturer with higher prices would suffer a significant
decline in sales and employment as customers switched to its lower cost rivals.
Should wages increase at all tire producers, however, the resulting rise in tire
prices would have a much smaller impact on sales since consumers have few
viable alternatives to substitute toward. Thus, because there are fewer substi-
\ecs Boche tutes for tires in general than for one particular brand of tire, a wage increase
or . throughout
the industry will have a smaller_relative effect on employment than
Wwelosin
Cif the wage increase were isolated to one individual firm in the industry.
Union wage policy offers many illustrations of the importance of this prin-
ciple.’ One well-known example concerns the United Mine Workers union. In
the 1960s, about 85 percent of domestic coal was mined by members of the
UMW. Because major buyers of coal such as utilities and steel companies had
few non-UMW sources of supply, the union faced a relatively inelastic product
demand curve and labor demand curve, allowing it to aggressively raise wages.
By 1984, however, the UMW’s share of industry employment had fallen to
around 45 percent, largely because of the union’s failure to organize the newly
opened strip mines in the western United States. This decline forced the UMW
to significantly moderate its wage demands in the 1984 negotiations lest many
of its members lose their jobs as buyers of coal switched to lower cost non-
union producers.
The Share of Labor in Total Cost ~The second law of derived demand concerns
the share of labor cost as a percentage of the total cost of production. It also
influences labor demand through the scale effect. A rise in wages, as discussed
previously, causes production costs and the price of the product to rise, reducing
sales and the demand for labor. The actual impact of a rise in wages on produc-
tion costs depends on how high labor costs are relative -
duction.In capital-intensive industries such as petroleum refining, labor cost is a
8This is shown in Bruce E. Kaufman and Paula E. Stephan, “The Determinants of Interindustry
Wage Growth in the Seventies,” Industrial Relations 26 (Spring 1987): 186— 194.
* Statistical evidence is provided in Richard B. Freeman and James L. Medoff, “The Impact of the
Percent Organized on Union and Nonunion Wages,” Review of Economics and Statistics 63, no. 4
(November 1981): 553-560. The case of the coal industry is discussed in William Miernyk, “Coal,”
in Gerald G. Somers, ed., Collective Bargaining: Contemporary American Experience (Madison, Wis.:
IRRA, 1980): 1-48.
small fraction of the total cost of refining a barrel of oil. In this case, a 20-percent
increase in wages for refinery workers would have a relatively small impact on
the level of employment since total production costs and the price of refined oil
would not be much affected. In labor-intensive industries such as apparel, how-
ever, the demand for labor is likely to be much more elastic_since the same
20-percent increase in wages would boost total costs considerably more, re-
sulting in much higher prices, lower sales, and a relatively large decrease in
employment.’°
One implication of this second law of derived demand is that the firm’s labor
demand curve for a narrowly defined occupation or craft will be more inelastic
than that for all employees in general. As an example, consider the case of
movie and television actors. Movie and television actors are represented by the
Screen Actors Guild. The Guild is a craft union that limits its membership to
actors.!! While labor costs in general represent a sizable share of total costs for
studios and television networks, the cost of actors themselves relative to total
costs is much smaller. As a result, the impact on the employment of actors from
a wage increase will be smaller
much than if the Guild were an industrial union
that bargained the same wage increase for all studio and network employees. It is
thus to the advantage of the Guild and the actors it represents to restrict its
membership to a narrowly defined occupation so that the disemployment effect
of
raising wages is reduced. (This is sometimes called the “importance of being
unimportant.”) The same conclusion applies to other craft unions such as airline
pilots, painters, carpenters, and plant guards.
Substitutability of Other Factor Inputs Ann increase in the wage rate causes a
reduction in the short-run demand for labor because of the scale effect. In the
long run, the demand for labor will decline even more as capital, is substituted
for
labor. The greater this substitution effect is, in turn, the more_elastic will be the
long-run labor demand curve.
The ease of substitution between capital and labor is determined by both tech-
nological and institutional constraints. The technological ability to substitute
between capital and labor in production is reflected in the relative curvature of
the isoquants; the more sharply curved the isoquant, the smaller is the ability
of the firm to substitute one factor input for another. Thus, &f the technology
of production is such that only one combination of capital and labor is able to
produce a particular level of output, such as that pictured by the L-shaped iso-
quants in Figure 5.3, there will be a scale effect but no substitution effect from a
wage increase, causing the long-run demand curve for labor to be relatively in-
elastic (but not completely inelastic).
An excellent example of how differences in the ease of substitution between
capital and labor can affect the elasticity of labor demand is provided by farm
J. R. Hicks (The Theory of Wages) has shown that this prediction may not always hold. In par-
ticular, this prediction is reversed when there are small substitution possibilities in consumption, but
large substitution possibilities in production.
ll
A craft union
:
organizes
.
workers who have a particular
: P :
skill or occupation (such as plumbers); an
industrial union organizes workers of all skills in a particular industry (such as auto workers).
laborers and airline pilots. A wide range of possible combinations of capital and
labor can be used by a land owner to produce a crop such as lettuce. When wages
are low relative to the price of capital, planting and harvesting will be done by
hand, with only the simplest of tools. Should wages of farm laborers rise, the
farmer is motivated to mechanize farm production, replacing labor with ma-
chinery. The ease of substitution between capital and labor in agriculture thus
makes the long-run demand curve for farm labor relatively elastic. The opposite
case is provided by airline pilots. As wages rise for pilots, airlines would like to
replace labor with capital much as the lettuce growers did. Are they able to do
so? While there is some room for substitution (instead of flying older jets that
require three pilots, the airline can buy newer planes that require only two pi-
lots), the technology of air travel places severe limits on the extent to which
pilots can be replaced by capital. What would be expected, therefore, is that an
equal increase in wages for farm laborers and pilots, other things being equal,
would result in a much larger decline in employment for the farm laborers be-
cause of the easier substitution possibilities in agriculture.”
Substitution possibilities between capital and labor and between different
types of labor (for example, union versus nonunion labor) are also constrained
by institutional rules such as union shop requirements, union manning restric-
tions, and government safety regulations. As one example, railroads switched
from steam to diesel locomotives partly because of the potential savings in labor;
diesel locomotives do not need firemen. The railway craft unions, however, suc-
cessfully bargained for work rules that required the railroads to keep firemen on
the trains, preventing the substitution of capital for labor.
A second example concerns union shop requirements. If wages of unionized
workers increase relative to nonunion workers, firms have an incentive to sub-
stitute from the former to the latter. To limit the disemployment effect on their
members, therefore, unions have frequently negotiated contracts mandating the
firm to operate a “union shop.” A union shop clause mandates that upon 30 days
from being hired a worker must join the union as a condition of continued em-
ployment, diminishing the ability of employers to substitute nonunion labor for
union labor. Substitution possibilities under a union shop agreement are still not
zero, however, since in the long run the firm can build new plants in less union-
ized states or states with “right-to-work” laws (laws that make union shop re-
quirements illegal).
Government safety regulations can also reduce the firm’s ability to substitute
between capital and labor. One example is in the airline industry. Before airlines
could fly the new generation of jets (the Boeing 757 and the McDonnell Douglas
DC-9 Super 80) that required only two pilots, the planes had to obtain safety
certification from the Federal Aviation Administration (FAA). While certifi-
2 An implication of this discussion is that unions should be strongest in those industries where
substitution possibilities in production are smallest. Evidence in support of this hypothesis is pre-
sented in Richard B. Freeman and James L. Medoff, “Substitution between Production Labor and
Other Inputs in Unionized and Nonunionized Manufacturing,” Review of Economics and Statistics 64,
no. 2 (May 1982): 220-233.
cation was ultimately granted, the Airline Pilots Association actively lobbied
against it, delaying certification for several years.”
The Elasticity of Supply of Other Factor Inputs The fourth law of derived de-
mand concerns the elasticity of supply of factors of production besides labor.
This fourth law also affects the elasticity of labor demand through the substitu-_
tion effect, As discussed previously, a rise in the wage rate relative to the price of
capital motivates the business firm to substitute capital for labor. As the demand
for capital increases, this may itself set off a large increase in the price of capital,
negating the original desire on the part of the firm to make the substitution of
—~ J& capital for labor. The extent to which the price of capital rises in response to an
increase in demand depends on the elasticity of its supply curve; the more in-_
elastic the supply curve, the higher will be the rise in the price of capital and the
smaller willbe the firm’s incentiveto substitute away from labor. Thus, the more
inelastic are the supply curves of other competing factors of production, the
smaller will be the substitution effect for a given wage increase and the smaller
the long-run elasticity of labor demand.
The importance of this fourth law is illustrated by the example of airline pi-
lots. In the early 1980s, some senior pilots for major carriers were earning as
much as $150,000 a year. Given the high cost of pilots, airlines had a strong
incentive to substitute capital for labor by purchasing new two-pilot planes such
as the Boeing 757. The supply curve of these new planes, however, was initially
fairly inelastic since increases in production were heavily constrained by long
lead times for parts and components, as well as by the relatively fixed supply of
certain highly skilled workers. If all airlines had attempted to purchase these
new planes, the large increase in demand would have resulted in a substantial
bidding up of the price, eliminating part of the original incentive to buy them.
TECHNOLOGICAL Over time, the demand for labor is influenced not only by changes in relative
CHANGE AND factgr prices but also by changes in technology. Improvements in the state of
LABOR DEMAND technology arise from advances in basic knowledge (for example, genetic engi-
neering) and improved techniques of production (such as robots, word pro-
cessors, and jumbo jets). Particularly during the 1960s, there was widespread
concern over whether the process of technological change or “automation” would
result in rising levels of unemployment in the economy, a concern that has re-
surfaced today as robots replace auto workers and word processors make the typ-
ing pool a thing of the past.'*
The impact of technological change on employment can be analyzed with the
long-run theory of labor demand. Improvements in the state of technology can
most often be incorporated into the production process only through the addi-
tion of new capital in the form of more modern, up-to-date plant and equip-
"This is described in “Do Three Pilots Make a Crowd?” Business Week (July 21, 1980): 179-180.
“See, for example, H. Allan Hunt and Timothy L. Hunt, Human Resource Implications of Ro-
botics (Kalamazoo, Mich.: W. E. Upjohn Institute, 1983).
Chapter 5 The Demand for Labor in the Long Run ow S09
FIGURE 5.8 The Effect of echegsoical Changeon Labor TEE G

Technological change
shifts the isoquant Q, ,
inwards to Q; +41,
illustrating that the same
Q, level of output can now
_ be produced with less
capital and labor (compare
points X and Y). The lower
costs of production this
makes possible, however,
result in a lower price
and an increase in sales to
Q> ++. The net result is
an increase in the demand
for capital and labor from
point X to Z.
Labor (L)
ment. Thus, in the short run the employment of a firm is largely unaffected by
technological change since, by definition, the amount of capital is fixed. In the
long run, however, the firm has the opportunity not only to substitute between
labor and capital but also to replace older, technologically outdated capital with
the technologically most advanced capital.
The impact of technological change on the demand for labor is best illustrated
in terms of its effect on employment within an industry rather than within a
single firm. Assume that each firm in the industry has an identical production
function of the form Q = f(K, L) that expresses the relationship between capital
and labor inputs and the maximum amount of output that can be obtained, given
the current state of technology. This production function can be used to generate a
series of isoquants, one of which is pictured in Figure 5.8 as Q, ,. It represents a
level of industry output of Q, produced in the current time period t. Given the
ratio of factor prices W,, R, in the industry and the resulting isocost line AB,
firms minimize costs by producing Q, ,with K, units of capital and L, units of
labor, shown as point X.
Technological change has a two-pronged effect on the demand for labor. The
initial effect is to reduce the demand for labor as better technology allows firms to
produce a given levelof output with fewer workers. The second effect is that
improved technology results in lower production costs and, thus, lower product
prices, increased sales, and a greater demand for labor.’
‘5 This two-pronged effect is well-illustrated in a recent case study of the impact of computers on
the employment of managers and clerical workers. See Paul Osterman, “The Impact of Computers
on the Employment of Clerks and Managers,” Industrial and Labor Relations Review 39 (January
1986): 175-186.
The Displacement The essence of technological change is that it opens up new, more efficient ways
of Labor by to produce a product with less labor or capital. This effect is represented in Fig-
Technologica! ure 5.8 by the shift toward. the origin of the isoquant Q, , to Q;,,4,- The iso-
Change quants Q, ,and Q, ,,; represent exactly the same level of output; the difference
is that with the advance in the state of technology, Q; can now be produced in
period t + 1 with less labor and capital. If factor prices remained at W,, Rj,
firms in this industry would minimize the costs of producing Q, by using only K,
units of capital and L, units of labor, given by the tangency of the isoquant
Q, ,+, and the isocost line CD (point Y).
The first impact of technological change, then, is to reduce the demand for
labor in the industry from L, to L, (point X to Y) as firms build new labor-saving
7OU CES plants or install more A henterng rise-saving equipment. Although it may seem
UIT CosTS counter-intuitive at first, technological change is also capital saving in many
cases, as shown by the decline in capital inputs from K, to K>. One example is
the railroad industry. When railroads switched from steam to diesel locomotives,
the same number of ton miles of freight service could be produced with less
capital since diesels required less time for repairs and could operate over much
greater distances without refueling. Similarly, computers have resulted in sub-
stantial savings in capital in the trucking industry since more efficient schedul-
ing permits fewer trucks to haul the same amount of freight. While technologi-
cal change frequently reduces both the absolute level of capital and labor needed
to produce a particular product, the_ratio_of-capitalto labor may change if tech-
nological change has a factor “bias,” that is, if it proportionately reduces the use
of one input more than the other.'*
If the only effect of technological change was to continually shift the iso-
quants in Figure 5.8 toward the origin, over time the economy would need fewer
and fewer workers to produce a given level of output. Despite the revolution in -
technology in the 20th century, however, the number of jobs in the economy
has not decreased but rather has increased by many millions. How can this be
accounted for?
Technological It is true that if firms were to produce the level of output of 1900 with the tech-
Change and nology of 1988, only a fraction of the work force would be employed today. How-
Product Demand ever, even as technological change reduces the amount of labor needed per unit.
of output, it also leads to a large expansion of total output demanded and thus of -
labor demanded in the economy.
Technological change leads to an increase both in the salesof the individual
industry and in total product demand in the entire economy. Consider first the
level of sales in the industry. Beginning microeconomics shows that the average.
a es
total cost of production puts a floor under the product price that firms can charge
if they are to earn a “normal” profit and remain in business in the long run. The
effect of technological change, however, is to reduce the unit costs of produc-
tion, since the same level of output can now be produced with less labor and
"See W. E. G. Salter, Productivity and Technical Change (Cambridge, England: Cambridge Uni-
versity Press, 1969).
Chapter 5 The Demand for Labor in the Long Run 73 bl
capital. In a competitive industry, as per unit costs of production fall, competi-

tion among firms leads to downward pressure on the price of the product until
the price is again just equal to average total cost.!” This fall in the price is quite
important for employment since itleads to expanded sales in the industry. The
extent to which output demand increases depends on the elasticity of the indus-
try’s product demand curve; the more elastic it is, the greater will be the increase
in sales in response to the fall in price. The effect on labor demand is illustrated
in Figure 5.8. The initial impact of technological change shifts the isoquant
from Q;,,to Q;,, 41, causing a decrease in employment in the industry of L,; — L).
The fact that Q; can now be produced at a lower cost, however, leads to a lower
price and greater sales in the industry. If the product demand curve is relatively
elastic, the increase in sales may be enough to make the isoquant Q, , , , the new
profit maximizing level of production. Given the same ratio of factor prices, the
tangency of the isocost line EF and isoquant Q, ,,, (point Z) yields a new de-
mand for labor of L; units and demand for capital of K; units. While L, — L,
workers initially lost their jobs due to the more efficient technology, the lower
prices this made possible resulted in the end in a net increase of.employment in
the industryof L,; — L,.
It is also possible, however, that employment may decrease. If the elasticity of
of_
product.demand is relatively low, or alternatively, if, due to a lack of competi-
tion in the industry, the cost savings are not passed on in the form of lower
prices, the decline in the product price may... nerate enough newsalesto
provide jobs for all the L, — Tj workers initially laid off. In this case, the new
profit maximizing isoquant Q), , , ; in Figure 5.8 would lie somewhere in between
Q,,+; and Q,,, and the level of employment would likewise lie somewhere
between L, and L,, resulting in a net decline over time in industry employment.
Where will these workers who are permanently displaced by the technological
change find jobs?
The answer is that technological change, by reducing the price of the indus-
try’s product, stimulates more sales not only in the industry itself, but also in
many other industries in the economy. A fall in the product price of one good
has the effect of increasing the real income of every worker in the economy. For a
given level of dollar income, a decline-in the price of one product means that
the consumer can now buy not only more of it, but also more of other products,
as well, leading to an expansion of sales and in labor demand across numerous
industries. The extent to which employment expands in each industry as a result
of the higher level of real income depends, as shown in the last chapter, on the
income elasticity of the industry's product demand curve.
—= ne
The Combined The net result is that in a particular industry, technological change may or may
Effect not lead to a long-run decline in the demand for labor. This is illustrated in
Figure 5.9, which shows the long-run industry labor demand curve D,jg,). At
the wage of W,, assuming the level of output is Q,, the demand for labor in
'7Perhaps the best example of this is the computer industry, where prices per unit of memory have
fallen 15 percent to 20 percent a year.
FIGURE 5.9 The Impact of Technological Change on the Long-Run Demand for Labor
Given the wage W,,, the first impact of technological change is to displace labor as better technology enables the firm to produce
the same Q, units of output with less labor. This shifts the long-run labor demand curve to the left, from Dyq,) to Dy + 1(a,). The
resulting decline in the product price, however, stimulates an increase in sales to Q, leading to a rightward shift of the labor demand
curve to D, , \a,). In this example, sales do not increase enough to offset the original displacement of labor, and employment
suffers a net decline of L; — L3. The opposite could also occur.
Ww,
Dz..1,) Divan Dia,
ba: ke Le Te Labor (L)
4 period t is L, (point A). The first impact of technological change is to reduce

@ the labor requirements per unit of output, represented by a leftward shift of the
long-run demand curve to D, , \(g,). At the same level of output of Q,, in period
t + 1 the demand for labor is reduced to L, (point B). The second impact of
technological change is to reduce per unit costs of_production and the price of _
the industry’s product, leading to an increase in sales and the demand for labor.
This is represented by the rightward shift ofD,,;,o,) to D, + 1(Q,)) Tesulting in an
increase in employment from L, (point B) to L; (point C). Whether this subse-
quent increase in labor demand of L; — L, is enough to offset the initial dis-
placement of labor of L; — L, depends on the extent to which the savings in
production-cest are passed on to buyers in the form of lower prices and on the
price and income elasticity of the industry's product demand curve. Figure 5.9
assumes a net decrease of L, — L, jobs in the industry. Due to higher real in-
comes of consumers, however, labor demand curves in other industries will also
shift to the right, providing jobs in other sectors of the economy to replace those
lost because of technological change.
The scenario of events traced out here has a real-world counterpart. As shown
in Chapter 4, employment in the goods producing sector of the economy has
grown only slightly over time while employment in the service sector has boomed.
One reason is that the pace of technological change has been greatest in the
goods producing sector, resulting in a significant decline in unit labor require-
ments. The resulting gains in real income have been disproportionately spent by
consumers on services, causing large gains in employment there. The combina-
tion of high rates of technological change in goods production and a high in-
Chapter 5 ~The Demand for Labor in the Long Run ae 213
come elasticity of demand for services has resulted, therefore, in a fundamental

shift in the locus of employment in the economy.
In the long run, technological change is a positive force leading to higher real
incomes aa growing
tou) employment opportunities. '* Society, therefore, gains from
technological progress. Not all of these gains are equally shared, however, and
for some workers there are significant costs.'? Workers who are displaced from
their jobs by new technology may have to seek work in a completely different
industry or occupation. For older workers or workers with families, this transi-
tion process may be a difficult one. Many of the new jobs created in service-
related industries, for example, pay considerably lower wages than the jobs in
manufacturing eliminated by new technology. An additional aspect of tech-
nological change is that it frequently imparts a skill twist to labor demand, de-
creasing labor demand for unskilled or semiskilled workers (such as assembly line
workers, telephone operators, and stenographers), while increasing the demand
for skilled, highly educated workers (for example, robot maintenance workers,
computer programmers, and word processor operators). To effectively compete
for these new jobs, displaced workers must invest in a new vintage of “human
capital” in the form of additional education or vocational training. While younger
workers can frequently undertake the financial and emotional burdens of this
transition, older workers often cannot, resulting in a serious unemployment
problem for this labor force group.”
PRODUCTIVITY The discussion of capital/labor substitution and technological change leads to a

