Chapter 7: Discrimination, Part 2
Econ 3050/WMST 3650 Fall
2000
From your book:
The
Overcrowding Model. Barbara Bergmann developed
the overcrowding model. Her model shows
that a consequence of occupational segregation may be a male/female pay
differential. This is independent of