GROWTH AND third subject—the effect of productivity growth on the demand for labor. The
THE DEMAND term “productivity” usually refers to labor productivity, defined as the amount
FOR LABOR of output produced in the firm, industry, or economy per employee hour. There
are other productivity measures, however, such as capital productivity (output per
unit of capital input) and total factor productivity (output per unit of total fac-
tor input).”!
An increase in labor productivity means that firms are able to produce each
physical unit of output with less labor input. The discussion of the long-run de-
mand for labor provides insight into two of the most important causes of the
~ secular increase in labor productivity..The first is the process of capital/labor
substitution that has taken place as wage rates have risen relative to the costs of
capital. The price of capital equipment (producer's durable equipment) rose by
'8See Richard Belous, “Technological Change and Its Effects on Labor Markets,” in Proceedings of
the 1986 Spring Meeting (Madison, Wis.: Industrial Relations and Research Association, 1986):
494-501.
See Philip L. Martin, Labor Displacement and Public Policy (Lexington, Mass.: Lexington Books,
1983); and Jeanne Gordus, Paul Farley, and Louis Ferman, Plant Closings and Economic Dislocation
(Kalamazoo: W. E. Upjohn Co., 1981).
This is graphically illustrated in “Mechanic’s Switch to High Tech Is Full of Personal, Financial
Strain,” The Wall Street Journal (May 31, 1983): 31.
21 These are described in more detail in J. A. Mark and W. H. Waldorf, “Multifactor Productivity:
A New BLS Measure,” Monthly Labor Review 106, no. 12 (December 1983): 3-15.
IN « THE « NEWS
High-Tech Allows American Firms to Stay
Competitive with Overseas Rivals
In many of America’s basic industries, the total of computer technology to every step of the manu-
compensation cost per hour of labor is in the $20 facturing process, beginning with product design
to $25 range. How, then, can American manufac- and ending with shipping of the finished good. For
turing firms hope to compete against their rivals in example, a factory might receive a new order for a
countries such as Taiwan, Korea, or Mexico where custom-made part on Monday. The part would be
wage and benefit costs are so much lower? One an- designed on a video screen and automatically ana-
swer is by reducing the amount of labor used per lyzed for performance and producibility. The de-
unit of output through more advanced technology. sign would be dispatched to an automated milling
A particularly promising approach is computer- system where the part would be made exactly to
integrated manufacturing (CIM). Perhaps the most specifications by a system of robots and lathes, and
interesting aspect of CIM concerns the type of then sent off to shipping and receiving where the
labor that it allows firms to economize on. Sur- central computer would have already printed the
prisingly, the workers in question wear white, not invoice and shipping label. On Tuesday, the part
blue collars, as described in a recent article in might be ready for shipment, rather than a month
Business Week. or two later, as in most factories.
The article states that although computer- Surprisingly, the chief goal of CIM is not to save
integrated factories are still a rarity, they are never- on direct-labor costs. For most products, such sav-
theless beginning to reshape the U.S. industrial ings would fall far short of justifying the huge in-
landscape. The essence of CIM is the application vestment required because factory automation has
296 percent between 1950 and 1985, for example, while wage rates in manufac-
turing rose by 565 percent. Business firms responded by trying to hold down pro-
duction costs as much as possible by substituting from labor to capital in produc-
tion. By doing so the firm produces the same level of output with fewer employee
hours, giving rise to an increase in labor productivity. A second important factor
leadi ng to increases in labor productivity is technological change. As shown in
Figure 5.8, technological change allows the firm to produce the same level of
output with less labor, leading to an increase in output per worker and, thus, in
labor productivity.”
The initial impact of the rising relative price of labor and the continual ad-
vance in the state of technology, therefore, is to increase labor productivity and
reduce the demand for labor. Since labor productivity has grown by about 2 per-
cent a year since 1950, does this mean that the economy offers 2 percent fewer
jobs each year? The answer is no for reasons already discussed. Although the
* According to Mark and Waldorf, “Multifactor Productivity,” 40 percent of the increase in labor
productivity between 1948 and 1981 was due to an increase in capital per employee hour, and
60 percent was due to technological improvements and other sources of increased efficiency.
IE
———————_———————__————— E
Ee
already reduced the direct-labor content of most automation could more than offset higher U.S.
U.S.-made goods to 5 to 15 percent. Instead, wage rates.
CIM’s primary benefits come from automating the The result? The old plant employed more than
flow of information through a factory. This elimi- 200 people, but the new CIM plant uses only a
nates not only direct labor, but also much of indi- handful of people to monitor the line and repair
rect labor—that 45 percent or so of cost attri- equipment. Instead of days and even weeks it took
butable to middle management, clerical staff, and to switch production from one type of motor starter
supervisory personnel. CIM also provides signifi- to another in the old factory, the new CIM plant
cant cost savings through improved production turns out 130 versions of the starters without ever
quality and reduced inventories of materials and shutting down. The company estimates that the
finished goods. CIM plant cut unit production costs 40 percent,
One of the best examples of CIM in action is an and that the plant is now the lowest-cost producer
Allen-Bradley plant in Milwaukee that manufac- of electric motor starters in the world. True, auto-
tures electric motor starters. In order to compete mation cost a lot of jobs in Milwaukee as the com-
with low-wage foreign labor, Allen-Bradley had puter and the robot replaced supervisors and as-
already transferred a large part of its manufactur- semblers, but without CIM there would have been
ing from Milwaukee to Mexico, and was consider- no jobs at all.
ing the same course of action for the electric motor
starters. After thoroughly investigating the bene-
fits and costs of CIM, however, the company de-
cided to invest $15 million in a new state-of- Source: “High Tech Comes to the Rescue,” Business Week
the-art plant in Milwaukee on the bet that total (June 16, 1986): 100-103.
initial impact of productivity growth is to reduce the demand for labor, it also
leads to greater employment because lower, product prices and higher real _in-
comes stimulate additional sales in the economy.” Rather than destroying jobs,
therefore, productivity growth is actually the wellspring of higher per capita in-
comes and more jobs in the economy.
What is true at the economywide evel may not be true for an individual in-
dustry, however. If an industry’s product has a low price elasticity and income
elasticity of demand, rapid productivity growth may lead to a net decline in
employment as the additional sales resulting from lower prices do not offset the
displacement of labor from capital/labor substitution and labor-saving techno-
logical change. The increase in real income resulting from the higher productiv-
ity growth in the one industry, however, will create additional jobs in other in-
dustries as that income is spent by consumers.
3 Capital/labor substitution in response to rising wages does not reduce production costs; rather it
minimizes the rise in per unit cost. The continuing process of technological change in the capital
goods industries, however, reduces the price of capital and gives rise to capital/labor substitution,
y which does result in higher productivity and lower per unit costs of production.
eee ed
TABLE 5.1 Telephone
The Relationship Agriculture Communications
between Productivity Index of Relative Index of Total
Growth and Output per Price of Total Farm § Output per Relative Price Telephone
Employee Agricultural Employment Employee of Telephone Employment
Employment in Goods (in thousands) Hour Service (in thousands)
: Hour
Agriculture and (a) (b) (c) (d) (e) (f)
Telephone — 1954 24.0 100.0 8,651 23.2 100.0 699.0
Communication, 1960 37.0 87.3 7,051 37.4 99.0 706.0
1954-1984 1970 66.0 76.9 4523 62.1 TAI 942.0
1984 139.0 68.0 3,461 144.0 51.8 985.7
sources: Data for agriculture from Economic Report of the President, 1987 (Washington: G.P.O., 1987), Tables B—-94,
B-95, B—96; data for telephone communication from Bureau of Labor Statistics, Handbook ofLabor Statistics, 1975
(Washington: G.P.0., 1976), Table 128; Employment and Earnings (March 1985), Table B—2; and Productivity Mea-
sures for Selected Industries, 1958-85, Bulletin 2277.
Empirical Evidence
Productivity Growth and Employment in Agriculture
and Telephone Communication
The relationship between productivity growth and labor demand is vividly illus-
trated by the different patterns of employment change in two industries, agricul-
ture and telephone communication. Both industries have experienced a virtual
revolution in the technology of production over the last 30 to 40 years. In 1950,
for example, a long distance telephone call was handled with mechanical switch-
ing equipment, long distance operators, and cable transmission systems; three
decades later the same long distance call was made with direct dialing equip-
ment, computers, and satellites.“ A similar transformation has taken place in
agriculture with the introduction of high-yielding hybrid grains, modern tilling
and harvesting machinery, and computerized egg and cattle production.
The process of technological change and capital/labor substitution in both in-
dustries resulted in dramatic increases in labor productivity. These productivity
gains are illustrated by the data in columns (a) and (d) of Table 5.1, which show
the growth in output per worker in each industry. Using 1977 as the base year
(1977 = 100.0 percent), output per worker in 1954 in agriculture was only
24 percent, and in telephone communication, 23 percent. Looked at another
way, between 1954 and 1984 output per worker in agriculture increased by
446 percent and in telephone communication by 520 percent. Such rapid gains
in productivity substantially reduced the amount of labor needed per unit of out-
put. Given the state of technology and level of capital in 1984, for example,
firms in both industries could have produced 1954's output with fewer than one-
fifth as many workers.
See “Telephone Communication,” in Bureau of Labor Statistics, Technology and Labor in Five
Industries, Bulletin 2033 (Washington: G.P.O., 1979): 28-40.
The previous section shows, however, that productivity growth also leads to
forces that increase employment in the industry. The first is that productivity
growth leads to declining unit costs of production, which should lead to lower
product prices, greater sales, and greater employment. This process has clearly
been operative in both agriculture and telephone communication, as revealed by
the data in columns (b) and (e). The figures in each column are a ratio of two
price indices: the numerator is the respective price index for the industry's prod-
uct or service (1954 = 100.0), and the denominator is an index of consumer
prices in the economy (the Consumer Price Index, 1954 = 100.0). Both ratios
have declined over time, meaning that the level of prices in both agriculture and
telephone communication increased less rapidly than all prices in general or, to
put it another way, the relative price of agricultural goods and telephone service
declined over time. By 1984, for example, the relative price of agricultural prod-
ucts had declined by one-third, and the relative price of telephone service had
been cut in half. This decline in the relative product price should have led to
greater sales and employment in both industries, the amount of increase depend-
ing on how price elastic each industry’s product demand curve was.
The second source of employment growth from increases in productivity comes
from increased product demand as a result of the rise in the real income of con-
sumers. Because productivity growth leads to lower product prices (or at least to
a slower rate of increase), the rise in wage rates over time will outstrip the rise in
product prices, leading to an increase in real income for consumers. Between
1954 and 1984, real per capita disposable income grew by 89 percent. This in-
crease in real income led to greater consumer expenditures and product demand
for agricultural goods and telephone service, as well as for many other goods
and services.
The net effect of productivity growth on industry employment depends on
whether the initial displacement of labor because of productivity growth is less
than or greater than the subsequent increase in employment because of greater
industry sales. For agriculture, column (c) of Table 5.1 shows that the initial
displacement effect of productivity growth has far outweighed the effect of in-
creased sales from lower prices and higher incomes. Between 1954 and 1984, for
-example, agricultural employment fell by 60 percent. The reason is that agricul-
tural products have both a very low price elasticity and income elasticity of prod-
uct demand. Even though relative prices of agricultural goods have fallen and
the real incomes of consumers have risen, demand for agricultural goods such as
bread, milk, and meat has increased only modestly. For agriculture, then, pro-
ductivity growth has led to a sharp decline in agricultural employment because
the savings in labor per unit of agricultural goods have outweighed the increase
in labor demand due to greater levels of production.”
In marked contrast to agriculture, employment in the telephone communica-
tion industry actually increased by 41 percent between 1954 and 1984. Tech-
See Patricia A. Daly, “Agricultural Employment: Has the Decline Ended?” Monthly Labor Re-
view 104, no. 11 (November 1981): 11-17.
nological change and capital/labor substitution significantly reduced unit labor

requirements in the telephone industry, but the demand for telephone service
has grown so rapidly that the net effect has been a rise in total employment. In
this situation, then, the price and income elasticities of demand for the product
have been sufficiently high that productivity growth has resulted in a net in-
crease in industry employment.
The telephone industry also offers a prime illustration of the skill twist that
technological change imparts to labor demand. A recent study of a large Cana-
dian telephone company, for example, found that technological change reduced
the demand for labor in inverse relation to the skill level of the occupation—the
least skilled category, telephone operators, suffered a decline in employment
from 39 percent of total company employment in 1952 to 10 percent in 1972,
while more highly skilled white-collar workers increased their share of employ-
ment from 29 percent to 38 percent.”° Similar trends have been projected to
continue in future years.
Policy Application
Employment Forecasting
What industries and occupations will offer the best job prospects for college
graduates in 1995? How many new jobs are created by an increase in defense
spending? What region of the country gains the most jobs from American ex-
ports to other countries? These are some of the questions for which policymakers
in Congress and state and local governments need answers and that economists
are trained to provide. How do economists make such employment projections?
Employment forecasting is an application of the theory of labor.demand. One
frequently used approach utilizes the input-output model.
The input-output model was first developed by the Harvard economist Wassily
Leontief, an accomplishment that earned him the Nobel prize. The heart of the
input-output model is the concept of the production function, symbolically rep-
resented here as Q = f(K, L). The traditional use of the production function is
to predict the level of output that will result from a given level of factor inputs.
Leontief’s key insight was to realize that the production function could also be
used_in reverse to predict, given a certain level of output, the labor and capital
requirements necessary. to produce te
Using the production function to forecast future employment trends is more
complicated than it might sound, however. The reason is illustrated in Fig-
ure 5.10. Assume the current level of employment is L,, given by the tangency
between the isocost line AB and the isoquant Q, (point X). The first step to
employment forecasting is to estimate how much the level of output will rise
**Michael Denny and Melvyn Fuss, “The Effect of Factor Prices and Technological Change on
the Occupational Demand for Labor: Evidence from Canadian Telecommunications,” Journal of Hu-
man Resources 18, no. 2 (Spring 1982): 161-176.
FIGURE 5.10 The Effect of rate Factor Prices on aia al adie

Given the existing cits
employment level of L,
(point X), to forecast the
future level of employment
it is necessary first to
predict the level of output.
Assuming the level of
output to be Q,, if
isoquants are convex
allowing substitution
possibilities between. K
and L, it is also necessary
to predict the change
in relative factor prices
and the curvature of the
isoquants. If factor prices
do not change, future
employment would be L,
(point Y); if wages rise
relative to the cost of
capital, however, the future
level of employment might
be only L3 (point Z).
de i, Lo Labor (L)
between the current period and the forecast date. In some cases one or more
estimates may be supplied to the forecaster; for example, Congress might ask for
projections of the number of new jobs created by three alternative defense bud-
gets of $300 billion, $350 billion, and $400 billion. In other cases the forecaster
has to project the future level of output (such as the level of GNP in 1995) based
on historical trends and current developments. Whatever the method, the ex-
ample considered here assumes that the future evel of output at the forecast date
is predicted to be Qo.
With the convex isoquants pictured in Figure 5.10, even if output is correctly
prediced to be Q, it is still impossible to predict the level of employment with-
out additional information. In particular, it is also necessary to predict first the
change in telative factor prices between the current date and the forecast date,
and second the relative curvature of the isgquants. If relative factor prices re-
main the same, for example, the isocost line CD would be parallel to AB and
the predicted level of employment would be L, (point Y). If wages increase
faster than the cost of capital, the slope of the isocost line would become steeper,
such as the broken line EF. In this case, at the same output level of Q, employ-
ment would only be L; (point Z). The extent to which changes in a
affect the predicted level of employment depends on the curva of the isg-
quants. The greater the substitution possibilities between capital and ae the
greater will be the sensitivity of labor demand to changing factor prices (that is,
the greater the distance L, — L;).
The assumption that isoquants are convex as in Figure 5.10 places a consider-
; able burden on the forecaster because he or she has to predict not only the future
level of output, but also the future level of wages and capital costs and the con-
vexity of the isoquants. These informational demands are, in fact, so large and
the resulting computational work so complex that the traditional production
function can generate employment forecasts only at a very aggregated level,
such as for the manufacturing sector.
By way of contrast, the input-output model is capable of generating employ-
ment forecasts for several hundred detailed industries or occupations. How is it
able to do this?
The Input-Output The key to the input-output model is the assumption that the technology of pro-
Model duction allows zero substitution possibilities between capital and labor, giving
rise to a “fixed proportions” type of production function. Such a production
function was illustrated in Figure 5.3. Each isoquant is L-shaped, implying that
a given level of production requires capital and labor in one and only one
proportion.
Given the assumption of fixed proportions, the crucial step in using the input-
output model is to calculate the input-output coefficients. The input-output
coefficient for labor (a,) is defined as a, = L/Q, the coefficient for capital (ax) is
defined as ay = K/Q where L, K, and Q are measured at the time of the forecast.
The labor input-output coefficient is the amount of labor required per unit of
output. For the technology illustrated in Figure 5.3, and assuming Q, is the ini-
tial level of output, a, = 20/100 = .2, and ay = 10/100 = .1. Since the labor/
output ratio is, by assumption, the same for all levels of output, the demand for
labor can be written as
= 4,Q, (5.3)
and the demand for capital as
K = a,Q. (5.4)
Thus, if the projected level of output in Figure 5.3 is Q; = 300, the forecast of
employment in that year would be L = .2(300) = 60 workers.
The demand for labor function in Equation 5.3 can be used to predict the
total amount of labor required to produce a given level of output. For many pur-
poses, it is also desirable to be able to forecast labor requirements for detailed
industry and occupational groups. This can be accomplished by defining a more
disaggregated input-output coefficient a, ;= L;/Q, where L; equals the amount
of labor in the ith occupation or industry (machinists or basic steel) needed to
produce one unit of output (an automobile).
Forecasting with The following example illustrates how the input-output model can be used. In
the Input-Output 1983 Congress passed a law raising the federal gasoline tax by 5 cents a gallon.
Model The revenues from this tax were to be used for new highway and mass transit
construction projects. How many new jobs would be created by this expen-
diture? The first step is to estimate the amount of construction expenditure (Q)
that would be funded by the gas tax. At the time this was reported to be $5.5 bil-
lion. The second step is to calculate the labor input-output coefficient (a,). For
highway construction, the Bureau of Labor Statistics estimated that the ratio
L/Q in 1976 was 80.7 hours of employment per $1,000 of construction expen-
diture.*’ Since the input-output coefficient is in terms of 1976 dollars, it is neces-
sary to deflate the 1983 expenditure level of $5.5 billion by the increase in con-
struction prices between 1976 and 1983, yielding a figure of $3.43 billion. The
final step is to plug the estimate of a, (80.7 hours) and Q ($3.43 billion) into
Equation 5.3. The predicted increase in employee hours from the additional
construction expenditure is 276.8 million hours. Assuming a full-time employee
works 1,900 hours per year, this figure translates. into a predicted increase of
146,000 new jobs in construction and related industries.8
The input-output model offers a relatively simple method for forecasting fu-
ture employment levels. The input-output model’s simplicity is also its weak-
ness, however. In particular, there are three potential sources of error in making
employment projections with it. The first, and the one the Bureau of Labor Sta-
tistics (BLS) has found to be the most serious, is inaccurate estimates of the
input-output coefficients themselves.’”? Obviously, if the data used to construct
the coefficients are inaccurate due to measurement error, the projections derived
from the model will also be inaccurate.
The second most serious source of error found by the BLS is caused by changes
over time in the value of the input-output coefficients. The “naive” version of
the input-output model, as in Equation 5.3, assumes that the input-output co-
efficient is a constant. For this to be true, over the forecast period there must be
no improvement in technology and no capital/labor substitution due to chang-
ing factor prices. Studies that have used the input-output model to explain his-
torical trends in employment have found, however, that the input-output coeff-
cients actually change a good deal over a period of a decade or more, causing the
forecasts from a naive version of the model to have a large amount of error.”
In its forecasts, therefore, the BLS adjusts the input-output coefficients to take
into account secular trends in productivity, technological change, and other
such factors.
Finally, the BLS found that the least important source of error in its employ-
ment forecasts involves incorrect estimates of the future level of output (the as-
sumed value of Q in Equation 5.3). These three factors do not invalidate the
usefulness of the input-output model, but they do suggest that users of employ-
ment forecasts be aware of the uncertainties involved in the forecasting process.
7Robert J. Prier, “Labor and Material Requirements for Federally Aided Highways,” Monthly La-
bor Review 102, no. 12 (December 1979): 29-34.
28This is counterbalanced by a decline in employment in other industries, however, since the gas
tax reduced consumption spending by households. If these other industries were more labor inten-
sive, a net decrease in employment would occur in the economy.
Ronald E. Kutscher, “New Economic Projections through 1990—An Overview,” Monthly La-
bor Review 104, no. 8 (August 1981): 9-17.
30This is shown in Richard B. Freeman, “Manpower Requirements and Substitution Analysis
of Labor Skills: A Synthesis,” in Ronald G. Ehrenberg, ed., Research in Labor Economics, vol. 1
(Greenwich, Conn.: JAI Press, 1977): 151-184.
SUMMARY In the long run, the firm has more flexibility in adjusting its employment needs
to the relative prices of capital and labor, since it is free to vary both factor in-
puts. One result of this greater flexibility is that the long-run labor demand
curve is more elastic than the short-run labor demand curve. A rise in the wage
rate, for example, leads to a reduction in employment in the short-run because
of the scale effect as the firm cuts its level of production, and a further reduction
in the long run because of the substitution effect as the firm replaces labor
with capital.
A second influence on labor demand in the long run is technological change
and productivity growth. The initial effect of technological change is to displace
labor as more efficient techniques of production allow firms to produce a unit of
output with less labor. Over the longer run, technological change and produc-
tivity growth also lead to more jobs in the economy, because they result in lower
prices and greater real income, fueling an increase in product demand and labor
demand. Technological change may have an adverse impact on workers in cer-
tain occupations by making their particular skills obsolete.
The final topic considered in this chapter is employment forecasting. To make
employment forecasts, economists frequently use the input-output model of la-
bor demand. The most important assumption of the input-output model is that
the technology of production allows zero substitution possibilities between capi-
tal and labor. Research shows that changes in factor prices and in the state of
technology do cause the input-output coefficients to change over time, requiring
that the forecaster adjust the projections for these potential biases.
GLOSSARY
Four laws of derived demand— Four specific fac- Labor productivity —The amount of output pro-
tors that determine the size of the elasticity of duced per employee-hour.
labor demand. Long-run demand curve for labor— A curve that
Input-output coefficients — The coefficients of depicts the relationship between the wage rate
the input-output model. The labor input-output and the firm’s desired level of employment, as-
coefficient is the amount of labor required per suming both labor and capital are variable.
unit of output; the capital input-output coefficient | Marginal rate of technical substitution (MRTS)
is the amount of capital required per unit of — The rate at which capital can be substituted
output. for labor, holding output constant.
Isocost line — A line that shows all the various Scale effect — The change in labor demand that
combinations of labor and capital that a firm results from the change in the firm’s optimal
can purchase with a particular amount of money level of production due to a wage increase or
expenditure, given the wage rate and cost of decrease.
capital. Substitution effect (of labor demand) — The
Isoquant—A curve that shows all the various change in labor demand that results from the
combinations of capital and labor that a firm substitution of capital for labor (or vice versa) in
can use to produce a particular level of output. response to a change in the wage rate.
Chapter 5. The Demand for Labor in the Long Run 223
REVIEW QUESTIONS
1. Demonstrate in a graph that if the produc- 3. The United Automobile Workers Union has
tion technology is of the fixed proportion type lobbied heavily for a “domestic content” bill that
(L-shaped isoquants), an increase in the wage rate would require American cars be produced with
will cause only a scale effect on labor demand and parts and components made in the United States.
no substitution effect. Will there be any difference What effect would passage of this bill have on the
between the short-run and long-run labor demand demand for labor in the auto industry? Consider
curves? both the scale and substitution effects.
2. Before the airline industry was deregulated, 4. One of the demands of railroads in recent la-
the Airline Pilots Association (ALPA) was gener- bor negotiations has been to eliminate the caboose
ally regarded as one of the most successful unions from trains, cutting train crews from four workers
in raising the wages of its members; the Interna- to three. The unions have resisted this because of
tional Ladies Garment Worker's Union (ILGWU), the loss of jobs. Are the unions right? Would re-
on the other hand, has had only limited success in moving the caboose cause a decline in total em-
raising wages. First, use the four laws of derived ployment? Consider the short-run and long-run
demand to explain this difference. Second, after effects. What factors does the answer depend on?
deregulation of the airline industry over 40 new 5. “If productivity growth occurs more rapidly
airlines began business. What was the likely effect than expected, then an employment forecast using
of this on the elasticity of demand for airline an input-output model will underestimate the ac-
pilots? Which law of derived demand does this tual level of future employment.” Is this statement
involve? true, false, or uncertain? Why?
ADDITIONAL READINGS
Bureau of Labor Statistics. “Projections 2000.” Hamermesh, Daniel S. “The Demand for Labor
Monthly Labor Review 110 (September 1987): in the Long Run.” In Orley Ashenfelter and
3-63. A series of articles that outline the bu- Richard Layard, eds. Handbook of Labor Eco-
reau’s methods for making long-term economic nomics. Amsterdam: North-Holland, 1986:
projections and the anticipated growth in em- 429-472. A comprehensive review of the the-
ployment by industry and occupation to the ory of labor demand and empirical estimates of
year 2000. the elasticity of demand.
Goldstein, Morris, and Robert S. Smith. “The Hunt, H. Allan. “Technological Change and Em-
Predicted Impact of the Black Lung Benefits Pro- ployment: Fears and Realities.” In Proceedings
gram on the Coal Industry.” In Orley Ashen- of the 1986 Annual Winter Meeting. Madison,
felter and James Blum, eds., Evaluating the Labor Wis.: Industrial Relations Research Associa-
Market Effects of Social Programs. Princeton, tion, 1986: 447-454. Discusses the impact of
N.J.: Princeton University Press, 1976. The computers and other types of automation on
study examines the extent to which coal mine employment and why fears about growing un-
operators would reduce employment if their pay- employment are exaggerated.
roll costs were increased due to a government Killingsworth, Mark R. “Substitution and Output
mandated black lung insurance program. A Effects on Labor Demand: Theory and Policy
good illustration of the scale and substitution Applications,” Journal of Human Resources 20
effects of labor demand. (Winter 1985): 142—152. Uses estimates of the
substitution and scale effects to analyze the em- Rees, Albert. The Economics of Trade Unions.
ployment effect of wage subsidies. Chicago: University of Chicago Press, 1962:
Miller, Ronald E., and Peter D. Blair. Input- Chapter 4. A nontechnical discussion of how
Output Analysis: Foundations and Extensions. the four laws of derived demand influence union
Englewood Cliffs, N.J.: Prentice-Hall, 1985. wage policy.
Provides an in-depth review of the input- Shaiken, Harley. Work Transformed: Automa-
output model and its applications to economic tion and Labor in the Computer Age. Lexington,
problems. Mass.: Lexington Books, 1984. An in-depth
Peitchinis, Stephen. Computer Technology and look at the employment impact of computer-
Employment. New York: St. Martin’s Press, integrated manufacturing techniques.
1983. Examines the impact of computers on the
occupational structure and the general level of
employment.
CSHUAGR.
TE R
The Determination
of Wages
In this chapter demand and supply are brought together in order to analyze the
process of wage determination. Wage determination is at the core of labor eco-
nomics since the structure of wages and the change in wages over time are re-
sponsible for efficiently allocating labor and maintaining a balance between
demand and supply in the market. The chapter begins by focusing on the deter-
mination of wages in perfectly competitive markets. The two major issues of
concern are how market forces determine an equilibrium wage rate for a particu-
lar type of labor and how the market responds to a condition of disequilibrium
caused by a change in demand or supply. The chapter then turns to a considera-
tion of wage determination in what economists call “imperfect markets.” Two
examples are monopsony labor markets (those with only one employer) and seg-
mented labor markets. The chapter also discusses three different ways in which
institutional forces affect wage rates. These are government mandated minimum
wages, labor unions, and internal labor markets.
THE PATTERN Industries, occupations, and geographic areas differ markedly in terms both of
OF WAGES the levels of wage rates and in the changes in wages over time. This diversity is
illustrated in Table 6.1. Shown in column (a) is the level of average hourly earn-
ings in 1987 for production workers in the U.S. nonagricultural economy and
in nine individual three-digit SIC industries. Shown in column (b) is the per-
centage change in average hourly earnings between 1977 and 1987 in these
industries.
Before examining these data, it is useful to briefly discuss the exact meaning of
average hourly earnings, how the data are collected, and the concept of SIC
(Standard Industrial Classification) numbers. To economists, the meaning of
the term “wage rate” is intuitively obvious—it is the price of labor per hour of
work. Measuring the actual wage rate can be difficult, however. One complica-
tion is that firms use several different types of payment schemes. While some
employees are paid on an hourly basis, others receive salaries or are paid on a
piece-rate basis. Whichever method is used, it is possible to calculate the aver-
225
226 Chapter 6 The Determination of Wages
a
TABLE 6.1 Average Percent
Average Hourly Hourly Change
Earnings, in Earnings,
Earnings, 1987, and 1987 1977-1987
Percent Change in sic Industry (a) (b)
Earnings, 1977-1987, — Nonagricultural economy $8.98 71.0%
in Nine Industries 524 Gas stations 5.51 60.2
314 Leather footwear 5.76 66.5
225 Knitting mills 6.54 76.2
806 Hospitals 9.85 113.7
271 Newspapers 10.15 52.2
421 Trucking 11.00 55.3
175 Carpentry 12.64 65.4
331 Basic steel 13.84 65.5
120 Coal 15.75 91.1
source: Bureau of Labor Statistics, Employment and Earnings (March 1988 and March 1978): Table C—2.
age compensation received per hour by dividing weekly earnings by weekly hours
of work. Doing so yields average hourly earnings. Do average hourly earnings mea-
sure the price of labor? Unfortunately, the answer is no for the data do not cap-
ture all forms of employee compensation, particularly fringe benefits. This com-
plication can be dealt with by calculating total compensation costs
per hour, where
total compensation includes both earnings and the dollar value of fringe benefits
pro-rated on an hourly basis. While this measure of labor cost comes closest to
the economists’ notion of the price of labor, the drawback is that data on fringe
SSS
benefits are often unavailable, making average hourly earnings the next best
measure of the wage rate.'
A second issue is the source of the data on average hourly earnings. As dis-
cussed in Chapter 3, the data on employment, unemployment, and labor force
participation are derived from the CPS or “household” survey. The data on aver-
age hourly earnings, as well as data on employment, hours of work, and produc-
tivity by detailed industry, are obtained from the “establishment” survey. This
survey, conducted monthly by the Bureau of Labor Statistics (BLS), is done by
mail questionnaire and includes in its sample establishments (industrial plants or
places of business) that employ over 40 percent of all workers in the non-
agricultural economy. The data from both the household and establishment sur-
veys are presented in the monthly BLS publication Employment and Earnings.
The wage data in Employment and Earnings are listed by Standard Industrial.
Classification (SIC) number, beginning with SIC 10, Metal Mining, and end-
ing with SIC 89, Miscellaneous Services. Each establishment in the United
States is put in a two-digit SIC group according to the type of product produced.
An iron ore mine would be in SIC 10; an establishment producing any kind of
food type product would be in SIC 20, Food and Kindred Products; a trucking
establishment would be in SIC 42, Trucking and Warehousing. Each two-digit
‘For more detail on the measurement of labor cost, see Jack E. Triplett, ed., The Measurement of
Labor Cost (Chicago: University of Chicago Press, 1983).
Chapter 6 The Determination of Wages yaphy/|
group is then broken down into individual three-digit SIC industries: all estab-
lishments in Metal Mining, for example, are subdivided into SIC 101, Iron Ore;
SIC 102, Copper Ore; and so on. SIC numbers thus provide a very useful way to
classify economic statistics such as wage and employment data.
The data in Table 6.1 show the great diversity in the behavior of wages among
these various industries. In 1987, the level of average hourly earnings varied
from a low of $5.5 Lin gasoline service stations to a high of $15.75 in bituminous
coal mining. The average of hourly earnings in the entire nonagricultural econ-
omy was $8.98. Likewise, between 1977 and 1987, wages grew much more rap-
idly in some industries than others. While the average rate of wage growth in
i the
economy was 71 percent, the rate varied from a low of 52 percent in newspapers
to 114 percent in hospitals.
What can account for the fact that average hourly earnings in coal mining
were 186 percent higher than in gas stations? Why did hourly earnings in hospi-
tals grow far more rapidly than in the newspaper industry? There is no single
answer to these questions since wages are influenced by a host of factors—the
strength of product demand, the capital intensity of production, the extent of
unionization, the skills and training required of the work forces, the nature of
the working conditions, and many others. Despite the complexity of the subject,
economic theory can explain much of this disparate behavior, as this and suc-
ceeding chapters attempt to show.
WAGE The starting point in developing the theory of wages is the model of perfect
DETERMINATION competition. This model best illustrates how market forces operating through
IN COMPETITIVE labor demand and labor supply interact to determine the level of wages and em-
MARKETS ployment. There are five key assumptions in the model of perfect competition:
1. Business firms seek to maximize dollar profits, and workers seek to maximize
utility. ;
2. Workers and firms have perfect information about wages and job opportuni-
ties in the labor market.
3. Workers in the labor market are identical with respect to skills and produc-
tivity; jobs offered by firms are identical with respect to working conditions
and other nonwage attributes.
4. The labor market is composed of many individual firms on the buyer's side of
the market and many workers on the seller’s side. The workers do not belong
to unions, and firms do not collude.
5. All jobs in the labor market are open to competitionby workers, institu-
no
tional barriers inhibit mobility of workers from one job to another (for
instance, seniority provisions and internal hiring rules). Costs of mobility
are zero.
Given these assumptions, the perfectly competitive model gives rise to one of
the most important predictions in labor economics—the law of one wage. The
law of one wage states that in a competitive labor market the competition be-
tween buyers and sellers will result in the establishment in the market of one
FIGURE 6.1 The Determination of Wages in a Perfectly Competitive Market

If the labor market—graph (a)—is perfectly competitive, the equilibrium wage will be W, where demand and supply are equal. At
a wage of W,, demand (point A) is greater than supply (point B), and competition will force the wage up. At a wage of W,, there
is an excess supply of labor (point C — point D) and the wage will fall until W, is reached. Given the wage W in the market, the
individual firm—graph (b)—can hire all the labor it wants at that wage, as illustrated by its perfectly elastic supply curve Si. If it
pays a wage less than W, (point G), it will lose all its workers; if it pays more than We, the competitive firm will lose profits arid
be forced out of business.
(a) The Market (b) The Individual Firm
Lo Le L, Labor
(L)
uniform “going” wage rate that will be paid by all the firms and received by all
the workers.’ ae
The determination of wages in a perfectly competitive market and the reason-
ing behind the law of one wage are illustrated in Figure 6.1. Represented in
graph (a) is the market for some particular type of labor; graph (b) represents one
individual firm in the market. The wage rate is determined by the interaction of
demand andsupply in the overall market.’ As shown in Chapter 1, at any wage
higher than the-equitibrium-wage We(say W,) a surplus of labor will occur
(point C — point D), resulting in a bidding down of wages until demand and
supply are equal. Likewise at any wage below W, (say W,) a shortage of labor
occurs (point A — point B), and wages will be bid up as companies compete for
additional workers. Only at W, will the labor market be at an equilibrium or
state of balance.
*The law of one wage is the labor market analog of the more familiar law of one price in competi-
tive product markets. The reasoning behind the law of one wage is discussed in Alfred Marshall,
Principles of Economics, 9th ed. (New York: Macmillan, 1961): 546-550.
‘Chapter 2 showed that the labor supply curve for a demographic group may be positively or
negatively sloped. In an individual labor market, however, the supply curve is always positively
sloped. A rise in the wage for truck drivers in Chicago, for example, will induce a greater supply of
truck drivers as persons move from other cities or switch occupations in order to obtain these higher
paying jobs.
Chapter 6 The Determination of Wages 229
The process of wage determination for one individual firm in the market is
illustrated in graph (b). What wage does the firm have to pay for labor? The
answer_is-W,, given by the intersection at point F of the firm’s downward sloping
demand. (marginal revenue product) curve D, and the horizontal or perfectly
elastic supply curve 8;. The reasonis that the wage W, is determined by the
beer (ecormeion tercen the aggregate of firms and workers in the market
(graph a) and the individual competitive firm, because it is so small relative to
the entire market, cannot pay less than W, and still attract labor, nor does it
have to pay more than W, to be able to obtain all the labor it needs. In the
language of economics, the competitive firm is a “wage taker”; it has to pay the
going wage, but can hire as much labor as it needs without paying more. This is
shown in graph (b) by the horizontal or perfectly elastic supply curve of labor to
the firm, S,. Given its downward sloping demand curve D, and the market de-
termined wage of W,, the firm’s optimal level of employment is L,.
According to the law of one wage, in equilibrium all workers in this labor
market will receive and all firms will pay the same wage of W,. It is important to
understand why. What would happen in the labor market pictured in Figure 6. 1,
for example if one firm decided to only pay W, (less than W,), while a second
firm decided to pay W, (more than W;)?
Given the assumptions of perfect information, maximizing behavior, and free
mobility, competitive pressure will cause both firms to ultimately change their
rates of pay to W,. At the wage of W, the supply of labor to the low-wage firm
will be zero, shown as point G in graph (b). All of that firm’s workers would quit
and seek employment at other firms that are paying more for the same job. To
attract and keep its work force, the low-wage firm would be forced by the pres-
sure of labor mobility to raise its wage to Wy.
An opposite situation faces the high-wage firm paying W,. This firm will have
a long line of job applicants wanting to work there. The problem for this firm is
that by paying a higher wage its labor costs are also higher than other firms, and
it will earn less profit. While this situation may be tolerable in the short run, in
the long run in a perfectly competitive product market this firm will eventually
be forced out of business by its lower-cost rivals. The pressure to maximize profits
in order to survive will thus force this high-wage firm to reduce its wage from W,
to W,. Thus, as predicted by the law of one wage, competition will ensure that
in equilibrium only one going wage is observed in each particular market for
labor.
Empirical Evidence
Wages of Secretaries
To what extent do wages in the labor market correspond to the predicted one
wage of competitive theory? Evidence on this issue is presented in Figure 6.2 for
workers in a highly competitive labor market, the market for clerical workers.
The data show the distribution of weekly earnings for full-time manufacturing
workers classified as Secretary I and Secretary IV in Chicago in May, 1986.
FIGURE 6.2 The Distribution of Workers by Weekly Earnings

in Secretary | and IV Occupations, Chicago, 1986
Percent in
earnings
class
30
25
20
Secretary |
15
10 Secretary IV
$200$220$240$260 $280 $300 $320 $340 $360$380$400 $420$440$460 $480 $500$520$540 Weekly
-219 -239 -259 -279 -299 -319 -339 -359 -379 -399 -419 -439 -459 -479 -499 -519 -539 -559 earnings
source: Bureau of Labor Statistics, Area Wage Survey, Chicago, Illinois, March 1986, Bulletin 3059-9 (Washington: G.P.0., 1986): Table A-1.
These data seem to offer only partial support for the predictions of the com-
petitive model. For each occupation, and particularly for the Secretary I work-
ers, the earnings distributions do trace out a distinct bell-shaped curve showing
that earnings for each occupation are centered around a specific level of pay.
Thus, in the Secretary I occupation, the single greatest number of workers (21
percent) received $300 to $319 per week; for the Secretary IV occupation, the
largest group of workers (16 percent) earned $400 to $419 per week. This result
suggests that rates of pay in each occupation are not randomly determined, but
rather are shaped by forces such as those envisioned in the competitive model.
Given this, it is also evident from Figure 6.2 that the earnings for workers in
each occupation, and particularly for the more skilled Secretary IV classifica-
tion, exhibit a good deal of dispersion about the mean.‘ In the Secretary I oc-
cupation, 38 percent of the workers had earnings below $260 per week or above
$340 per week. The dispersion in earnings is significantly greater for workers
who were classified as Secretary [V; 60 percent of these workers had earnings
*Although the sample is limited to full-time workers, part of the dispersion in weekly earnings
may be due to differences in hours worked. A detailed study by Francine Blau, Equal Pay in the Office
(Lexington, Mass.: Lexington Books, 1977), that used data on hourly earnings of clerical workers
also found large wage dispersion among firms, however.
that diverged $40 or more from the mean level of weekly earnings for the
occupation.
These data suggest that the labor market gives rise to clear differences in aver-
age tates of pay among occupations, but within occupations a good deal of disper-
sion in earnings separates individual workers. From this it must be concluded
thatthe law ofone wage is refuted if literally interpreted as requiring a single rate
of pay for each worker in the market; as a description of the central tendency for
wages in each occupation to be grouped about a common rate, however, the law
serves as a good approximation for competitive markets such as secretaries.
—é
Market It is worthwhile to pursue the example of secretaries further and ask why the
Imperfections labor market did ngt give rise to the predicted one wage. One possible reason is
because the law of one wage describes a situation of long-run equilibrium, while
real-world labor markets are in a constant state of change. A second reason is
because the market for secretaries in Chicago violated one or more of the five
assumptions of the perfectly competitive model. A factor or circumstance that
causes a market to diverge from the perfectly competitive ideal is called a market
imperfection. The more serious the imperfections in a labor market, the more
the outcome of the wage determination process will diverge from the predicted
outcome of competitive theory.
What market imperfections account for the wide dispersion in earnings of sec-
retaries exhibited in Figure 6.2? One possible imperfection would be if the fourth
assumption of the competitive model were violated: for instance, if many secre-
taries belonged to a union or if there were relatively few firms on the buyer's side
of the market. This is clearly not the case in the Chicago market, however, and
attention must be focused elsewhere. Another possibility is that the fifth as-
sumption (no barriers to mobility) is violated. If strict seniority provisions or in-
house promotion rules were an important feature of the labor market for secre-
taries, for example, the competitive forces of labor mobility would be seriously
impeded, allowing a dispersion in wage to persist. It is probably the case,
however, that competition for jobs is.relatively wide-open, particularly for the
Secretary I occupation, and thus this market imperfection is not likely to be
significant.
Heterogeneity of Workers and Jobs One assumption of the perfectly com-
petitive model that is likely to be in error with respect to the market for secre-
taries is the third assumption, that-all-workers and jabs are identical. Competi-
tive theory predicts that a single going wage will prevail in the market for a set of
homogeneous jabs.and workers. It is reasonably certain, however, that workers
and jobs in the secretary’s market are heterogeneous or “differentiated,” introduc-
ing an imperfection into the market. One source of heterogeneity is with respect
to the supply of labor offered to firms. Despite the fact that our example exam-
ines earnings for two narrowly defined occupations, not all workers classified as a
Secretary I or Secretary IV are likely to be of the same skill and productivity
level, reflected by differences among them in years of education, experience, job
SZ Chapter 6 The Determination of Wages
skills, and innate ability, for example. Part of the dispersion in earnings in Figure
6.2, therefore, may reflect the fact that employers pay different wages to workers
of different productivity. This result is quite consistent with the law of one wage
if the law is slightly reinterpreted; that is, in terms of efficiency wages (wage
rates per unit of work performed) the labor market may still give rise to equality
of earnings once differences in worker productivity are taken into account.’
A related reason for the dispersion in earnings is that not all jobs in the Secre-
tary I or Secretary IV occupations are identical in the Chicago labor market.
Some jobs will have more pleasant working conditions, shorter commuting dis-
tances, or more fringe benefits while other jobs will have just the opposite. As
shown in more detail in Chapter 8, workers presumably would accept a lower
wage or “compensating differential” to obtain the job with good working condi-
tions, but would demand a higher wage to take the same secretarial job with
poor working conditions, leading to a pattern of earnings as shown in Figure 6.2.
Albert Rees and George Shultz, in a detailed study of 12 different occupations
in the Chicago area, attempted to determine to what extent these factors could
account for the apparent discrepancy between the prediction of the theory and
the actual observed pattern of wages.° In their study they controlled for differ-
ences among workers with respect to age, sex, race, seniority, experience, educa-
tion, commuting distance, type of neighborhood (where the job was located),
and type of industry. They found for typists and keypunch operators that differ-
ences in these variables among individual workers could account for, respec-
tively, 47 and 55 percent of the difference in the wage rates these workers were
paid, substantially reducing the actual dispersion in earnings.’
If the wage data for secretaries in Figure 6.2 could be standardized as Rees and
Schultz did in their study, the variance (relative spread) of wage rates around the |
mean of each frequency distribution would be narrowed considerably, perhaps by
50 percent or more. Decreasing the variance of each distribution by even one-
half still leaves a dispersion in wage rates for individual workers that hardly fits
the predicted one wage of competitive theory. What else can explain this appar-
ent deviation between theory and fact?
Imperfect Information Another assumption of the competitive model that is
violated in this example is the second, the existence of perfect information. Im-
perfect information can lead to a dispersion.in wage rates in two ways. The first’
has to do with the job search process. If workers in the Secretary I or Secretary
IV occupations knew the wages paid by all firms in the Chicago market, they
An unfortunate source of confusion comes from the two different meanings of the term “eff-
ciency wage” in labor economics. The first is as given above; the second denotes the idea that em-
ployee work effort or “efficiency” is positively related to the wage rate paid by the firm. This second
meaning is discussed at several points in this and subsequent chapters, although it is not formally
identified with efficiency wage theory until Chapter 13.
*Albert Rees and George P. Shultz, Workers and Wages in an Urban Labor Market (Chicago: Uni-
versity of Chicago Press, 1970).
‘Although controlling for sex and race reduces measured wage dispersion, these variables may
reflect discriminatory differences in wages rather than productivity differentials. Evidence on this
point is provided in Blau, Equal Pay in the Office.
would flock ta.the high-wage firms and shun the low-wage firms, driving wages
to_equality in the market just as the theory predicts. As emphasized by Nobel
laureate George Stigler, however, information about wages, working conditions,
and job openings arenot.announced to workers like prices at an auction; rather,
acquiring information about these factors requires a process of job search by
workers as they sequentially contact one employer and then another in the labor
market.®
The theory of job search is examined in Chapter 13. The essential point to
note here is that the acquisition of information in the labor market through job
search is costly, in terms of both direct out-of-pocket costs and the opportunity
cost of the time devoted to it. While each worker would like to obtain the high-
est paying job possible, beyond some point the additional search costs from con-
tacting yet another firm will surpass the probability of finding a higher wage
offer. Rather than searching in the labor market until the highest wage is found,
the worker will find it profitable to search only until he or she finds a job paying a
wage equal to or greater than some minimum acceptable wage. With imperfect
information, in an otherwise perfectly competitive market, two identical work-
ers may be paid different wage rates. Part of the dispersion in earnings of secre-
taries in Figure 6.2 reflects, therefore, the fact that imperfect information and
search costs prevent the process of labor mobility from fully competing away dif-
ferences in wages among similar jobs in the labor market.
Imperfect information can also lead to a dispersion of wages by affecting the
pay policies of employers. In a world of perfect information, a firm would know
the performance level_of each secretary it employs and could set the wage in
proportion to his or her productivity. In the real world, however, firms can only
imperfectly monitoran employee's job performance, both because supervisory
staff is costly and because many aspects of job performance (e.g., effort and dili-
gence) are difficult to measure. The problem posed for employers in this situa-
tion is that both hourly and salaried workers are faced with the temptation to
work at less than peak efficiency since the employer cannot fully detect loafing or
malfeasance. How can the employer combat this problem? One solution is to
deliberately pay a.wage above the market level, The higher the wage, the more a
secretary stands to lose from being fired for poor job performance and, thus, the
harder he or she will work. A dispersion in wages for a specific occupation such
as Secretary I will emerge, in turn, if firms differ either in their ability to monitor
employee performance or in the amount of profit that is at risk from acts of mal-
feasance by workers. This partially explains the fact that workers in any given
occupation generally tend to be paid more in large-sized plants or companies
relative to smaller ones.? The management in large-size organizations finds it
8George J. Stigler, “Information in the Labor Market,” Journal of Political Economy 70, pt. 2 (Oc-
tober 1962): 94-105.
°Numerous studies find that wage rates vary positively with the size of the plant or company,
other things being equal. See Lucia F. Dunn, “The Effects of Firm Size on Wages, Fringe Benefits,
and Worker Disutility,” in H. Goldschmid, ed., The Impact of the Modern Corporation (New York:
Columbia University Press, 1984): 5-58; and Charles Brown and James Medoff, “The Employer Size
Wage Effect,” Harvard Institute of Economic Research Discussion Paper 1202 (Cambridge, Mass.:
Harvard University, 1986).
more difficult to detect shirking on the job and, accordingly, pays a higher wage
as a means to increase the incentive for employees to work diligently.'° In this
case, both efficiency wages and money wages will systematically differ across
employers.
Nonmaximizing Behavior A final imperfection in the labor market may be if
the first assumption of the perfectly competitive model is violated, that is, if
firms do not always maximize profits.
The perfectly competitive model assumes that firms always act to maximize
profits. Given this, a firm will never pay workers a wage higher than the mini-
necessary
mum to attract a sufficient supply of labor since to do so would raise
labor costs and reduce profits. This motivation to maximize profits is reinforced
not only by the quest for pecuniary gain on the part of the owners of the firm,
but also by the threat of bankruptcy in the long run if the firm allows its labor
costs to exceed those of its rivals in a highly competitive industry.
While the maximization assumption is central to economic theory, some
economists argue that firms frequently do not maximize profits, but rather satis-
fice with respect to profits. "' Proponents of the theory of satisficing behavior pro-
vide two reasons why firms do not always maximize profits. The first is that the
separation of ownership and control in the modern corporation allows the man-
agement to pursue other goals (for example, promoting a good company image
or avoiding confrontation with unions) that may be at the expense of maximum
profits as desired by the shareholders. A second reason is that firms in
oligopolistic or regulated industries may not face heavy competitive pressure and
do not have the incentive to hold down wages and other costs to a minimum in
order to survive.
The consequences of satisficing behavior are that some firms_may consciously
pay_wages higher than the minimum going rate, giving rise to a dispersion of
wage rates in the market. In a classic study of 50 manufacturing firms in the New
Haven, Connecticut, area, for example, Lloyd Reynolds found that firms had a
considerable latitude in their pay policies and that the larger, more profitable
firms paid substantially higher wages for a given grade of labor relative to their
smaller, less profitable counterparts.'? Reynolds's explanation for this dispersion
in wage rates was in part that the companies with high profits were generally
those having barriers to entry in the product market. A barrier to entry is some
factor that impedes the ability of new firms to enter a market, usually resulting in
a high level of market concentration (few firms in the market). In oligopolistic
industries in manufacturing, the main barrier to entry is large capital require-
ments; in nonmanufacturing industries such as trucking or airlines (prior to de-
regulation in 1979), the principal entry barrier was the difficulty firms had in
"See Lawrence Katz, “Efficiency Wage Theories: A Partial Evaluation,” in Stanley Fischer, ed.,
NBER Macroeconomics Annual 1986 (Cambridge, Mass.: MIT Press, 1987): 235-276.
"Herbert Simon, “Rational Decision Making in Business Organizations,” American Economic Re-
view 69, no. 4 (September 1979): 493-512.
“Lloyd G. Reynolds, The Structure of Labor Markets (New York: Harper & Row, 1951), Chap-
ter 9.
FIGURE 6.3. The Area of lndetpemmnaney in wage Rates

The area of indeterminancy
in wages is represented by
the band extending from
the upper limit W, to the
lower limit W,. Most firms
~ will pay close to the going
market wage W,, but some
will be in either tail of the
wage distribution as high-
wage or low-wage firms.
The more closely the labor
market fits the perfectly
competitive ideal, the
smaller will be the area of
indeterminancy.
to eho ay
gaining regulatory permission to compete in new markets. In either case, bar-

riers to_entry allow firms in concentrated or regulated industries to pay wages
higher than the market level, since higher labor costs can more easily be passed
on in the form of higher pri ithout precipitating the entry of new, lower cost
rivals. These high wage firms will also be able to attract the best “quality” work-
ers in the labor market, making efficiency wages more equal than money wage
rates. Whether the increased productivity of workers employed by high wage
firms completely offsets their higher rates of pay has been investigated in a num-
ber of studies. While evidence has been found on both sides of this issue, the
most recent studies find that workers in concentrated and regulated industries do
receive a wage premium that cannot be totally accounted for by higher worker
productivity. ”
A More Realistic This discussion of the wages of secretaries suggests that a more realistic represen-
Model tation of the wage determination process would resemble that given in Figure
6.3. If the labor market satisfied all five assumptions of the perfectly competitive
model, the equilibrium wage in the market would be W;, and this wage rate
would be the one and only wage paid by each firm in the market. Most real-
world labor markets, however, feature imperfections such as limited informa-
tion, heterogeneous workers and jobs, and nonmaximizing behavior, that pre-
vent the forces of competition and labor mobility from completely eliminating
all wage differentials. The dispersion of wage rates that results is illustrated in
Figure 6.3 by the band of wage rates bounded on the top by an upper limit W,)
3 See John E. Kwoka, Jr., “Monopoly, Plant, and Union Effects on Worker Wages,” Industrial and
Labor Relations Review 36, no. 2 (January 1983): 251-257; and John S. Heywood, “Labor Quality
and the Concentration—Earnings Hypothesis,” Review of Economics and Statistics 68 (May 1986):
342-346. Besides being able to obtain higher quality workers, high-wage firms will also incur lower
search costs and lower costs from turnover, further reducing the dispersion in hourly labor cost.
and on the bottom by a lower limit W,. The firm cannot pay more than Wy
because to do so would lower its profits below the minimum necessary level; like-
wise it cannot pay below W, because it would not be able to attract or keep a
work force. Within the band, however, is an area of indeterminacy where the
imperfections in the competitive process allow some room for discretion by indi-
vidual firms with respect to their pay policies."
As shown by the bell-shaped curve in Figure 6.3, most_firms will pay close to
the average market wage W,; others, however, will be in either tail of the distri-
bution as high-wage or low-wage firms, depending on their size, industry, profit-
ability, and management attitude toward employee compensation.” Of these
factors, research has found that a firm’s industry affiliation is a particularly im-
portant determinant of its position in the wage hierarchy. '® A good illustration of
this point is provided in a well-known study by John Dunlop in which he exam-
ined the wage rates paid to truck drivers by various firms in Boston in 1953."
Since driving a truck is a fairly standard type of job, it might be thought that all
truck drivers would earn a similar wage per hour. He found, however, that truck
drivers’ wages not only had a wide dispersion, but also varied systematically by
the industry of the particular firm. The lowest-paid drivers worked for scrap
metal firms and earned only $1.27 per hour while the highest-paid drivers deliv-
ered magazines and earned $2.49 per hour. He concluded that firms in a labor
market are arrayed along a series of distinct “wage contours” within the area of
indeterminacy, with there being a separate contour for firms in each industry. In
Figure 6.3, the horizontal line at W,, would be the wage contour for magazine
delivery firms, the line at W, would be the contour for scrap metal firms. Ac-
cording to Dunlop, the position of each contour in the area of indeterminacy is
largely determined by product market characteristics of the industry such as de-
grees of competition, firm size, profitability, and so on. .
'*See Richard A. Lester, “A Range Theory of Wage Differentials,” Industrial and Labor Relations
Review 5, no. 4 (July 1952): 483-500. The existence of an area of indeterminacy in wage rates in a
nonunion labor market was well-accepted by the neoinstitutional labor economists of the 1950s.
With the ascendancy of the neoclassical school in the 1960s and 1970s, the idea largely dropped
from sight in economic research. That the concept is of real-world importance, however, is sug-
gested by the prominent role it still retains in the more applied field of compensation management.
See Marc Wallace, Jr. and Charles Fay, Compensation Theory and Practice (Boston: Kent Publishing,
1983): 25-45.
A recent article in Business Week [“Airline Wages Are Set for a Long Slide” (April 9, 1984):
127-128] provides an excellent example of this. Because of airline deregulation, pilots’ wages at
established airlines have come under heavy downward pressure from new, nonunion carriers. Where
will be the new equilibrium wage for pilots? According to one industry economist, “The equilibrium
point at which wages finally settle will vary within perhaps a 20 percent range from one airline to
another, depending on their route structure, quality of management, economies of large scale opera-
tions, and marketing skills.”
'°See Alan Krueger and Lawrence Summers, “Reflections on the Interindustry Wage Structure,”
and William Dickens and Lawrence Katz, “Inter-Industry Wage Differences and Industry Character-
istics,” in Kevin Lang and Jonathan Leonard, eds., Unemployment and the Structure of Labor Markets
(New York: Basil Blackwell, 1987): 17-47, 48-89.
John Dunlop, “The Task of Contemporary Wage Theory,” in George Taylor and Frank Pierson,
eds., New Concepts in Wage Determination (New York: McGraw-Hill, 1957): 117-139.
The size of the area of indeterminancy in market wage rates will vary from one
labor pee. to ie depending on the de
egreeto which the market conforms
change or a market for a commodity such as cats the analle: will be the dis-
persion in wage rates. Perhaps the best example of this was documented in a
study by Melvin Reder of the starting salaries of MBA graduates from the Uni-
versity of Chicago.'* This market was found to have many of the characteristics
of an auction market (the buying and selling took place through one centralized
exchange, buyers and sellers had considerable information about each other,
and the product was fairly homogeneous) and thus competition resulted in a
very small dispersion in starting salaries, once productivity differences among
students (for example, differences in grade point average) were accounted for.
Many other labor markets, however, diverge in one or several respects from a
true auction market; the more they do so, the greater will be the area of indeter-
minacy in wages. Several studies, for example, have found a large degree of wage
dispersion among semiskilled factory workers such as machine operators.'? One
important reason for this finding (the internal labor market) is examined later in
this chapter.
CHANGES IN An important function of the labor market is to allocate labor to its most_effi-
WAGE RATES cient use. This task is gargantuan. On the demand side of the market, the de-
mand for labor by business firms is in a constant state of flux. Some business firms
are growing rapidly and need many additional employees, while others are laying
off workers because of declining sales. The supply side of the labor market is also
‘quite dynamic. Each year over one-and-a-half million people enter the labor
force looking for work, and several million more who are already employed
change jobs. Given the great heterogeneity of labor demand in terms of industry,
occupation, and geographic area, and the great heterogeneity of labor supply in
terms of education, skill, job preference, and desired geographic location, how
does the labor market efficiently match the job openings in firms with the indi-
viduals seeking work? As discussed-in Chapter 1, the answer to this question
involves two key processes: changes in wage rates and the process of labor mobil-
ity. This section examines in more detail how these processes work to efficiently
allocate labor.
Market Shown in Figure 6.4 are two labor markets. The demand curve for labor in each
Adjustments market is D,, and the supply curve is S,. Assuming each market is perfectly com-
to Disequilibrium petitive, the equilibrium wage and level of employment will be W,, L, (point
A). Assume that each market is perturbed by some event that shifts the demand
'8Melwin W. Reder, “An Analysis of a Small, Closely Observed Labor Market: Starting Salaries
for University of Chicago MBA's,” Journal of Business 51, no. 2 (1978): 263-297.
Robert Raimon, “The Indeterminateness of Wages of Semiskilled Workers,” Industrial and Labor
Relations Review 6, no. 2 (January 1953): 180-194.
a
FIGURE 6.4 The Change in Wages Due to Excess Demand or Supply in the Market
Graph (a) illustrates the market adjustment to a situation of excess demand. The original equilibrium is point A. Assuming the
demand curve shifts to D,, at the wage W, there is an excess demand for labor of L) — L; (point B — point A). As a result, wages
rise to W, (point C), but the net increase in employment is only to L3. Graph (b) illustrates a situation of excess supply. Assuming
the supply curve shifts to the right to S2, at the wage of W,, there is an excess supply of labor of L, — (point B — point A). As
iS
D,
Ly ets Ly aabes Li Leos ba Labor
(L) te
or supply curve for labor. How will the market regain an equilibrium so that
demand and supply are again equal? There are two cases to consider: first, the
market reaction to a situation of excess demand; and second, the reaction to a
situation of excess supply.
Excess Demand Starting from a point of equilibrium such as point A in Fig-

ure 6.4, a situation of excess demand in the labor market can arise from one of
two sources: an increase in labor demand or a decrease in labor supply. The for-
mer case is illustrated in graph (a). Assume for this example that a rise in con-
sumer spending has led to an increase in sales
for firms in this labor market, with
sales in other markets remaining constant. As a result, the demand curve for
labor will shift to the right from D, to Dj, illustrating that at the going wage of
W, the desired level of employment by firms has increased from L, to L; (point
A to B). At the wage W,, there is now an excess demand for labor of LP a a
condition that will cause firms to bid up wages as individual firms compete for
scarce labor. A new equilibrium will be achieved only when the wage rate has
risen to W, (point C) where demand is again equal to supply. At this new equi-
librium, the level of employment L; is higher than L, but less than L,. The
competitive model predicts that three things will happen in this labor market:
(1) the absglute dollar level of the wage will rise;(2) the wage zelativetowages in
other labor markets will rise; and (3) the level of employment will rise. If the
excess demand originated from a leftward shift of the supply curve instead, pre-
dictions (1) and (2) remain the same, but prediction (3) is reversed.
Excess Supply The second form that disequilibrium in the market can take
is excess supply. This may be caused by either a leftward shift of the demand
curve or a rightward shift of the supply curve. The latter case is illustrated in
graph (b) of Figure 6.4. Assume for this example that a sizable migration of job
seekers enters this labor market. The effect is a shift of the labor supply curve
from S, to S,, illustrating that at the going wage of W, the number of people
ee, work has increased from L, to L,(point A to B). At the wage W,, there
is an excess supply of labor of L, —Ly, bidding down wage rates in a competitive
market as unemployed workers compete for the scarce jobs. The new equilibrium
is at the lower wage of W, where the demand curve D, and supply curve S,
intersect (point C). The level of employment increases to L,. The competitive
model predicts that in this situation three dine ogee (1) the absolute
dollar level of wages will decline; (2) the wage rate relative to the wage in other
markets will decline; and (3) the level of employment will increase. If the excess
supply originated from a leftward shift of the demand curve, predictions (1) and
(2) remain the same, but prediction (3) is reversed.
Empirical Evidence
Wages of Oil Field and Packinghouse Workers
Do changes in demand and supply actually lead to changes in wages, as the com-
petitive model predicts? Some interesting evidence on this question is provided
by two examples discussed below. The first concerns the change in wages for oil
field workers; the second pertains to wage changes for packinghouse workers.
Oil Field Workers A marked increase occurred in the demand for labor in the
oil and gas extraction industry (SIC 13) during the 11-year period of 1970
through 1980. As shown in the first row of Table 6.2, the number of oil and gas
wells drilled in the United States doubled over the course of the decade, from
28,000 in 1970 to 61,000 in 1980, reflecting the large incentives given to do-
mestic exploration by both the rise in foreign oil prices and the deregulation of
~
TABLE 6.2 1970 1972 1974 1976 1978 1980

Wages and Oil and gas wells
Employment in the drilled (in 5
Oil and Gas thousands) 28 27 32 40 47 61
Average hourly
Extraction Industry, earnings $3.57 $4.04 $4.87 $5.85 $7.07 $8.59
1970-1980 Industry wage as
percent of
nonagricultural
wage 107% 111% 115% 120% 123% 129%
Production
workers (in
thousands) 178 177 203 237 291 381
EEE
source: Bureau of the Census, Statistical Abstract of the United States, 1984 (Washington: G.P.0., 1983): Table 1280;
Bureau of Labor Statistics, Employment and Earnings (Washington: G.P.0.): various issues.
domestic crude oil prices under President Carter. Since the demand for labor is
derived from the demand for the product, the increase in the number of drillings
over the decade led to a large increase in the demand for drilling crews, geolo-
gists, and petroleum engineers. This increase caused the demand curve for labor
to shift sharply to the right in this industry.”°
The competitive model makes three predictions concerning the change in
wages and employment in response to a situation of excess demand in the mar-
ket. The first is that the absolute level of wages in the industry should rise. This
prediction is borne out for this industry as shown by the data in the second row
of Table 6.2. Average hourly earnings in the oil and gas extraction industry were
$3.57 in 1970 and $8.59 in 1980, an increase of 133 percent. If the rise in prices
between 1970 and 1980 is subtracted out, earnings in real terms still advanced
significantly, from $3.67 in 1970 to $4.03 in 1980.
The second prediction of the competitive model is that wages in this industry
relative to wages in other industries should also increase. The data in the third
row confirm this prediction. They show the ratio of average hourly earnings in
the oil and gas extraction industry relative to average hourly earnings in the en-
tire nonagricultural economy. In 1970, wages received by oil and gas field work-
ers were 7 percent higher than the wage received by the average worker in the
economy; by 1980 this relative wage premium had risen to 29 percent. While
product demand increased across nearly all industries in the 1970s, the relative
increase was greater in the oil and gas extraction industry, leading to a rise in
relative wages as the competitive model predicts.
The third prediction of the competitive model is that the level of employment
should increase in the industry. This prediction is also confirmed by the data in
the fourth row of Table 6.2, which show the number of production workers em-
ployed in the oil and gas extraction industry. It is evident that employment ex-
panded sharply during this period, from 178,000 employees in 1970 to 381,000
in 1980.
The pattern of wages and employment in this particular industry over the
1970s matches quite closely the predictions of the competitive model. This ex-
ample also illustrates how changes in wages serve to efficiently allocate labor re-
sources in the economy. Given higher oil and gas prices, oil companies found it
to their advantage to increase domestic exploration of oil and gas. This explora-.
tion, however, was frequently in remote and inhospitable areas, such as the
north slope of Alaska, or offshore in the Gulf of Mexico. How could oil com-
panies find the thousands of additional workers they needed, and how could they
induce these workers to move to where the drilling was? The obvious, but never-
theless profound, answer is that wages rose until enough oil field workers in
Texas and Oklahoma and schoolteachers and unemployed auto workers in
Michigan, for example, found it to their advantage to voluntarily move to
Alaska or to seek work on the off-shore rigs in the Gulf of Mexico. Without any
“See Richard Greene, “Employment Trends in Energy Extraction,” Monthly Labor Review 104,
no. 5 (May 1981): 3-8.
TABLE 6.3 TTT
1958 1959 1960 1961

Wages and
Manufacturing employment 2,100 1,900 1,600 2,100
Employment in Fargo, Average hourly earnings,
North Dakota, before manufacturing $2.11 $2.14 $2.23 $2.41
and after the Armour Wage in Fargo as percent of wage
in Minnesota 97% 94% 94% 98%
Plant Shutdown a ee I
source: Bureau of Labor Statistics, Employment and Earnings: States and Areas 1939-1970, Bulletin 1370-8 (Wash-
ington: G.P.0., 1971).
centralized direction, the operation of the labor market brought about a redistri-
bution of labor resources that served the self-interest of both the oil companies
and workers, and the social interest of the country at large.
Packinghouse Workers The case of oil and gas field workers illustrates the
reaction of wages and employment to excess demand in the labor market. A sec-
ond example illustrating the opposite case of how wages and employment adjust
to excess labor supply is the shutdown of an Armour Company meatpacking
plant in July 1959 in the city of Fargo, North Dakota.?!
The adjustment of wages and employment in the Fargo labor market in re-
sponse to the plant shutdown is shown by the data in Table 6.3. Before the clo-
sure of the Armour plant, manufacturing employment in Fargo in 1958 was
2,100 (row one), and the level of average hourly earnings was $2.11 (row two).
The closing of the Armour plant resulted in an immediate loss of approximately
450 jobs. In terms of Figure 6.4, this loss is represented by a leftward shift of the
labor demand curve, illustrating that at the going wage of W, there was now a
smaller demand for labor in the market.
The competitive model predicts that three adjustments will take place in the
labor market in a disequilibrium situation such as this. The first prediction is
that employment will decline. This prediction is borne out by the data in row
one, which shows that manufacturing employment in Fargo declined from 2,100
in 1958 to 1,600 in 1960.
The second prediction is that the absolute dollar level of wages will decline.
This prediction is not borne out by the behavior of wages in the Fargo labor mar-
ket, as shown in row two. Wage ratés in Fargo actually rose from $2.11 in 1958
to $2.23 in 1960 and continued to increase in 1961. This increase is true even if
inflation is subtracted out. In real terms, wages rose from $2.11 in 1958 to $2.18
in 1960.
The third prediction of the competitive model is that the level of wages in
Fargo relative to wages in other labor markets should also decline. This predic-
tion is confirmed by the data in row three of Table 6.3. These data show the
ratio of wages in Fargo to the level of wages in the adjacent state of Minnesota,
assuming Minnesota represents the major alternative labor market for Fargo
1 This example is drawn from Richard C. Wilcox and Walter H. Franke, Unwanted Workers (New
York: The Free Press of Glencoe, 1963). Although somewhat dated, the Fargo example was chosen
because the decrease in labor demand was quite large relative to the size of the market, the Fargo
labor market was largely nonunion, and macroeconomic conditions were stable.
workers. In 1958, wages in Fargo were 97 percent of wages in Minnesota. After

the plant shutdown, relative wages in Fargo declined to 94 percent of Minnesota
wages in 1960, a pattern consistent with the prediction of the theory.”
The net result, then, is that in response to the excess supply of labor created
by the Armour plant shutdown, both relative wages and employment in Fargo
did decline as predicted by the competitive model. The absolute level of wages,
however, not only did not decline as the theory predicts, but instead continued
to increase.
The failure of money wage rates to decline in the face of an excess supply of
labor represents one of the most important deviations of actual labor market be-
havior from the predictions of competitive theory. When President Carter im-
posed the embargo on grain sales to the Soviet Union in 1979, the price of
wheat received by North Dakota farmers dropped precipitously in a matter of
days. Why, then, when the Armour plant closed did wages received by workers
in Fargo not also drop to clear the market?
The answer to this question revolves around differences in the two types of
markets. The wheat market is a classic example of an auction market. Prices rise
and fall as buyers and sellers bid against each other in an attempt to strike the
best bargain possible. The flexibility of prices is encouraged by several features of
the market and the good in question. Buyers and sellers of wheat, for example,
engage in constant “shopping around” in the market as they search for the best
price. Likewise, since the wheat is a standardized commodity, the only variable
of concern to both parties is the price. Finally, since wheat is an inert com-
modity, it does not care at what price it is bought and sold, nor does it care who
the buyer is or how many times it changes hands. The labor market, on the
other hand, is far different.” If the labor market were like a commodity market,
firms would auction jobs on a daily basis to the lowest bidders. Competition
among unemployed workers would result in the wage being bid down until
everyone who wanted a job had one. Sixty years ago factory labor was often
hired this way as a foreman would stand at the plant gate and offer work to those
outside at whatever wage the market would bear.** This type of buying and sell-
ing is seldom done in the labor market today for several reasons. One is the
growing prevalence of long-term jobs in the labor market. While the plant fore-
man 60 years ago might hire and fire workers on a daily or weekly basis, many
firms today would find this prohibitively expensive. Firms invest substantial sums
of money in their employees in the form of hiring and training costs. Thus, even
if an unemployed worker offers to work at a lower wage than an existing em-
ployee, the firm will generally not find it profitable to hire him or her. A second
* Wages in Fargo also declined relative to the U.S. manufacturing wage (from 100 percent in
1958 to 98 percent in 1960).
» See Daniel J. B. Mitchell, “Explanations of Wage Inflexibility: Institutions and Incentives,” in
Wilfred Beckman, ed., Wage Rigidity and Unemployment (Baltimore: Johns Hopkins University Press,
1986): 43-76.
See Sanford Jacoby, Employing Bureaucracy: Managers, Unions, and the Transformation of Work in
American Industry, 1900-1945 (New York: Columbia University Press, 1985).
FIGURE 6.5 Adjustment of the Labor Market to Excess Supply When Wages Are Rigid Downward
The original equilibrium is at point A with the demand curve D, and supply curve S,. Assuming the demand curve shifts to Dz, if
wages were flexible downward they would fall to W2, restoring equilibrium at point C. With inflexible wages, however, an excess
supply of labor will persist of L; — L, (point A — point B). Although the money wage remains at W,, this wage will decline relative
to wage rates in other labor markets. This decline in the relative wage restores equilibrium (point E) by causing people to move
out of this labor market, shifting the supply curve leftward to S,, and by inducing a flow of capital investment into the area, shifting
the demand curve rightward to D3.
Wage
(W)
kk Labor (L)
reason why money wages are downwardly inflexible is because unemployed work-
ers frequently refuse to lower their “asking wage,” preferring instead to remain
unemployed until a job opening is found at the desired wage rate. This reflects
the expectation of many workers that their loss of job is temporary, the psycho-
logical resistance of workers to accepting a wage less than what they were accus-
tomed to, and the availability of unemployment insurance benefits. For these
and other reasons, then, money wage rates typically do not fall even in the pres-
ence of considerable unemployment.
The Role of Relative Wage Adjustments If the level of money wages does not
fall in response to an excess supply of labor, is there an alternative route by
which the labor market can reach a new equilibrium? The answer is yes and
involves the role of relative wages and factor mobility. This process is illustrated
in Figure 6.5 for the case of the Fargo labor market.
Before the closure of the Armour plant, the labor demand curve in Fargo was
D,, the supply curve was S,, and the equilibrium wage and level of employment
were W, and L, (point A). When the Armour plant closed, the demand curve
for labor in Fargo shifted to the left from D, to D,, leading to an excess supply of
labor in the Fargo market of L, — L, (point A — point B). In a perfectly com-
IN « THE « NEWS
Market Adjustments to a Labor Shortage
Take Many Forms
Simple models sometimes yield extremely power- parting a certain amount of inertia to the equi-
ful insights. This is certainly true of the model of libriating process in the short run.
demand and supply. In response to an excess de- The article says that although much attention
mand for labor, for example, the demand and sup- has been focused on the growing scarcity of service
ply model predicts that the wage rate should rise in workers for restaurants, hotels, and stores, a far
the market until a new equilibrium is established. more serious labor shortage is emerging for skilled
The example of oilfield workers bears testimony blue-collar craftsmen ranging from machinists and
to the validity of this prediction. Unfortunately, electricians to shipbuilders and bricklayers. One
however, simple models can rarely capture the full example cited to illustrate the severity of the labor
range of reality, particularly in labor markets. This shortage of craftsmen was the Electric Boat Com-
is amply illustrated, for example, by a recent ar- pany in Groton, Connecticut. At the time the ar-
ticle in The Wall Street Journal about the severe la- ticle was written, Electric Boat Company had over
bor shortage of skilled craftsmen that existed in 300 openings for machinists, welders, pipefitters,
many parts of the country in 1987. The article and other skilled craftsmen. Competition was so
points out that firms attempt to deal with excess fierce for people with these skills that when 97
demand pressures in a variety of ways besides rais- craftsmen at a nearby plant were laid off, Electric
ing wages, and that while wages will eventually Boat offered jobs to 30 of the workers, but was able
rise in the face of a labor shortage, as the de- to hire only two.
mand/supply model predicts, they often respond To attract craftsmen, the Electric Boat Com-
to changing market conditions rather slowly, im- pany and other employers have adopted a number
petitive market, this excess supply of labor would have been eliminated by a fall
in the wage to W,, giving rise to a new equilibrium level of employment of L,
(point C). If wage rates are inflexible downward, however, this path to equi-
librium in the labor market is blocked. How is a new equilibrium restored?
Even if the level of money wages in Fargo did not decline, the relative wage -
did as wages in Fargo grew more slowly than wages in other, more buoyant labor
markets. The decline in the relative wage provided an incentive, in turn, for
some of the L,; — L, unemployed workers to migrate to other cities or states to
find work. The effect of this, in Figure 6.5, is to cause a shift of the supply curve
of labor to the left from S, to S,, depicting that at the prevailing wage of W,
there is less labor available in the Fargo market. Equilibrium would be restored
when the process of relative wage decline and out-migration of labor had caused
the supply curve to shift leftward until it intersected the demand curve D, at
point B. The decline in relative wages will also restore equilibrium in the labor
market through a second mechanism. The fall in the relative wage not only in-
duces an outflow of labor from the Fargo market but also an inflow of capital as
i
SSS
i ee eee eee ee
of recruitment practices that in a slack labor mar- tisen sharply as companies competed for labor.
ket they would not bother with. Electric Boat Surprisingly, however, such has not been the case.
Company, for example, placed help-wanted adver- According to the article, average hourly wage
tisements as far away as Sturgeon Bay, Wisconsin, rates for machinist jobs in 1987 ranged from $14 to
and sent company recruiters to western Pennsyl- $16, about the same as in 1982. Likewise, over the
vania in search of laid-off craftsmen from the steel previous 4 years bricklayers’ wages remained un-
industry. The company also sent representatives to changed, as did those of tool and die makers.
area high schools to discuss job opportunities with Allowing for inflation, real wages for craftsmen ac-
graduating seniors. tually declined.
Other companies have tried different tactics. How can a labor shortage and stable wages co-
The Laureno Lumber and Mill Work Company, exist? According to one well-known labor econ-
for example, extended its average workweek to omist, “What we'te seeing is the legacy of the
55 hours from 40. The company also offered new last economic recession and the ongoing pressure
recruits a $1,500 “hiring bonus,” despite its fear of foreign competition. Many companies have
that the bonuses would encourage long-time em- chosen to ignore issues of [labor] supply and de-
ployees to seek pay increases. The company also mand for fear that higher labor costs will make it
admitted that it had lowered its hiring standards more difficult to compete.” He goes on to say that
and was willing to take people with little experi- a number of companies decided to operate short-
ence as door makers and forklift truck operators, handed or institute mandatory overtime rather
despite the additional costs of training and lower- than raise wages, on the hope that the labor short-
quality production. age would prove to be temporary.
Given the scarcity of labor, one would think The article notes, however, that gradually the
that the wages of skilled craftsmen would have continued
firms decide to locate new plants in the area to take advantage of the lower price
of labor. The new plants shift the solid demand curve D, to the right to the
broken demand curve D,. The combination of labor outflow and capital inflow
works to restore equilibrium in the market, illustrated by the intersection of the
demand curve D, and supply curve S, at a wage level of W, and employment
level of L,; (point E).
While a change in relative wages is thus able to restore equilibrium in the
labor market given an initial situation of excess supply, several caveats concern-
ing this process should be pointed out. The first is that the process of labor out-
flow and capital inflow can take a relatively long time to restore equilibrium. In
Fargo, for example, 6 months after the plant shutdown, 44 percent of the pack-
inghouse workers who were laid off remained unemployed. This high rate of un-
employment reflects the slow rate of out-migration of workers from the Fargo
area and the even slower process of attracting new firms into the area.
: A second feature of the adjustment process that should not be glossed over
concerns the economic and social costs suffered by individuals and families in-
IN « THE « NEWS
pressure of excess demand was beginning to show fits. If the shortage remains severe for a sufficiently
up in craftsmen’s wages. The Laureno Lumber and long time, these large employers also begin to ex-
Mill Company, for example, couldn't get anyone perience high separation rates, forcing them to
to answer its ads at $6 an hour, so it raised its wage consider boosting pay rates, as well.
scale to $7 to $9 an hour. The pressure to raise
wages affects smaller companies first, the article
states, because they typically pay less and, thus, in
Source: “A Growing Shortage of Skilled Craftsmen Troubles
a labor shortage tend to lose their craftsmen to big- Some Firms,” The Wall Street Journal (September 14, 1987):
ger employers with higher wages and fringe bene- 1, 20.
|TR a eee Se a ES SS a
volved in the plant shutdown. The Armour workers who obtained employment
after the shutdown were working at new jobs that paid, on average, 23 percent
lower wages. For the 44 percent of the Armour work force that remained unem-
ployed after 6 months, family savings were depleted and, in some cases, homes
lost. The adverse impact of the shutdown was most strongly felt by older workers
whose job skills were suddenly obsolete and for whom the options of going back
to school or out-migration to a new city were far more difficult than for younger
workers. A recent government survey of workers who lost their jobs due to plant
closings in the 1979 to 1983 period found a similar pattern.”* Of the 5.1 million
workers with job tenure of 3 years or more who were displaced from their jobs
between 1979 and 1983, 2.0 million remained unemployed or had dropped out
of the labor force as of 1984, 45 percent of the displaced workers with new jobs
were earning less than on their old jobs, 35 percent of the displaced workers no
longer were covered by any form of health insurance, and workers over the age
of 44 suffered the longest bouts of employment and the largest drop in earnings.
Finally, an interesting question to ponder, and one that is the focus of Chap-
ter 13, is what happens to the equilibrating process if the decrease in demand for
labor occurs across all local labor markets rather than just one as in the case of
Fargo? While a decline in relative wages can restore equilibrium in one market
through the mobility of labor to other expanding job markets, if labor demand ‘
decreases in all markets together (such as in a recession or a depression) there is
an opportunity neither for changes in relative wages nor for workers to migrate
to other areas in search of better job opportunities. In this case the downward
rigidity of wage rates may prevent the labor market from reaching a new equi-
librium, leading to the possibility of persistent unemployment.
*See Paul O. Flaim and Ellen Sehgal, “Displaced Workers of 1979-83: How Well Have They
Fared?” Monthly Labor Review 108 (June 1985): 3-16. For a more recent case study of the impact ofa
plant closing on an individual community, see Charles Craypo and William I. Davisson, “Plant
Shutdown, Collective Bargaining, and Job and Employment Experiences of Displaced Brewery
Workers,” Labor Studies Journal 7, no. 3 (Winter 1983): 195-215.
TABLE 6.4 Percentage of Minimum W: age

Level and Coverage Date of pfedinatdy Relative to Average Hourly Wage
of the Minimum Minimum Covered in i Manufacturing
Wage, 1938-1981 Wage Change Minimum Wage Private Industry Before After
October 1938 $0.25 43.4% — 40.6%
October 1939 0.30 47.1 39.6% 47.6
October 1945 0.40 55.4 30.8 41.1
January 1950 0.75 53.4 28.7 53.7
March 1956 1.00 53.1 39.3 52.3
September 1961 1.15 62.1 43.1 49.6
September 1963 Ws 62.1 46.6 50.6
February 1967 1.40 75.3 44.8 50.2
February 1968 1.60 72.6 47.6 54.4
May 1974 2.00 83.7 37.8 47.3
January 1975 2.10 83.3 427 44.9
January 1976 2.30 83.8 41.7 45.6
January 1978 2.65 85.1 38.5 44.4
January 1979 2.90 85.1 40.8 44.6
January 1980 3.10 86.1 41.7 44.5
January 1981 3.35 86.0 40.1 43.3
source: Bureau of the Census, Statistical Abstract of the United States, 1982-1983, Table 677; Employment and
Earnings, various issues.
Policy Application
The Minimum Wage
An issue of long-standing controversy in labor economics is the minimum wage.
Ever since the minimum wage was first established, economists have been debat-
ing the benefits and costs of the law. This section uses the competitive model of
wage determination to analyze just what those benefits and costs are.
History of the The federal minimum wage was established in 1938 by the Fair Labor Standards
Minimum Wage Act (FLSA). As shown in Table 6.4, when first established the minimum wage
was set at 25 cents per hour and coyered only a minority (43 percent) of the
work force, primarily workers in larger firms involved in interstate commerce.
Over the years the FLSA has been amended a number of times to raise both the
level of the minimum wage and the number of workers covered by the law. In
1988 the minimum wage stood at $3.35 per hour and covered approximately 86
percent of the work force. The major groups of workers not now covered under
the federal minimum wage law are executive, administrative, and professional
personnel, employees of small retail firms, household workers, workers in certain
recreational industries, and agricultural workers.
The reason for the periodic upward adjustment of the minimum wage is illus-
trated in the last two columns of Table 6.4, which show the minimum wage as a
percentage of the average wage in manufacturing before and after the date of
, each amendment to the FLSA. Since the minimum wage is stated as a fixed
dollar amount per hour, the process of economic. growth and inflation. con-
tinually erodes the minimum wage as an effective floor on the wages and pur-
chasing power of low income workers. Between January 1981 (the last time the
minimum wage was increased) and January 1988, for example, the minimum
wage declined 31_percent in real terms. As a consequence, Congress has in the
past periodically raised the level of the minimum wage, generally to a level of
about 50.percent of the average manufacturing wage. The result is to give a saw-
toothed pattern to the relative value of the minimum wage, a pattern which will
be repeated if Congress passes legislation introduced in 1987 to raise the mini-
mum wage to $4.65 per hour.
Purpose of the In assessing the benefits and costs of the minimum wage, one important con-
Minimum Wage sideration is to what degree the minimum wage actually fulfills its intended
purposes.
As stated in the Fair Labor Standards Act, the primary purpose of the mini-
mum wage is the “maintenance of the minimum standard of living necessary for
health, efficiency and general well-being of workers.” The important point to
note is the phrase “minimum standard of living,” which pertains to annual in-
come, not the wage rate per se. Much like other government tax and transfer
programs (for example, AFDC and Social Security), the minimum wage is an
attempt to put a floor under the income of a particular subgroup of the popula-
tion, in this case the “working poor’—people with jobs who receive very low
wages.
Who are the working poor, and how many are affected by the minimum wage?
Two recent studies have analyzed these questions with data from the Current
Population Survey.” In 1986, 5.1 million workers, or about_9_percent of workers
who were paid an hourly wage, reported earning less than or equal to the mini-
mum wage of $3.35. This group of workers was comprised largely of young
people and women, the majority worked part-time, and most were employed in
service and sales occupations. If the minimum wage had been set at $4.65 in
1986, 17 percent of hourly workers would have been affected. It is apparent,
then, that a significant group of people work, yet receive wages sufficiently low
that they might still have annual incomes below the government's official pov-
erty line. In 1986, for example, the official poverty line for a nonfarm family of
four was $11,203. A head of a household with three dependents who earned the
minimum wage in 1986 and worked full-time (2,000 hours per year) would have
earned only $6,700.
It would appear from this example that if Congress intended the minimum
wage to insure a minimum standard of living, it has failed. The minimum wage
still leaves a head of a household working full-time far below an adequate level of
family income. The studies cited above found, however, th that in actuality only
one-fifth of the low-wage workers (earning $3.35 or less) were found to be living
in poverty households. The reason for this discrepancy is that many low-wage
si Ralph E. Smith and Bruce Vavricek, “The Minimum Wage: Its Relation to Incomes and Pov-
erty,” and Earl F. Mellor, “Workers at the Minimum Wage: Who They Are and the Jobs They
Hold,” Monthly Labor Review 110 (June and July 1987): 24-30 and 34-37.
workers are not single-earner, heads of households, but rather so-called “second-_
ary” earners, suchasas teenagers and spouses. “It was found, for example, that 72.
percent of all low--wage workers were either children or spouses of household
heads. In this case, while the individual incomes of these low-wage workers was
quite low, total family income was often above the poverty line because of earn-
ings of other ffamily members,p particularly males heads of households.
These data point out a critical issue for Congress in deciding on the appropri-
ate level of the minimum wage. For the minority of low-wage workers who are
single-earner family heads, particularly female family heads, the minimum wage
is sufficiently low that even with full-time work their annual income may still fall
far below the poverty line. For these people, the minimum wage fails to fulfill its
objective of insuring a minimum standard of living. To raise the minimum wage
further, however, might work to the detriment ofother low-wage workers such
as teenagers who do not have the minimum income requirements of a household
head and who might, at the higher minimum wage, be unable to find jobs
atvall:
Economic Analysis — Given the intended purpose of the minimum wage law, the competitive model
of the Minimum of wage determination can be used to identify the benefits and costs of the law
Wage and who the gainers and losers are likely to be. This analysis is illustrated in
Figure 6.6, graphs (a) and (b).
FIGURE 6.6 The Impact of a Minimum Wage on the Covered and Uncovered Sectors
Before the enactment of a minimum wage law, the wage rate in both the covered—graph (a)—and uncovered—graph (b)—
sectors is W,. lf employers in the covered sector are required to pay a minimum wage of W,, employment will decline from
L, to L, (point A to B), but L3 (point C) people will now want to work. If the L; — L, people continue to search for work in the
covered sector, unemployment will rise by that amount. Some of these workers, however, will probably migrate to the uncovered
sector, shifting the labor supply curve from Sy to Sy,. lf wages are flexible downward, the wage in the uncovered sector will fall
to W; (point F).
(a) Covered Sector _ (b) Uncovered Sector
-Ls Labor - L, Le L; Labor

L,
: esL
Figure 6.6 shows the market demand and supply curves for two low-wage sec-
tors, a “covered” sectorin graph (a) and an “uncovered” sectoringraph (b). To
begin, assume that there is no minimum wage law and that the workers and jobs
in each sector are identical. Competition in the labor market would insure that
the wage rates in each sector are also equal, shown as W, in both graphs. Now
assume that Congress passes a minimum wage law that mandates all employers in
the covered sector to pay a wage of at least W,. What will be the impact on the
two sectors?
The first and most basic prediction of economic theory is that the imposition
of a minimum wage above the market-determined wage will cause a decline in
employment in the covered sector. In graph (a) this prediction is represented by
the decline in employment from L, (point A) to L, (point B) on the demand
ae he curve De. In the short run this decline in employment results from the scale
effect of labor demand as the higher minimum wage forces up labor costs and
product prices with a resulting decline in sales and employment. In the long run
the negative employment effect is even larger since the demand curve becomes
more elastic due to the substitution effect as firms replace labor with capital.
A second, less certain prediction is that the imposition of a minimum wage
will also cause a rise in unemployment in the covered sector. At the minimum
wage of W,, employment is only L, (point B), but the supply of people wanting
to work has increased to L; (point C). If this excess supply of workers continues
to search actively for jobs, then measured unemployment in the covered sector
would increase by L; — L,. The reason this prediction is less than certain is that
some or all of these unemployed workers may give up their job search in the
covered sector, either by dropping out of the labor force as discouraged workers,
pursuing some nonwork alternative such as school, or remaining in the labor
force but moving to the uncovered sector in hopes of obtaining jobs there.”’ To
the extent that any of these three events happen, measured unemployment in
the covered sector will increase less, or possibly not at all, in reaction to a mini-
mum wage.
A third predicted impact of a minimum wage involves wages and employment
in the uncovered sector. Assuming some of the workers who lose their jobs in
the covered sector decide to seek jobs in the uncovered sector, the supply of
labor at the prevailing wage of W, would increase, represented in graph (b) by a
rightward shift in the supply curve from S, to Sy,. At W, there is now an excess
supply of labor of L; — L, (point E — point D) in the uncovered sector, result-
ing in downward pressure on wages. If wages are downwardly flexible in this mar-
ket, the wage would decline from W, to W;, leading to a new equilibrium at
point F with a lower wage, but a higher level of employment (L,) than origi-
nally. If wages ae nortatt and assuming some of the unemployed workers con-
tinue to search actively for jobs in the uncovered sector, unemployment will reg-
ister a net increase.
“Jacob Mincer, “Unemployment Effects of Minimum Wage Changes,” Journal of Political Econ-
omy 84, no. 4 (August 1976): S87—S104.
The winners from the minimum wage law are the L, workers in the covered
sector who kept their jobs and are working at higher wages than before. There
are several groups of losers. The L, — L, workers in the covered sector who are
laid off by employers because of the increase in the wage lose most. Two other
groups of losers are consumers who have to pay higher f tices for the goods and
services produced by firms in the covered sector, and the business firms them-
selves who, at least in the short run, will have lower_profits. If wages should fall
in the uncovered sector, a fourth group of losers would be those workers who had
originally been working for the wage W,.
Has the minimum wage law achieved its objective of raising the income of low
wage workers? The answer to this depends on the elasticity of the demand curve
in the covered sector. If labor demand is inelastic (Ep < 1) the answer is yes; if it
is elastic the answer is no. By definition, when demand is inelastic the percent-
age decline in employment is smaller than the percentage increase in the wage,
causing total wage income paid by firms to low-wage workers in the covered sec-
tor to increase. This increase in income is not equally shared, however, since a
minority of the workers in the covered sector will lose their jobs altogether,
while the majority will keep their jobs and earn more per hour. Conversely, if
wages fall in the uncovered sector and demand is inelastic, total income re-
ceived by workers in that sector would fall.
Research Findings Economists have conducted considerable empirical research in an attempt to de-
termine the benefits and costs of the minimum wage. Several conclusions from
these studies stand out.
First, as the theory predicts, a rise in the minimum_wage (or the coverage
ratio) does lead to a net decline in employment. This negative employment
effect is greatest for teenagers and least for adult men. For teenagers, for ex-
ample, a 10-percent rise in the minimum wage has been found to reduce their
employment by about | to 3 percent.”
Second, the decline in employment understates the total decrease in labor
demand by firms, since the response of employers to a higher minimum wage is
often to reduce the hours of work of employees. Evidence suggests that a higher
minimum wage leads firms to turn full-time jobs for teenagers and adult males
into part-time_jabs to save on labor cost.” For women, a rise in the minimum
wage leads to a decrease in part-time jobs and an increase in full-time work. One
explanation is that at a higher wage firms substitute full-time adult female work-
ers for part-time teen-aged workers.
Third, firms also respond to a higher minimum wage with a combination of
higher_prices, shorter business hours, and smaller_profit margins. There is also
28Charles Brown, Curtis Gilroy, and Andrew Kohen, “Time-Series Evidence on the Effect of the
Minimum Wage on Youth Employment and Unemployment,” Journal of Human Resources 18, no. 1
(Winter 1983): 3-31.
Edward Gramlich, “Impact of Minimum Wages on Other Wages, Employment, and Family In-
comes,” Brookings Papers on Economic Activity 2 (1976): 409-462.
Lae, Chapter 6 The Determination of Wages
some evidence that a higher minimum wage leads to increased management effi-
ciency due to attempts to minimize the adverse impact of the wage hike on the
firm, a reaction commonly referred to as a shock effect.*° The impact of a higher
minimum wage is also offset by the fact that a small minority of employers (ei-
ther knowingly or unknowingly) will continue to pay wages lower than what is
legally required.
Fourth, of the workers who lose their jobs in the covered sector, only one-
third or fewer remain unemployed; the majority withdraw from the labor force.™
The major or response for teenagers is to stay in school longer.
Fifth, a rise in the minimum wage increases the total income received by adult
males and females despite the negative employment effect. For teenagers, the
elasticity of demand is sufficiently high that as a group the minimum wage may
reduce their total income. The chief beneficiaries of the minimum wage are
ult females, and the chief losers are teenagers, particularly nonwhite teen-
agers. Adult males, because of the small number working at minimum wages, are
not much affected.”
Sixth, and finally, an additional cost to teenagers from the minimum wage is a
decrease in job experience and on-the-job training. According to a number of
economists, the most important consideration for a teenager is not the wage it-
self, but rather a chance to get a first job, acquire some experience and training,
and then use this as a springboard to upward mobility in the job market.* Be-
cause it raises labor costs to employers, critics of the minimum wage contend
that it causes a reduction in job opportunities and training for teenagers and thus
has a detrimental impact on their transition from school to work.
The evidence on the minimum wage suggests that the benefits and costs of the
law vary considerably for different labor force groups—adult women gain the
most and teenagers are hurt the most. In an attempt to lessen the adverse impact
of the minimum wage, the Reagan administration at one point proposed a spe-
cial subminimum wage of $2.50 per hour for teenagers. Proponents argued that
this law would lead to a large increase in job opportunities for teenagers. Critics
argued that the subminimum wage would lead to more jobs for teenagers, but
only at the expense of fewer jobs for other groups. The main beneficiary, in this
view, would be business firms who could pay lower wages for the same work and
earn more profits. One empirical study that investigated this issue found support
for both points of view. It was estimated that a subminimum wage would create
as many as 430,000 new jobs for teenagers, but would reduce jobs for adults by as
many as 107,000.*
**Rise in Minimum Wage Spurs Some Firms to Cut Work Hours and Hiring of Youths,” The Wall
Street Journal (August 15, 1978); E. G. West and Michael McKee, “Monopsony and ‘Shock’ Argu-
ments for Minimum Wages,” Southern Economic Journal 46, no. 3 (January 1980); 883-891.
*'Mincer, “Unemployment Effects of Minimum Wage Changes.”
*Gramlich, “Impact of Minimum Wages on Other Wages, Employment, and Family Income.”
*Masonori Hashimoto, “Minimum Wage Effects on Training on the Job,” American Economic
Review 72, no. 5 (September 1982): 1070-1087.
“Daniel Hamermesh, “Minimum Wages and the Demand for Labor,” Economic Inquiry 20, no. 3
(July 1982): 365-380.
MONOPSONY The discussion up to this point has focused on wage determination in competi-
tive markets. There are a number of other types of market structures, however.
This section examines the polar opposite to perfect competition—monopsony,
meaning a labor market with only one buyer of labor.*®
There are many_firms in a perfectly competitive market, giving workers a wide
range of alternative places of employment. Because workers have choice, each
firm has to pay the going market wage lest its employees quit and go elsewhere.
A different situation arises in a monopsony market, however. Rather than
having many firms to choose from, a monopsony has only one firm in the mar-
one firm
ket. The classic example of a monopsony is a one-company town where the only
source of jobs in the local area is a textile mill or acoal mine. A situation ap-
proaching monopsony can also occur in other contexts. One is when numerous
employers compete in the local labor market, but only one employer demands a
specific set of skills. An example is a city fire department. Once trained, the
firefighters may find few, if any, other employers in the area that could use their
skills. Finally, monopsony-like conditions may arise because of factors such as
seniority and pension rights, company loyalty, marriage and family obligations,
or a fear of the unknown, that prevent labor mobility by tying workers to a par-
ticular firm or community. *°
Just as a monopolist in the product market is able to charge a higher price
than it could if it were a perfectly competitive firm,
lower wage than if it were a perfect competitor in the Pres ercord
of wage determination for a monopsonist is shown in Figure 6.7. The monop-
sonist, like the competitive firm, has a downward sloping marginal revenue
product schedule, shown as MRP,. What is different between the monopsonist
and the competitive firm is the labor supply curve. The supply curve for a com-
petitive firm is perfectly elastic (horizontal), meaning that it can hire as many
workers as it wants at the going market wage. A restaurant or a retail store in a
large city are good examples. The supply curve facing the monopsonist is far
different, however. Since the textile mill in the one-company town is the only
firm in the market, it has the entire labor supply exclusively to itself. The supply
curve facing the monopsonist is thus the upward sloping market supply curve,
shown as S, in Figure 6.7. While the competitive firm can hire all the labor it
wants at the going wage, the monopsonist cannot; to hire additional workers,
the supply curve S, shows that the monopsonist must paya‘successively higher
wage rate to attract additional workers. Thus, as the textile mill tries to expand
employment, it must pay a higher wage to attract housewives into the labor
force, for example, or to induce persons in outlying rural areas to accept employ-
ment at the plant. Conversely, if the textile mill lowers the wage it will pay, it
35 A labor market with only a few buyers of labor is called an oligopsony; a market with many
employers, but where jobs in one firm are an imperfect substitute for jobs in another firm is called
monopsonistic competition. In both cases the labor supply curve to the firm will be upward sloping,
providing the firm some market power over wages much as in a monopsony market. For a study of
oligopsony in labor markets, see Charles R. Link and John H. Landon, “Monopsony and Union
Power in the Market for Nurses,” Southern Economic Journal 41 (April 1975): 649-659.
36See Martin Bronfenbrenner, “Potential Monopsony in Labor Markets,” Industrial and Labor Re-
lations Review 9 (July 1956): 577-588.
FIGURE 6.7 Wage Determination in a Monopsony Market

The profit maximizing level
of employment for the
monopsonist is L,. This is
determined by where the
firm's MAP schedule
intersects its MCL
schedule (point B). Given MCL,
that the firm desires to hire
L,, the supply curve S,
shows that the firm must
pay a wage of W, (point A)
to attract that many
workers. If the labor W2
market were competitive,
the wage rate and level of
W,
employment would be W,,
L, (point C). The amount of
monopsonistic exploitation
is the difference between
the value of what labor
produces (point B) and
what it is paid (point A).
Lat Ale Labor (L)
will not lose all of its work force since they have few, if any, other employment
Opportunities.
Given the marginal revenue product curve and the supply curve for the
pay? As shown in Chapter 4, the decision rule for calculating the optimal level
Cw JIT-
of employment is to hire additional workers as long as the marginal increase in
WARING
revenue to the firm is greater than the marginal increase in labor cost. Thus, the
perfectly competitive firm in Figure 6.1 hired L, workers (point F) since at that
point the extra revenue from one additional worker as measured by the marginal
revenue product is just equal to the extra cost as measuredby the wage_rate per
hour. This decision rule holds true for the monopsonist, but the resulting level
of wages and employment turns out to be considerably different than for a per-
fectly competitive firm. The reason has to do with the marginal cost of labor.
The Marginal Cost of Labor Schedule To determine the monopsonist’s op-
timal level of wages and employment, it is necessary to derive in Figure 6.7 a
marginal cost of labor (MCL) schedule, shown as the line MCL,. It shows the
change in the firm’s total_dollar outlay for labor per hour for each additional
worker hired. For the competitive firm, the marginal cost of labor schedule is
horizontal and coincides with its supply curve, reflecting the fact that it can hire
as many hours of labor as it wants at the going market wage. For the monop-
sonist, the marginal cost_of labor schedule lies above the supply curve and is
more steeply sloped. “Co;
Why does the marginal cost of labor schedule for a monopsonist have this
shape? To explain this, assume the wage is $5.00 per hour and the firm hires 10
workers. Total labor cost is $50. If the firm has to pay $5.25 an hour to attract an
additional worker, what will be the marginal increase in labor cost? Total cost of
10 workers was $50; the total cost of 11 workers is (11 X $5.25) = $57.75. The
% increase in cost is $7. 75, higher than the actual wage of $5.25. Why? Hiring the
last worker cost $5.25, which meant paying the previous 10 workers $5.25, not
$5.00, causing the total increase in labor cost to be ($5.25 + $2.50) = $7.75.”
Thus, with an upward sloping supply curve such as S,, to hire worker L, requires
a wage of W, (point C), but the actual increase in labor cost is MCL, (point E).
The Equilibrium Level of Wages and Employment How many workers should
the monopsonist hire? The answer is L,. At point B in Figure 6.7 the extra reve-
nue brought in from hiring an additional worker (the MRP, ) is just equal to the
marginal increase in labor cost (the MCL). To expand employment to L, would
result in lower profits since the extra cost of hiring that last worker (point
E) is
far greater than the revenue brought in (point _C). Once the level of employ-
ment is determined, it is an easy matter to determine the wage the monopsonist
will pay for labor. To hire L, workers, the labor supply curve shows that a wage
of WZ(pointA) has to be paid. ae
To appreciate the implications of this outcome, it is useful to compare the
wage and employment level under monopsony with what would result if the
market were competitive. Under perfect competition, the wage rate and level of
’ employment in the market would be W, and L,, determined by the intersection
of the marginal revenue product schedule MRP, and the supply curve S, (point
C). Under monopsony, the wage and employment levels are only W, and L,.
The result is that both wages and employment are lower in a monopsonistic mar-
ket than in a competitive market. iworkers had alternative sources of employ-
ment, the monopsonist would not be able to take advantage of its employees by
paying them less than competitive rates. Since they do not, this gives the firm
power to practice monopsonistic exploitation of labor. The amount of exploita-
tion: is the difference between what labor is worth to the firm (shown by the
MRP, at point B) and the wage that labor receives (W,). Note that in perfect
competition there is no monopsonistic exploitation since labor is hired up to the
point where W = MRB.
Empirical Evidence
Monopsony in Baseball
The most important prediction of the monopsony model is that the wage rate
paid to workers will be lower than if the labor market were competitive. One
example that offers striking-support for this prediction is the dramatic increase in
37]t is assumed the firm must pay all workers a uniform wage. If it was able to limit the wage
increase to only the last worker hired (a form of price discrimination), the increase in marginal labor
. cost would be much lower. One way firms attempt to accomplish this is by limiting the amount of
information that workers have about each other's rates of pay, a second (as illustrated in the previous
In the News section) is to pay one-time hiring bonuses or other lump-sum payments such as those for
moving or housing expenses.
player salaries in baseball following the introduction of the free agent system in
197i
Prior to 1977, a baseball player was a “free agent” (could negotiate with any
club) until he signed his first contract with a major league team or its minor
league affiliate. Once signed, all baseball contracts contained a renewal or “re-
serve” clause stipulating that thereafter the player's services were the sole prop-
erty
of that club for his entire career in baseball unless the club decided to sell or
trade his contract. The reserve clause effectively turned the market for individ-
ual players into a monopsony by preventing them from negotiating with any but
their own club
On December 23, 1975, the reserve clause was struck down by an arbitrator
ruling on a grievance brought by the Major League Baseball Players Association
(MLBPA). Under a new agreement negotiated by the MLBPA and the owners,
players were no longer permanently tied to a team; players who had signed con-
tracts prior to 1976 were able to become free agents in 1977; players signing con-
tracts after 1976 could become free agents after 6 years. The effect of the new
contract system was to turn what had been a monopsony into a competitive
market.
What effect did this development have on players’ salaries? According to one
study, average player compensation rose from $54,330 in 1976 to $77,292 in
1977—a 42-percent increase in 1 year! Opening the market to competitive bid-
ding led to a substantial rise in salaries, as the theory would predict.
A second prediction of the monopsony model is that workers are paid less
than their marginal revenue product; in a competitive market, by contrast,
workers are paid wages equal to their marginal contribution. This same study
also found clear evidence of this prediction; under the free agent system salaries
eo rose an estimated 37 percent for “below average” hitters but 69 percent for the
“star” hitters—the ones whose marginal contributions to the teams (and owners’
revenues) were the greatest.
UNIONS AND The impact of unions on the labor market is considered in much more detail in
WAGES Chapters 10, 11, and 12. This section focuses solely on the methods used by
unions to gain higher wages for their members. The analysis examines first the
wage impact of Unions in a competitive labor market and second the wage im-
pact of unions in a monopsonistic market.
Unionism The three different ways in which unions influence the level of wages in the
in a Perfectly labor market are illustrated in Figure 6.8, graphs (a), (b), and (c).
Competitive Market
sid attele ead The Strike Threat The first and most important way in which a union is able
to raise the wage for its members is through the threat of a strike. A strike gives
workers additional bargaining power vis a vis the firm that none of them as indi-
“The discussion in this section is drawn from James R. Hill and William Spellman, “Professional
Baseball: The Reserve Clause and Salary Structure,” Industrial Relations 22, no. 1 (Winter 1983):
1-19.
FIGURE 6.8 Three Ways Unions Gain Higher Wages

A union has three ways to raise wages. The first—graph (a)—is through the threat of strike. This induces the firm to increase
wages from W, to W,, resulting in a decline of employment of L, — Ly and an excess suppiy of job seekers of L; — L,. The
second way—graph (b)—is by restricting the supply of labor, which shifts the supply curve from S, to S», driving up the wage to
W,. The third way—graph (c)—is to increase the demand for union labor. This shifts the demand curve from D, to D>, driving up
the wage to W,.
(a) Strike Threat (b) Supply Restriction (c) Demand Increase ie AK
Wage
(Ww)
vidual workers have. The originof bargaining power is the ability to impose a
cost on someone if that person does not follow a prescribed course of action. Ina
nonunion setting, the individual worker’s bargaining power is derived from his or
her ability to quit and seek work at some other firm; in an attempt to attract and
keep a work force, the firm is motivated to pay at least the going market wage.
By the same token, labor mobility gives the individual eter baieaining
power to force the firm to pay a wage higher than the going market rate since the
firm could easily replace the person and recruit someone else from its perfectly
elastic supply of labor.
~ What each worker as an individual cannot do, workers as a group can. The
essence of a union is that individual bargaining is replaced by collective bargain-
ing where the workers negotiate as a single group through their elected union
representatives. The principal weapon which a union has to induce the em-
ployer to pay a higher wage is the threat of a strike. A strike, by shutting down
production at the firm, imposes a cost on the employer that is far larger than the
threat of an individual worker to quit. To avoid this cost, the firm is motivated
to compromise with the union by offering a wage higher than the going market
wage:
This sequence of events is illustrated in Figure 6.8, graph (a). Assuming the
labor market is perfectly competitive and initially nonunion, the market forces
of demand (D,), and supply (S,) determine the equilibrium wage of W, (point
A). If the workers in this labor market form a union, the threat of a strike and its
attendant costs will induce firms to agree to a higher wage of, say, W,. As the
For example, in the 1984 GM—UAW negotiations it was estimated that the company would
lose $200 million of earnings per week from a complete shutdown of its facilities.
wage is forced up to W,, employment falls in the market from L, to L, (point A

to B) as the least senior union members are laid off. Even as employment de-
clines, the higher wage of W, leads to an increase in the number of people seek-
ing jobs of L; — L, represented in graph (a) by the movement up the supply
curve S, from point A to point C. Firms in this market face an excess supply of
labor of L, — L, willing to work at the higher union wage.
Restriction of Supply A union has two other means to raise wages. One is
the restriction of the supply of labor. Many cgaft_unions restrict admittance into
apprenticeship and training programs operated by the union. Particularly in con-
struction, a number of highly skilled occupations such as electrician, plumber,
and iron worker require extensive training and on-the-job experience. To the
extent that the upian-operated apprenticeship program is the major source of
supply of such workers, it provides the union a channel to raise wages by ar-
tificially restrictingthe supply of labor to the market.* A second method used by
unions to restrict the supply of labor is support for various types of legislation
such as restrictions on immigration, compulsory retirement laws, or laws short-
ening the hours of work.
The impact on wages of restrictions of the supply of labor is illustrated in Fig-
ure 6.8, graph (b). If the labor market is nonunion, the equilibrium wage is at
W, (point A). Restriction of enrollees in apprenticeship programs or passage of
restrictive labor legislation results in a shift to the left of the market supply curve
from S, to S,. At the initial wage W, a shortage of labor now exists, resulting in
wages being bid up to W, (point B) and employment declining from L, to L,. At
the higher wage W,, employers in graph (a) face a long line L,; — L, ofjob ap-
plicants; in graph (b), however, the market is in equilibrium at W, with demand. fe
just
Saal to supply. In this case the long line of applicants is in the form of
people wanting to enroll in the apprenticeship training program or seeking to
immigrate into the country.
Increased Labor Demand The third method used by unions to raise wages
involves increasing the demand for union labor. One method of doing so is
to induce consumers or other union members to buy only union-made products.
One example is the well-known advertising campaign of the International
Ladies Garment Workers Union urging consumers to “look for the unjon label”
in buying apparel products; a second example is the practice of unions to pur-
chase products or use the services only of firms that employ union members.*!
*Union control of training in construction has weakened considerably with the growth of the
nonunion or “open shop” sector. Increasingly, therefore, restrictive apprenticeship programs serve
not to raise wages, but to ration scarce job opportunities. For a formal analysis of union restrictions
on training opportunities, see Yoram Weiss, “The Effect of Labor Unions on Investment in Training:
A Dynamic Model,” Journal of Political Economy 93 (October 1985): 994-1007.
"The union label originated in San Francisco in 1875 with the Cigar Maker's International
Union. The union members were white males whose weekly wage of $12 was threatened by Chinese
cigar makers, who earned only $6. To promote the demand for their product, the unionists affixed a
label to each cigar box identifying it as union-made by white men. This is described in Ernest R.
Spedden, The Trade Union Label (Baltimore: Johns Hopkins Press, 1910): 9-10.
An alternative method of increasing the demand for union labor is through re-
strictive work rules or “featherbedding.” Union contracts frequently stipulate
how many workers must be used for a particular task (for example, three workers
to guide a passenger plane to the terminal ramp), job classifications that prevent
all but a certain individual or group of workers from performing a particular task,
or restrictions on subcontracting work to nonunion firms.
The impact of these union policies on wages in the labor market is illustrated
in Figure 6.8, graph (c). In a nonunion market the wage is again W, (point A).
The effect of a “buy union” campaign or restrictive work rules is to shift the de-
mand curve for labor to the right from D, to D,, causing an excess demand for
labor at W, and a bidding up of wages in the market to W, (point B). At the
higher wage of W,, employment in the market increases to L).
The Monopoly View of Unions Given the assumption that the labor market
is competitive, the foregoing analysis yields a generally negative view of the ac-
tivities of unions.** The purpose of the union is to improve the economic posi-
tion of its membership, which is accomplished by raising wages in one or a com-
bination of the three ways outlined above. The winners from this process are the
union members who remain_employed at higher wages. The losers include sev-
eral groups. One group of losers are the union members who lose their jobs as
employers cut back on employment in response to higher wages. A second group
of losers are consumers who buy the higher-priced union-made products. A third
group of losers are firms and their owners who suffer reductions-in profits. A final
loser is society in general because the distortion of relative wage rates caused by
the union will lead to an inefficient use of resources.
In many ways the impact of a union in a competitive labor market is analo-
gous to the impact of a monopolistic. firm in the product market. The union
replaces a system of individual bargaining with a system of collective bargaining
where it is, in effect, the single seller of labor. Its control over the supply of labor
provides the union with market power with which it can raise wages for its mem-
bers at the expense of higher prices and lower output than if the economy were
perfectly competitive. From this perspective, unions are seen as another, albeit
important, market imperfection that causes the labor market to work less_effi
ciently than it might.
Unionism ina A second view of unions yields a markedly different set of conclusions. This is
Monopsony Market best seen if the labor market is assumed to be monopsonistic.
The monopsony model is reproduced in Figure 6.9. The labor market is com-
posed of only one firm; its marginal revenue product curve is MRP, ,the supply
curve of labor is S,, and the resulting marginal cost of labor schedule is MCL).
The equilibrium level of employment is L,, and the equilibrium level of wages is
W, (point A).
See, for example, Morgan O. Reynolds, Power and Privilege: Labor Unions in America (New
York: Universe Books, 1984). The monopoly view of unions is also discussed in Richard Freeman
and James Medoff, What Do Unions Do? (New York: Basic Books, 1984).
FIGURE 6.9 The Effect of Unionization in a Monopsony Market

In the absence of a union, the monopsonist would pay a wage of W, and hire only L, workers (point A). By unionizing, the workers
could achieve a higher wage such as W, and employment would also increase to L, (point C). The reason is that by establishing
the wage of W,, the union changes the firm's marginal cost of labor schedule from MCL, to W,CDE. Up to L» workers, the firm can
hire each additional worker at a constant wage of W,; past /,, however, it must pay a successively higher wage as given by the
supply curve. Should the union raise the wage beyond W3, employment would decline below the original level of L;.
Wage
_ (W)
MRP,
0 | bi _sbo ce (Labor dh)
In a monopsony, workers are exploited in the sense that the wages they re-.
ceive are less than the value
of their contributions to production (their MRP, )
and less than they would receive if the labor market were competitive. The rea-
son the firm can pay lower wages is because it has market power derived from the
Jack of alternative sources_of employment in the local labor market. Are any»
avenues open to workers to improve their wages? One option is to move to an-
other city or state where the labor market is competitive and wages are higher. A
second option is to join together and form a union. By forming a union, the
workers gain through the threat of a strike a countervailing force with which to
confront the employer's market power. What will be the impact of the union on
wages and employment in the monopsonistic market?
The introduction of a union in a monopsonistic market can result in both_
higherwages and greater employment, as shown in Figure 6.9. Assume that the
threat of a strike induces the
employer to raise the wage from W, to W,. At the
wage W, employment had been L,, determined by the intersection of the curves :
MRP, and MCL, (point B). At the higher, union-imposed wage of W,, the
monopsonist will actually increase employment from L, to L,. The reason em-
ployment increases is that up to an employment level of L, the firm is able to
hire additional workers at a constant rate_of W,; after employment exceeds L,,
the firm reaches the supply curve S,, forcing it to raise wages further to hire
additional workers. Since each worker up to L, can be hired at the same wage,
the monopsonist’s marginal cost of labor schedule is no longer the upward slop-
ing line OE, but rather the kinked line W,CDE containing the horizontal por-
tion from Hatopoint Given this, the monopsonist’s optimal employment
level is determined where the marginal revenue product schedule MRP, inter-
sects the marginal cost of labor schedule W,CDE. This occurs at point
C. where
the level of employment is L,. As a Cea cee workers in the monop-
sonistic market now enjoy both higher wages and greater employment oppor-
tunities. (The same result would occur if the wage W, were established by a
minimum wage law instead. )
Three points about this conclusion should be noted. First, the ability of the
union to’raise wages and employment in a monopsonistic market is not un-
limited. A rise in the wage above W; in Figure 6.9 would cause employment to
decline below the initial monopsonistic level of L,. Second, a union wage in-
crease below W; may still not result in the predicted increase in employment if
the higher labor costs force the firm out of business. Even though a monop-
sonistic firm pays low wages, it may still be operating at a loss. A wage increase,
even if only to competitive levels, might force the firm to shut down. A classic
example is the unionization of a marginally profitable textile firm in a rural
‘southern town. Finally, if the union happened to pick the wage W,, as in the
example in Figure 6.9, the resulting wage and employment level would
be exactly the same combination that would prevail in the case of perfect
competition.
The picture of unionism that emerges from this analysis is quite different than
in the case of perfect competition. Rather than introducing an element of inefh-
ciency into what had been a perfect labor market, unionism in a monopsonistic
market can lead to a net gain in social welfare.** The clear loser from unionism is
the firm that is forced to pay higher wages, with a consequent reduction in prof-
its. One group of winners from unionism are the workers who gain both higher
wages and increased employment opportunities. Consumers also gain from the
unionization of the monopsonistic employer since a higher wage such as W,
induces the firm to increase. employment and production, leading to lower prices
in the product market. Finally, the countervailing power of the union in a
monopsonistic market leads to a net gain in overall economic efficiency by mov-
ing the employer to hire the level of employment that would have prevailed if
the market had been perfectly competitive to begin with.
Two Views of Unionism“ Unionism through the force of collective bargain-
ing provides workers a means to increase wages that is unavailable to each of
For this point of view, see Lloyd Reynolds and Cynthia Taft, The Evolution of Wage Structure
(New Haven, Conn.: Yale University Press, 1953): Chapter 13; and James E. Meade, Stagflation
Volume 1—Wage Fixing (London: Allen and Unwin, 1982): Chapter 4.
“This title is usually associated with the work of Richard Freeman and James Medoff “The Two
Faces of Unionism,” The Public Interest 57 (Fall 1979): 69—93. Freeman and Medoff distinguish be-
tween the monopoly view of unionism and the “collective voice” view. By giving workers a voice in
the operation of the firm, Freeman and Medoff argue that unionism even in a competitive market
may increase efficiency through higher productivity, lower quits, and so on. This rationale for union-
ism is completely different from the monopsony argument given here. The collective voice aspect of
unionism is discussed in Chapters 10 and 12.
Chapter 6 The Determination of Wages
them individually. The issue of importance for economics is whether this in-
crease in wages is, on net, beneficial or harmful with respect to the efficiency of
the labor market and the allocation of resources in the economy. The answer
depends significantly on the competitive nature of the labor market. If the labor
market is highly competitive, union wage gains lead to a distortion in relative
ype factor prices and an inefficient allocation of resources, just as a monopoly does in
the product market. Conversely, if the labor market is monopsonistic (broadly
defined to include oligopsony and other forms of imperfect competition) union-
ism equalizes the bargaining power between a dominant employer and its work
force. In this case, the higher wages that result actually lead to greater employ-
ment opportunities and a more efficient allocation of resources.
Part of the controversy over unionism, therefore, revolves around whether
unions are “monopoly creating” or “monopsony reducing.” Critics of unions
generally take the former view; proponents of unions take the latter view. The
critical issue in differentiating between the two sides is to what extent workers
have mobility in the market. If numerous employers compete in the market and
workers can easily move from one to another, the labor supply curve-to the indi-
vidual firm will be quite elastic, providing little room for monopsonistic exploi-
tation of labor by business firms. Conversely, if workers have few employers to
choose from in the market, or if mobility is limited due to financial constraints,
discrimination, or the employer-specific nature of skills, the supply curve of la-
bor to the firm will be upward sloping, providing the employer a dominant bar-
gaining position vis a vis the worker. Thus, from one point of view unions are an
i
imperfection that interfere with the operation of a competitive market, from an-
iy
other point of view unions are necessary precisely because the market itself is
imperfect.
THE FIRM’S Up to this point the process of wage determination has been examined at the
INTERNAL WAGE market level and the level of the firm. The next topic is the determination of
STRUCTURE wages within the firm, or what is known as the firm’s internal wage structure. The
internal wage structure refers to the set of wage rates paid for particular jobs and
workers in the business firm and the.system of differentials
wage that separate
each job and worker from others. The work force of nearly every firm is com-
prised of people performing varied and distinct jobs, ranging from the waitress,
cook, and cashier in a small restaurant to the hundreds and thousands of distinct
jobs in a large corporation. Part of the personnel function of the management of
every firm is to decide what wage rate each individual in the organization will be
paid. The interesting questions for labor economics in this subject are how the
internal wage structure is established, whether it serves to efficiently allocate la-
bor within the firm, and the extent to which market forces of supply and demand
influence the i structure.
Determination of the internal wage structure in a medium- to large-sized firm
typically involves two steps. The first is to establish wage grades or ranges of pay
for particular jobs in the firm; the second is to determine the r. ~and-
method of compensation for individual workers in each job. Both steps are
briefly described below.
FIGURE 6.10 The Job Structure in a Chemical Plant
: Comptroller Personnel | Community

Research :
Relations
hire hire
- First-Line Supervision hire
i Bath Operations ae | Process Maintenance hire

Senior Clerks | Preparation | Cluster Cluster | Operator Mechanic
Operator ee
hire ‘Accountant Production Maintenance hire
Coating aQuality Trainee
| Technical Inspector
etc.
Tower
_ Operator — Assistant
Trainee — hire
Laboratorian
Operators | Area
. Clerks Coating Operator
hire Trainee hire
Stenos
hire hire
source: Peter B. Doeringer and Michael J. Piore, /nternal Labor Markets and Manpower Analysis (Lexington, Mass.: Heath Lexington Books, 1971): 46.
Determining Job The Firm's Job Structure The set of wage rates paid for particular jobs in a
Rates firm is heavily contingent on its job structure.* The job structure is defined as
————
the set of jobs in terms of skill and function required by the firm. The job struc-
ture in a particular firm is heavily influenced by the technology of production; if
the company is an airline, for example, technology dictates that the firm’s job
structure must include pilots, flight attendants, and so on. The types of jobs in a
firm are not completely determined by technology, however, for as shown in
Chapter 5, one of the major long-run decisions of the firm is what proportions
and kinds of labor, capital, and other factor inputs to use. Once these are de-
cided and a specific plant is built, then the technology of production generally
determines within fairly narrow bounds the job structure for which the firm will
need to recruit labor.
A particular job structure is illustrated in Figure 6.10 for a chemical plant.
This job structure shows one of the most fundamental features of modern tech-
nology—specialization and the division of laber- As Adam Smith pointed out
200 years ago, production can be made more efficient by breaking the task into
eT Scion in this section is drawn from George H. Hildebrand, “External Influences and
the Determination of the Internal Wage Structure” in J. L. Mey, ed., Internal Wage Structure
(Amsterdam: North-Holland, 1963): 260-299.
separate identifiable parts and having workers specialize in each one. This is re-
flected in the organization of modern firms into functional departments that spe-
cialize in one particular part of the production process. The chemical plant
in Figure 6.10 requires five separate functions or job clusters (jobs related to a
common activity): maintenance, quality control, operations, clerical, and
management.
One additional feature of a firm’s job structure has an important influence on
the internal wage structure, the concept of a job ladder. Each job cluster con-
tains a group of tasks that are related to a common function or activity in the
production process. For many types of production, the common set of jobs in a
job cluster also form a definite hierarchy of tasks extending in a vertical direction
from the least skilled job up to the most skilled job. This vertical hierarchy of
jobs is known as the job ladder and is illustrated in Figure 6.10 by the upward
line of progression in jobs for both the “operations” and the “quality control”
cluster.
What gives rise to these job ladders? Peter Doeringer and Michael Piore have
identified three factors: the specificity of skills for the jobs, the importance of
on-the-job training, and work place custom.* In many firms, some jobs require
specific skills or knowledge that are at least to some degree, only of use to that
one firm and that can be learned only by actual training on the job. An example
is a locomotive engineer who starts out as a yardman and works up to brakeman,
conductor, and finally engineer for a particular railroad. An engineer with the
knowledge of the specific route and operating procedures of that railroad cannot
be bought by the firm in the external labor market—the knowledge and skills
have to be learned and the progression up the job ladder is the manner in which
the training is accomplished. According to Doeringer and Piore, these economic
reasons for job ladders are also augmented by the force of custom at the work
place as informal or unwritten practices regarding promotion to higher jobs be-
come solidified over time into a recognized job ladder.
The Internal Labor Market The effect of a job ladder is to create an internal
labor market within the firm. The internal labor market refers to those jobs in
the firm for which competition to fill a vacancy is limited to workers already
employed with the firm. The beginning of the internal labor market is the port
of entry, which connects the internal job ladder to the labor market outside the '
firm. In Figure 6.10, the ports of entry are denoted by the word “hire” at the
bottom of the operations and quality control clusters. If someone at the top of
the job ladder retires, workers at each step of the ladder move up one rung,
creating a vacancy at the bottom or port of entry. To fill this vacancy, the firm
goes to the external market, where it competes with other firms for workers at
the entry level. The decision of whom to hire to fill the entry level positions is
quite important for both the firm and the job applicants in the external market.
It is important for the firm because the person chosen is likely to be with the firm
“Peter B. Doeringer and Michael J. Piore, Internal Labor Markets and Manpower Analysis (Lex-
ington, Mass.: Heath Lexington Books, 1971).
a considerable time, and the firm will invest much money in that person for
additional training. For these reasons, firms often conduct elaborate screening
processes to select workers for entry level positions in the internal labor market
(for example, management trainee). They are certainly much more careful with
these positions than with secretarial or custodial positions that require less in-
vestment in firm-specific training. The selection process for these entry level
positions is also quite important for the persons in the queue of job applicants
because the chosen person gains a job with the potential of considerable earn-
ings growth, promotion possibilities, and job security. Not surprisingly, these at-
tractive features often result in cut-throat competition as several hundred job
candidates compete for a handful of job openings at the ports of entry of large
firms.
Given the structure of jobs and promotion ladders in the firm, how does the
firm determine the appropriate set of wage rates for workers in each job cluster
and among job clusters? One solution would be to pay workers in each job classi-
fication the going wage for that occupation and skill level as determined by sup-
ply and demand in the external labor market. This solution was pictured in Fig-
ure 6.1, where demand and supply in the market determined the equilibrium
wage for each type of labor and the firm acted as a passive wage taker.
For some or all of the job clusters within a firm, the internal wage structure is,
in fact, largely determined by competitive pressures in the external labor mar-
ket. An example is the clerical-job cluster in the chemical plant shown in Figure
6.10. For clerical jobs there is no job ladder to speak of and no internal labor
market. The skills required of clerical workers are generally standardized and
transferable from one employer to another. In this case, the area of indeter-
minacy in wages will be relatively small and the chemical plant, if it is not to
lose all of its workers, must pay a rate fairly close to the going wage in the
market.
For other job clusters in the firm for which job ladders and on-the-job training
are important, the area of indeterminacy in wages may be much larger.*’ This is
because jobs in the internal labor market are not all standardized as are the cleri-
cal jobs; rather, each one is idiosyncratic in that it involves a specific set of skills
unique to the firm that can be learned only by a process of on-the-job training as
the worker moves up the job ladder. The upper limit to wages W, for each job in
the internal labor market is determined by the marginal revenue product of the
worker; at a wage higher than the MRP,, the firm would find it unprofitable to
fill the job. The lower limit to wages W, that the firm must pay for each job in
the internal labor market is determined by the best wage offer the worker could
get if he or she quit and went elsewhere. Because of the firm-specific nature of
job skills and training, the value of the worker to another firm, and thus the
wage received, is likely to be considerably lower than the wage paid by the
present firm. In the chemical plant in Figure 6.10, for example, if a worker in
the operations job cluster were to quit, he or she would lose the position on that
47See O. E. Williamson, M. L. Wachter, and J. E. Harris, “Understanding the Employment Rela-

tion: The Analysis of Idiosyncratic Exchange,” Bell Journal of Economics (Spring 1975): 250-278.
job ladder and would have to begin again at a lower position (possibly the port of
entry) at another chemical firm. The consequence is that even if the firm does
not pay the worker in the internal labor market a wage equal to the value of his
or her productivity, the worker would still accept a range of lesser wage rates
before going to some other firm. For clerical workers, on the other hand, the
lack of job ladders and firm-specific job skills makes the worker equally valuable
to every firm in the market and thus this type of labor will trade at or relatively
close to the going wage.
Job Evaluation Because the external labor market cannot determine a
unique wage for jobs in the internal labor market, firms have to use some non-
market device to establish an explicit wage structure. One technique that is fre-
quently used is job evaluation. The primary objective of a job evaluation system
is to achieve an internal wage structure that promotes equity and rewards eff-
ciency while at the same time being consistent with competitive conditions in
the external labor market. Most job evaluation systems have the same basic
methodology.** The first step is to write a detailed description of each job. The
second step is to rate or evaluate each job according to one or more “compen-
sable factors” such as the skill and education required, degree of physical dis-
comfort or risk, amount of supervisory responsibility, and so on. The third step is
to add these ratings together in some way to create a total “job worth score.”
Based on the job worth score, each job in a job cluster can then be ranked from
highest to lowest and assigned a wage rate or pay range. In actual practice, most
job ladders contain certain “key” jobs that are closely linked to the external la-
bor market and for which market rates of pay are readily available (such as from
local area wage surveys). In devising an internal wage structure, the firm will
take these wage rates as reference points and then use job evaluation scores to
establish pay rates for the positions between the key jobs.
Determining The structure of job rates provides a rough outline for the internal wage structure
Individual Rates of the firm, but it seldom completely determines it. The reason is that the firm
of Pay must decide on the specific rate of pay for individual workers in each job cate-
gory and the method of compensation.
Two basic types of compensation methods are used to determine individual
pay rates. The first bases pay on the amount of worker input supplied to the job,
the second bases pay according to some measure of output produced by an indi-
vidual worker or a group of workers.” The first type of arrangement commonly
takes the form of time payments such as a wage per hour or salary per year, while
the second encompasses pay plans related to individual output such as piece rates
and sales commissions and group output such as profit-sharing plans.
*See Richard I. Henderson, Compensation Management, 4th ed. (Reston, Va.: Reston Publishing
Co., 1985).
*See John Pencavel, “Work Effort, On-the-Job Screening, and Alternative Methods of Re-
muneration,” in Ronald Ehrenberg, ed., Research in Labor Economics, vol. 1 (Greenwich, Conn.:
JAI Press, 1977): 225-258.
A variety of factors influence the firm’s particular kind of compensation plan

or mix of plans. Generally, payment plans based on individual output provide
the strongest incentive for individual effort and most closely tie pay to perfor-
mance. While these features would seem to be attractive to both workers and
firms, the use of traditional piecework pay plans has decreased significantly over
time until today they cover no more than 15 to 20 percent of the work force.
Several reasons can be cited. One is that a piecework system is only practical for
jobs where the amount of output is both directly measurable and subject to
worker control. These types of jobs are becoming increasingly rare, however, as
a greater number of workers produce hard-to-measure services, such as teaching
or management, or work at jobs requiring considerable teamwork, such as on
assembly lines. A second consideration is that piecework plans often fail to de-
liver the anticipated increases in production because workers collectively restrict
output fearing that increased production will only cause management to raise
the output quota further or lower the compensation per unit produced.” A third
problem is that a piecework payment plan encourages workers to emphasize
quantity produced at the expense of product quality.
As piece rates have declined in significance, time rates have grown more
popular. Traditionally, blue-collar and lower-level service workers have been
paid on an hourly basis, while white-collar workers have more often received
monthly or annual salaries.*! This difference in compensation methods reflects,
in part, the tendency of companies to view production-type workers as a variable
cost and higher-level management and professional workers as a quasi-fixed cost.
Time rates are easy for the firm to administer, but they reward workers for the
time spent on the job rather than for the actual amount of work accomplished.
Lack of work effort on the part of employees is a major potential problem with
time rates, therefore.” Firms have attempted to spur work effort of workers on
time rates through a carrot-and-stick approach. The stick is close supervision of
workers with the threat that a worker can be fired for less-than-adequate effort.
The carrot takes two forms.
One form is the merit pay plan where wage and salary increases are awarded
on the basis of each worker's performance as determined by a supervisor or man-
ager.°? Since workers generally regard performance as an equitable basis for de-
termining pay, a merit plan that results in different wages for different workers in
the same job category does not necessarily violate employee considerations of
0The classic reference on this subject is Stanley Mathewson, Restriction of Output among Unor-
ganized Workers (New York: Viking, 1931).
1 Of the 106 million people employed in 1984, 92 million were paid a time rate and 54 million
were paid on an hourly basis. See Earl F. Mellor and Steven E. Haugen, “Hourly Paid Workers: Who
They Are and What They Earn,” Monthly Labor Review 109 (February 1986): 20—26.
This subject is explored in more detail in Edward Lazear, “Salaries and Piece Rates,” Journal of
Business 59 (July 1986): 405-432.
53,4 detailed discussion of merit pay plans is given in Robert L. Heneman,“Pay for Performance:
Exploring the Merit System, Work in America Institute Studies in Productivity no. 38 (New York:
Pergamon Press, 1984).
equity and fairness. In practice, however, merit pay plans can easily engender
worker dissatisfaction if employees perceive that individual performance is not
being accurately measured or consistently rewarded. One frequent outcome of
such dissatisfaction is that workers vote to join a union. A union, in turn, will
attempt to restrict management's ability to award pay on the basis of individual
performance and will instead seek to give objective factors such as seniority and
changes in the cost of living greater weight in pay determination.
The second carrot used by management to induce greater work effort by em-
ployees on time rates is to set up some type of group incentive plan. One that has
received considerable attention in recent years is gain-sharing. The central fea-
ture of gain-sharing is that at least a portion of employee pay is directly tied to
some objective measure of firm performance such as profits, productivity growth,
or cost reduction.”
The most prevalent type of gain-sharing program, profit-sharing, has several
attractive features. First, it stimulates employee job performance by giving each
worker a direct stake in the amount of profit earned by the firm. Second, it in-
creases employee pay without raising base rates (i.e., the wage per hour or an-
nual salary). This saves the firm money since cost-of-living adjustments and
merit increases are typically awarded as some percentage of base rates. A third
advantage is that the firm’s labor cost becomes flexible in the downward direc-
tion since the profit-sharing component of employee compensation will typically
fall during a recession.”
Profit-sharing also has disadvantages, however. The firm faces drawbacks in
that profit-sharing is more difficult to administer than straight time rates and
that it must reveal more of its cost and revenue data to employees. An important
drawback of profit-sharing for workers is that their annual incomes are much less
certain.
SEGMENTED The final topic considered in this chapter is segmented labor market theory. One
LABOR MARKETS of the basic assumptions of the perfectly competitive model is that all jobs in the
labor market are open to competition. This does not mean that a waitress can
quit her job and immediately become a doctor or that an unskilled laborer can
effectively compete for an ironworker’s job since each person's sphere of competi-
tion is limited in the short run by his or her education, job skills, geographic .
location, and mental and physical abilities. The presumption of competitive
theory, however, is that the walls separating labor markets are relatively porous
so that workers, given enough time to acquire the training or to move to a differ-
ent city, can effectively compete for jobs on an equal footing with persons al-
ready in that labor market.
*A detailed description of gain-sharing plans is provided in Carla O'Dell, People, Performance,

and Pay (Houston: American Productivity Center, 1987). She found in a survey of 1,600 firms that
75 percent used some form of a bonus or gain-sharing plan; of these, 69 percent had instituted the
plan within the previous 5 years.
* This aspect of profit-sharing is examined in more detail in Chapter 15.
Many economists, while recognizing that there are some barriers to mobility
across individual labor markets, believe that most jobs are, in fact, relatively
wide open to competition. Others disagree, however, believing that the labor
market is segmented or balkanized into noncompeting groups.* The idea of la-
bor market segmentation is that strong barriers to entry prevent certain groups
of workers from effectively competing for jobs in a given market. These barriers
may take several forms. One is the internal labor market, which prevents work-
ers in the external market from competing for job openings above the entry level
in the firm. A second is discrimination, which causes workers of certain race or
sex groups to be excluded from competition for desirable jobs. A third is union
hiring rules that restrict the competition for jobs to those workers who are regis-
tered with the union hiring hall or are next in line on the company seniority list.
A fourth is occupational licensing laws that restrict the entry of new practi-
tioners into an occupation by making the licensing exam difficult, by imposing
lengthy residency requirements, or by requiring large licensing fees. The com-
mon effect of these entry barriers is to give a marked advantage to the persons
who are the “ins” over those who are the “outs” in the competition for jobs.
There are a number of different theories of segmented markets. One that has
received considerable attention in recent years is the dual labor market theory.
Dual Labor Market Dual labor market theory was developed in the late 1960s by institutionally ori-
Theory ented economists such as Peter Doeringer, Michael Piore, Barry Bluestone, and
Bennett Harrison to explain continuing poverty and unemployment among dis-
advantaged workers, particularly blacks and other minorities in the inner city.”
From the perspective of competitive theory, inner city workers have low wages
and poor employment prospects because they lack the human capital (education
and training) necessary for high productivity and high earnings in the labor mar-
ket. The solution to the poverty problem from this perspective is improved edu-
cational systems and job training programs for inner city residents so that these
workers can leave low-wage “bad” jobs and effectively compete in markets offer-
ing higher-wage “good” jobs.
While this approach to the poverty problem seems promising, the dual labor
market theorists concluded from their case studies of inner city labor markets
that it would not succeed. To explain why, they developed the theory of dual
labor markets. Dual labor market theory posits that the labor market is seg-
mented into a primary labor market and secondary labor market. The primary
sector consists of the good jobs in the labor market, meaning those that offer
high wages, good working conditions, employment stability, chances of promo-
*6See Marcia Freedman, Labor Markets: Segments and Shelters (Montclair, N.J.: Allanheld, Os-
mun and Co., 1976); and R. Loveridge and A. L. Mok, Theories of Labor Market Segmentation
(Boston: Martinus Nijhoff, 1979).
*7See, for example, Doeringer and Piore, Internal Labor Markets and Manpower Analysis; Barry F.
Bluestone, “The Tripartite Economy: Labour Markets and the Working Poor,” Poverty and Human
Resources (July/August 1970): 15-36; and Bennett Harrison, “The Theory of the Dual Economy,”
in B. Silverman and M. Yanowitch, eds., The Worker in the Post Industrial World (New York: Free
Press, 1971).
tion, and equity and due process in the administration of work rules. Primary
jobs are generally found in large firms and firms that are unionized. A key feature
of employment in these firms in their well-developed internal labor markets,
owing to the substantial amount of on-the-job-training in firm-specific skills that
is required of the workers. Because of the substantial amount of time and money
that primary firms invest in their workers, these firms screen very carefully the
persons hired into the entry level positions in the internal labor market, hoping
to find persons possessing traits such as reliability, discipline, company loyalty,
an ability to be a member of the company “team,” and an ability to learn quickly
on the job.
The secondary sector contains the bad or undesirable jobs in the labor mar-
ket. These low-paying, “dead-end” jobs offer little in the way of promotion pros-
pects or job security and little protection from the arbitrary or unilateral exercise
of management's authority to discipline or fire workers. Secondary jobs are typi-
cally found in smaller firms that lack the capital assets to have an extended inter-
nal labor market, or that are in industries such as retail trade and services where
the production technology is such that the product can be produced without
much firm-specific on-the-job training. Because firms in the secondary market
do not offer much training or have substantial job ladders, the wages they pay are
low, as are the incentives for job stability among the employees.
Given the existence of a primary and a secondary sector in the labor market,
the second crucial assumption of dual labor market theory is that mobility be-
tween the two sectors is limited. Some workers in the secondary sector prefer
those types of jobs and would not compete for primary jobs, even if given a
chance. Examples would be teenagers who seek part-time work after school or
married women who desire jobs that are easily interrupted for a period of time.
According to dualists, however, many workers in the secondary sector desire —
work in primary jobs, but are effectively excluded from the competition. The
major barrier to entry they face is discrimination.
Primary employers who offer good jobs have a long queue of job applicants. In
making its hiring decision, the primary employer ranks the job applicants from
highest to lowest based on an estimate of who will most likely possess the desired
traits of trustworthiness, low turnover, trainability, and so on. The dualists argue
that because employers perceive that white males are more likely to possess these
traits than blacks or women, a white male will be systematically favored over an ‘
equally qualified minority worker. Due to this type of discrimination, white
males gain access to the firm’s internal labor market and primary jobs, while mi-
nority groups are forced to accept jobs in the secondary sector. Thus, improving
the education or job skills of inner city workers would not improve their eco-
nomic status because these workers would still be denied access to primary jobs
because of the stigmatizing effects of their race, sex, and previous employment
history. The dualists argue that because minority workers are trapped in the sec-
ondary sector, they then adapt to the low wages and dead-end nature of second-
ary jobs by engaging in frequent job changing, tardiness, sloppy work habits, and
so on. These traits reinforce the perceptions of primary employers that minority
workers would not be good employees, setting off a vicious circle that confines
these workers to the secondary sector.
Dual labor market theory set off a considerable debate in labor economics in
the 1970s because it offered a markedly different view of the wage determination
process.’* One issue was whether or not there are identifiable primary and sec-
ondary sectors in the labor market. At an aggregate, economy wide level, little
evidence has been found of a bimodal distribution of earnings among individual
workers, nor has a ranking of earnings among individual industries revealed any
type of two-humped distribution as might be predicted from the dual model. As
several of the dualists have themselves admitted, a strict dichotomy of the econ-
omy into primary and secondary sectors is too much of a simplification to be
analytically useful. Piore, for example, expanded the model into four sectors: a
secondary sector, a primary sector divided into an upper and lower tier, and a
craft sector.” This too may be overly simplified; as the example of the chemical
plant in Figure 6.10 suggests, primary and secondary jobs often exist within the
same firm. While empirical evidence does not offer a great deal of support for a
dual segmentation of the labor market, the more general proposition that labor
markets are segmented, albeit it in finer detail, can be supported by more per-
suasive evidence.© Much of this evidence is reviewed in Chapters 8 and 9,
where discrimination and differences in occupational attainment among race
and sex groups are discussed.
A second major implication of dual labor market theory is that the financial
returns to years of education and experience should differ significantly depend-
ing on whether the individual is employed in the primary or secondary sector.
Two persons with identical years of education would have very different lifetime
levels of income if one were hired into an entry level position in the primary
sector and another were forced to work in a secondary job. Not only would the
level of income differ between the two, so would the rate of increase of earnings
with additional years of experience as the primary worker moved up the job lad-
der in the internal labor market and the secondary worker’s earnings stagnated in
a dead-end job. This prediction of dual labor market theory has found at least
some empirical support, and is a subject which is examined in more detail in the
next several chapters.°! >
58See Glen G. Cain, “The Challenge of Segmented Labour Market Theories to Orthodox The-
ory,” Journal of Economic Literature 14, no. 4 (December 1976): 1215-1257.
°° Suzanne Berger and Michael J. Piore, eds., Dualism and Discontinuity in Industrial Societies (New
York: Cambridge University Press, 1980), 1-12. Paul Osterman also distinguishes between four dif-
ferent types of employment systems, but calls them, respectively, the industrial, salaried, craft, and
secondary sectors. See “Choice of Employment Systems in Internal Labor Markets,” Industrial Rela-
tions 26 (Winter 1987): 46-67.
One recent statistical study of labor mobility among women, for example, found evidence of 16
distinct labor market segments. See Alisa Wilson, Women’s Interindustry and Occupational Mobility
Using a Multidimensional Model of Economic Segmentation (unpublished Ph.D. dissertation, University
of Southern California, 1984).
‘Empirical evidence on this aspect of dual labor market theory is presented in William T.
Dickens and Kevin Lang, “A Test of Dual Labor Market Theory,” American Economic Review 75
(September 1985): 792-805.
IN « THE « NEWS
Firms Develop Two-Tier Work Forces

Large Japanese companies have been famous for tablished some tenure with the firm. Driven by the
their two-tier employment systems composed of need to reduce costs and increase flexibility, how-
“core” groups of permanent employees and larger ever, many corporations have replaced thousands
groups of “contingent” workers with little or no of regular employees with temporary workers and
job security. In some cases, the contingent workers part-timers, or have bid-out the work to smaller
are directly affiliated with the parent company, in subcontractor firms. This brings two advantages.
others the contingent workers are employed by One is that the contingent workers cost less, partly
subcontractors or work at home. In either case, because their hourly pay is often only half that of
the result is something of a dual labor market regular employees, and partly because they receive
structure where one group of employees have far fewer fringe benefits such as paid vacations and
“good” jobs providing high wages, extensive fringe health insurance. A second benefit of contingent
benefits, considerable opportunities for promo- employees is that they can easily be let go when
tion, and job security, and another group of em- the company no longer needs them, in effect turn-
ployees have less desirable jobs paying low wages ing the quasi-fixed cost of regular employees into a
with few fringe benefits, and providing little, if variable cost.
any, job security. According to some estimates, the ranks of con-
Several recent articles in the popular press have tingent employees have doubled since 1980 to in-
noted a similar development in the United States. clude nearly 17 percent of the work force. One of
Throughout most of the post-World War II period, the most rapidly growing segments of the con-
a blue-collar or white-collar job with a large tingent work force is temporary help workers, a
American corporation generally meant lifelong group that now numbers around 800,000 people.
economic security, at least once the employee es- Almost half of temporary workers are employed as
SUMMARY This chapter begins with a look at wage determination in a perfectly competitive
labor market. If real-world labor markets satisfied all the assumptions of the per-
fectly competitive model, demand and supply would give rise to one unique
equilibrium wage for a particular type of labor. Since real-world labor markets
are to one degree or another imperfect, wage rates at any one point in time are
dispersed around the mean level of hourly earnings.
Wage rates change over time with shifts in the demand and supply curves of
labor. In a situation of excess demand, competition results in wage rates being
bid up. While in a perfectly competitive market an oversupply of labor would
result in a bidding down of wage rates, this typically does not occur because of
the existence of sizable hiring and training costs and other such factors. As long
as relative wages are flexible, demand and supply can be brought into balance by
competitive forces.
The polar opposite of perfect competition is monopsony. In a monopsony
market the firm is able to set a wage rate lower than would prevail in the case of a
competitive market, because workers lack alternative sources of employment. A
degree of monopsonistic exploitation may also occur in situations of oligopsony.
rr
a
SS a i a eee eee
office help, another 25 to 30 percent work in low- of the work to outside public relations firms and
skilled industrial jobs. In 1985, Apple Computer independent consultants. Through corporate re-
laid off 20 percent of its work force due to a slow- structuring such as this, the company cut operat-
down in sales. When business picked up again, ing costs in half.
rather than hire more regular clerical staff, the Perhaps the biggest worry from the development
company hired temporaries. The company’s man- of a two-tier work force is its greater economic in-
ager of staffing said of the decision, “If we bring equality among “insiders” and “outsiders,” and a
someone on board full-time, there is an implied sense that one group of workers is destined to live
obligation that the job won't disappear. But that with the revolving door of contingent work while
can't happen in an industry as volatile as ours.” another enjoys the security of high pay and perma-
Another source of rapid growth in the con- nent jobs. To some degree this fear is mislaid, for
tingent work force is the decision of companies to a sizable portion of contingent workers choose
subcontract out work that had formerly been done temporary or part-time work in order to gain job
in-house. The USX Corporation, for example, experience or more flexible work schedules. For
lowered the number of in-house hours of labor another portion, however, contingent work was
needed to produce a ton of steel from 10 in 1982 to thrust upon them when their regular jobs fell vic-
4 in 1986, in part by subcontracting out mainte- tim to corporate cost cutting, and for these work-
nance-type jobs. Rather than pay a union pipefit- ers the lower wages and skimpy benefits of con-
ter $13 an hour plus ample benefits, the company tingent work are accepted more out of necessity
got the same work done by a smaller nonunion than choice.
company that paid its workers $5 an hour and pro-
vided no benefits. Similarly, the Owens-Corning Sources: “The Disposable Employee Is Becoming a Fact of Cor-
porate Life,” Business Week (December 15, 1986): 52—56; and
Fiberglass Corporation chopped its public affairs “As Big Firms Continue to Trim Their Staffs, 2-Tier Setup
department from 31 people to 4, and bid-out much Emerges,” The Wall Street Journal (May 4, 1987): 4.
EELS ES ETE ES IT SET RES ELE II TSE EEE IEE IEA EOE, SELLE EEE LEEGE IDI ALPE LEI TEE O IEOL I LOI DEELI ITEP PE TEPE A LAE ELIES
Wage rates are determined not only by demand and supply, but also by labor
unions. Labor unions are able to raise wages through the strike threat, a restric-
tion of labor supply, and an expansion of the firm’s demand for union labor. In a
perfectly competitive market, the higher wages that unions gain result in a mis-
allocation of resources and economic inefficiency. In a monopsonistic market,
however, union wage gains may benefit workers and consumers and result in a
better allocation of resources.
If the labor market were perfectly competitive, the firm’s internal wage struc-
ture would be dictated by demand and supply in the external labor market.
Many firms have an internal labor market where rates of pay are somewhat shel-
tered from outside competition. Wage rates for the port of entry jobs at the bot-
tom of the job ladder are largely market determined, but within the internal la-
bor market the specific nature of job skills creates an indeterminacy in wage
rates. Firms often use job evaluation techniques to set precise wage rates for jobs
, in the internal labor market. Individual rates of pay for workers with similar jobs
are then determined by the type of compensation system used by the firm, such
as piece rates, time rates, profit-sharing, and so on.
If competitive forces are to allocate labor efficiently, there must be relatively

few impediments to labor mobility. Real-world labor markets are sometimes seg-
mented by factors such as internal labor markets, discrimination, and occupa-
tional licensing laws that create barriers to entry. One extreme version of market
segmentation is dual labor market theory that hypothesizes a primary sector
of good jobs and a secondary sector of bad jobs, with little mobility between
the two.
GLOSSARY
Area of indeterminacy— The range of wage rates Marginal cost of labor (MCL) schedule— A line
within which demand and supply allow firms that shows the change in the firm’s total cost of
discretion as to their pay policies. labor per hour for each additional worker hired.
Barriers to entry— Factors that impede the abil- Market imperfections — Factors such as imperfect
ity of new firms (workers) to enter a product information or nonmaximizing behavior that
(labor) market. cause labor markets to diverge from the per-
Efficiency wages— The wage rate per unit of out- fectly competitive model.
put produced. Merit pay plan—A compensation system in
Gain-sharing— A compensation system which which pay increases are based on the level of
bases at least a portion of employee pay on im- employee performance or judged by a supervisor
provements in the firm’s profits, productivity, or or manager.
cost reduction. Monopsonistic exploitation — The difference be-
Internal labor market — The group of jobs in the tween the dollar value of the output produced
firm that are filled through internal promotion by labor and the wage labor is paid by a monop-
rather than hiring from the external market. sonistic firm.
Job cluster — Jobs related to a common activity in Monopsony — A labor market with only one buyer
the firm. of labor.
Job evaluation — A technique used by firms to es- Port of entry— The entry level position on a
tablish an internal wage structure that ranks firm’s job ladder where hiring is done from the
jobs on the basis of various attributes such as external market.
skill, responsibility, and so on. Primary labor market — The sector in the econ-
Job ladder — A vertical hierarchy of jobs in the omy that contains the desirable jobs with high
firm which workers move up as they acquire wages, good promotion prospects, and employ-
training at each rung. ment stability.
Job structure — The set of jobs in terms of skill Secondary labor market — The sector in the econ-
and function required by a firm. omy that contains the undesirable jobs with low
Labor market segmentation—A development wages, few promotion prospects, and unstable
that impedes the mobility of workers between employment.
labor markets because of barriers to entry such Shock effect — The positive impact on manage-
as internal labor markets and occupational li- ment efficiency of an increase in the minimum
censing laws. wage or other such factor.
Law of one wage— An economic law that states Standard Industrial Classification (SIC)—A
that in a perfectly competitive labor market all classification system in which each firm in the
firms will pay the same wage rate for a particular nonagricultural economy is assigned to a particu-
type of labor. lar group based on the type of product produced.
a
OOOO RRRE>D}?)?» |S
eae
REVIEW QUESTIONS
1. How closely does the labor market for the fol- reconcile Kerr’s statement with Hicks’? Ana-
lowing type of worker approximate a_ perfectly lyze each statement in terms of the model of
competitive market? What, if any, are the major perfect competition, drawing graphs to illus-
imperfections in each market? trate your answer.
a. migrant farm worker 3. In recent years a large number of steel plants
b. steelworker in Pennsylvania have closed. Illustrate the impact
c. nurse of this development on the Pennsylvania labor
d. bank executive market. How, if at all, will the labor market return
2. Consider the following quotations: to equilibrium? Is the persistence of high unem-
ployment in Pennsylvania consistent with the
“for the general tendency for the wages of
model of perfect competition? Why or why not?
labourers of equal efficiency to become equal-
4. What has happened to the extent of monop-
ized in different occupations (allowances
sony in the labor market in the last 40 years? Re-
being made for the advantages and disadvan-
late your answer to the ease of labor mobility and
tages of employment) has been a common-
the slope of the supply curve of labor for the indi-
place of economics since the days of Adam
vidual firm.
Smith. . . The movement of labour from
5. When discussing unions the following state-
one occupation to another, which brings it
ment is often heard: “Unions were necessary when
about, is certainly a slow one; but there is
they started, but now they have gone too far.” Use
no need to question its reality.” [J. R. Hicks,
the model of monopsony to analyze what this
The Theory of Wages (New York: Macmillan,
statement means.
1932): 3.|
6. Every large firm has a personnel department
“Abundant evidence now testifies that it responsible for setting specific rates of pay for each
would, in the absence of collusion, be almost job. In performing this function, how much dis-
more correct to say that wages tend to be un- cretion does the personnel department have in
equal rather than the other way around. . . . terms of the rates it sets? To what extent do out-
Occupational wage rates, locality by locality, side market forces impinge on this decision? For
in the absence of collective bargaining dis- what types of jobs would market forces leave little
played no single going rate but a wide disper- room or much room for discretion? Why?
sion.” [Clark Kerr, “Labor Markets: Their 7. A common finding is that women’s earnings
Character and Consequences,” American Eco- increase at a slower rate with each additional year
nomic Review 40, no. 2 (May 1950): 280.] of work experience than those for men of similar
educational backgrounds. How would dual labor
Are the statements by these two prominent
market theory explain this?
economists in conflict? Why? Is it possible to
ee ene e
eee EE
ee
ADDITIONAL READINGS
Brown, Martin, and Peter Philips. “The Decline California canneries and why employers later
of Piece Rates in California Canneries: 1890— replaced piece rates with time rates.
1960.” Industrial Relations 25 (Winter 1986): Dunlop, John. “Labor Markets and Wage Deter-
81-91. Identifies the reasons for the widespread mination: Then and Now.” In Bruce E. Kauf-
use of the piece-rate system in the late 1890s in man, ed. How Labor Markets Work: Reflections
on Theory and Practice by John Dunlop, Clark tracts and compensation methods such as profit-
Kerr, Richard Lester, and Lloyd Reynolds. Lex- sharing and pay for performance.
ington, Mass.: Lexington Books, 1988: 50-103. Osterman, Paul. Internal Labor Markets. Cam-
Reviews the significant features of wage deter- bridge, Mass.: MIT Press, 1984. A collection of
mination that theoretical models must incorpo- articles describing the origins of internal labor
rate and critiques current research on wages. markets and their impact on the labor market.
Farber, Stephen, and Robert Newman. “Account- Reder, Melvin. “Wage Determination in Theory
ing for South/NonSouth Real Wage Differen- and Practice.” In Neil W. Chamberlain, et. al.,
tials and for Changes in Those Differentials A Decade of Industrial Relations Research—1946—
over Time.” Review of Economics and Statistics 69 1956. New York: Harper & Row, 1958: 64-97.
(May 1987): 215-223. Reviews previous re- A nontechnical discussion of the issues and
search on the South/non-South wage differ- controversies separating neoinstitutional and
ential and attempts to explain its causes. neoclassical theories of wage determination.
Mangum, Garth, Donald Mayall, and Kristin Rottenberg, Simon, ed. The Economics of Legal
Nelson. “The Temporary Help Industry: A Re- Minimum Wages. Washington: American Enter-
sponse to the Dual Labor Market.” Industrial prise Institute, 1981. A large collection of stud-
and Labor Relations Review 38 (July 1985): 599— ies that examines all aspect of the minimum
611. Examines the reasons for the rapid growth wage.
in the temporary help industry and the implica- Topel, Robert. “Local Labor Markets.” Journal of
tions for workers and firms. Political Economy 94, part 2 (June 1986): S111-
Nalbantian, Haig R., ed. Incentives, Cooperation, S143. Examines wage and employment dynam-
and Risk Sharing. Totowa, N.J., Rowman and ics in local labor markets and finds that wage
Littlefield, 1987. Contains a series of articles by rates are sensitive to changes in demand and
economists and other compensation specialists supply conditions.
on the theory and practical design of labor con-
APPENDIX 6A,
Data and Reference Sources in Labor Economics
Many students are required to write a paper or complete some type of research
project in their labor economics class, a task that is usually begun at about this
point in the course. To help the student in this process, this appendix provides a
list of the major data and reference sources in labor economics.
GUIDES 1. Azevedo, Ross E. Labor Economics: A Guide to Information Sources. De-

troit: Gale, 1978. An annotated section on textbooks and general works is
followed by listings of journals, government publications, and information
services. Books and articles are then listed under such specific topics as
collective bargaining theory, manpower, and workers’ compensation. Au-
thor, title, and subject indexes.
2. Hoel, Arline, Kenneth Clarkson, and Roger Leroy Miller. Economics
Sourcebook of Government Statistics. Lexington, Mass.: Lexington Books,
1983. Provides a complete listing of the statistical series published by the
federal government that are germane to economics. Contains one chapter
on statistics in the labor area. Discusses the definition, method of collec-
tion, and issuing agency for each statistical series.
3. Paradis, Adrian, and Grace Paradis. The Labor Almanac. Littleton, Colo.:
Libraries Unlimited, 1983. Provides a thorough listing of sources of infor-
mation on labor topics as well as the names of labor leaders, organizations,
law, and government agencies.
DICTIONARIES 4. Ammer, Christine. Dictionary of Business and Economics. New York: Free
Press, 1977. Contains over 3,000 entries and includes information on im-
portant economists and schools of thought.
INDEXES AND 5. The New York Times Index. New York: New York Times. Semimonthly with
ABSTRACTS: cumulations. A comprehensive index to one of the country’s leading
NEWSPAPERS newspapers.
6. The Wall Street Journal Index. New York: Dow Jones. Monthly with annual
cumulations. The first half of the index is arranged by company name; the
second half by subject headings.
INDEXES AND 7. Business Periodicals Index. New York: Wilson. Monthly with cumulations.
ABSTRACTS: A subject index to articles pertaining to business and economics. Over 170
PERIODICALS periodicals are covered.
8. The Conference Board Cumulative Index. New York: The Conference
Board. Annual. Indexes publications of the Conference Board, an inde-
pendent nonprofit business research organization; covers numerous sub-
jects in the labor area.
Dilek
278 Appendix 6A Data and Reference Sources in Labor Economics
Human Resource Abstracts. Beverly Hills, Calif.: Sage. Quarterly. Ab-

stracts books, reports, and periodicals on topics such as employment and
unemployment, earnings and benefits, labor force participation, and man-
power policy. Abstracts are arranged by broad subject category, with au-
thor and subject indexes.
10. Index to U.S. Government Periodicals. Chicago: Infordata International.
Quarterly with cumulations. Covers all U.S. government periodical ar-
ticles with research value.
IL. Journal of Economic Literature. Nashville: American Economic Associa-
tion. Quarterly. Indexes scholarly articles in the labor area. Includes anno-
tated listing of new books, abstracts of major articles, and a large number
of book reviews.
12. Monthly Labor Review. Washington: Bureau of Labor Statistics. Monthly.
Indexes books and publications received. Also includes book reviews.
13. Work Related Abstracts. Detroit: Information Coordinators. Monthly with
cumulative annual index. Abstracts 250 management, union, govern-
ment, and academic periodicals. Arranged into 20 broad subjects such as
employment, union organizing, and collective bargaining.
BIBLIOGRAPHIES 14. International Bibliography of Economics. Chicago: Aldine. Annual. Pub-

lished under the auspices of UNESCO. Covers a wide range of economic
publications from around the world. Indexed by subject.
15. BLS Publications. Washington: Bureau of Labor Statistics. Contains a nu-
merical listing of bulletins with some annotations; a numerical listing of
reports; and a list of BLS periodicals. Updated by the semiannual Publica-
tions of the Bureau of Labor Statistics.
16. Major Programs, Bureau of Labor Statistics. Washington: Bureau of Labor
Statistics. Presents in condensed form the scope of the major programs of
the BLS, the principal data-gathering agency of the federal government in
the labor area. The following is given for each major program: the data
available, the coverage, the source of the data, the reference period, the
publications issued, and the principal use of the data. Data available from
individual states are also listed.
ACADEMIC 17. The major academic journals in the field of labor economics and industrial
JOURNALS relations are:
Industrial and Labor Relations Review
Industrial Relations
Journal of Human Resources
Journal of Labor Economics
Journal of Labor Research
Labor History
Labor Studies Journal
Proceedings of the Industrial Relations Research Association
Appendix 6A Data and Reference Sources in Labor Economics 279
NONACADEMIC 18. The articles in the academic journals listed in no. 17 often require a fairly
JOURNALS high level of technical expertise in economic theory, mathematics, and
statistics. Two nonacademic journals that contain articles written for the
layperson are:
International Labor Review
Monthly Labor Review
STATISTICAL Statistical Abstract of the United States. Washington: Department of Com-

SOURCES merce, Bureau of the Census. Annual. Provides a wealth of statistical data
on labor subjects as well as for many other areas. Also serves as a guide for
obtaining more complete data from other sources.
20. Census of the Population. Washington: Department of Commerce, Bureau
of the Census. Every 10 years. Provides the most detailed information
available on earnings, employment, occupational attainment, and other
such subjects for both the nation and individual states.
zie Money Income of Households, Families, and Persons in the United States. Cur-
rent Population Report P-60. Washington: Department of Commerce, Bu-
reau of the Census. Annual. Presents detailed data on annual earnings and
income of Americans classified by demographic, industry, and occupa-
tional characteristics. Derived from the Current Population Survey.
Vip Historical Statistics of the United States: Colonial Times to 1970. Washington:
Department of Commerce, Bureau of the Census, 1975. A compilation of
historical statistics on the labor force, employment, earnings, and union
membership, among other subjects. Many statistical series extend back to
the 1800s.
Zs Employment and Earnings. Washington: Bureau of Labor Statistics.
Monthly. Presents monthly data obtained from the Current Population
Survey and the survey of business establishments. The basic source for cur-
rent labor force and earnings data.
24. Monthly Labor Review. Washington: Bureau of Labor Statistics. Monthly.
Also presents current labor force data, as well as data on consumer and
producer prices, productivity, and strikes.
La. Handbook of Labor Statistics. Washington: Bureau of Labor Statistics. An-
nual, but 1985 is latest edition. Provides a compilation of much of the
statistical data gathered by the BLS. The 1975 “Reference Edition” pro-
vides complete historical data.
26. Current Wage Developments. Washington: Bureau of Labor Statistics.
Monthly. Provides details on the contract terms of major union agree-
ments, plus statistical data on various measures of compensation. April or
May issue has an annual review article.
Dis Economic Report of the President. Washington: G.P.O. Annual. Appendix
contains detailed information on income, employment, and production.
28. Annual Report of the National Labor Relations Board. Washington: G.P.O.
Annual. Presents detailed data on number of unfair labor practice charges,
their disposition, petitions for representation elections, the union win-rate
by state, industry, and so on.

770159799 Kaufman 2e the Economics of Labor Markets and Labor Relations (1)

Uploaded by

Copyright:

Available Formats

770159799 Kaufman 2e the Economics of Labor Markets and Labor Relations (1)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

770159799 Kaufman 2e the Economics of Labor Markets and Labor Relations (1)

Uploaded by

Copyright:

Available Formats

Celpoue

The Labor Market

Economics iis concerned wee ae and

Labor economics is both an interesting and a useful subject. One reason is

FIGURE 1.1 A Circular Flow Model of the econo

spend. Pe 1987, he ee miaedan rc by penercan ores amounted

FIGURE 1.2 Trends in Three Dimensions of Labor Supply

(b) Labor Force Participation Rate (LFPR)of = =«_—

1900 20 College educated _

TABLE 1.1 Percentage Share of

. During EE aerecession th as 1974

FIGURE 1.3 Level and Composition of Unemployment, 1950-1987

and Handbook of Labor Statistics, 1985, Table 26.

ers’ paychecks were increasing by an average of 3 to 4 percent a year; by the late

FIGURE 1.5 The Pattern of Poverty and Income Inequality, 1986

Poverty line (family

a eT Gee Much as with the business firm, the indi-

FIGURE 1.6 The Determination of Wage Rates by Demand and Supply

FIGURE 1.7. The Market panne toseit in oLreny and Ey

the wage until a new

librium of W, (point X) to a new equilibrium of W, (point Z) where demand and

350 Newly qualified ,.

1960 1964 1968 1972 1976 1980 1984

(b) Relative Earnings Ratio

1960 1964 1968 1972 1976 1980 1984

institutional forces. ecu Porees represent the influence of various or-

inscompetition with each other to win the

See O. E. Williamson, M. L. Wachter, and J. E. Harris, “Understanding the Employment Re-

The neoclassical school of labor economics is distinctive in terms of both its

to derive theoretical predictions and testable hypotheses.

as clear-cut as it sounds because of the complexity of human behavior and the

As Ray Marshall, President Carter's secretary of labor, argued, every proposed

= treads iol iin

' eanyt ‘ speciemea

. Graph (b), on the other hand, provides an illustration of time-series

1-4 5-14 15-29 30-34 35-39 40 41-48 49-59 60+ Hours

(b) Time-Series Pattern

classified as a full-time worker, a person must work 35 or more hours a week,

In the terminology of Chapter 1, the cross-sectional and time-series pattern of

change simply because of a change innae preferences or tastes concerning it.

Preferences and about

Te ac several important properties. First, as a matter of

Same Vera o& Okt :

(a) An Individual with (b) Two Individuals with

A third property ofindifference curves isthat there isawhole set ofseparate _

es. This is illustrated in

A UTRRNA TR RE nee «fi B, for example, may have a rela-

FIGURE 2.4 A Set of Budget Constraints

0 100 Hours of leisure —>

A third property of the budget constraint imanecntinondlithinlitOalntis

in the wage and flatter for a fall in the wage.

FIGURE 2.5 Equilibrium Hours of Work

0 50 100 Hours of leisure >

FIGURE 2.6 The Effect on Hours of Work of an Increase in Nonlabor Income

Hours of leisure —>

effect. There is a second force that —_ in i opposite mii sere

FIGURE 2.7 A Wage Change Leading to Increased Hours of Work

FIGURE 2.8 A Wage Change Leading to Decreased Hours of Work

t a higher wage rate income is also greater, leading a person to “buy”

H, H, Hours of 0 H’, H, Hours of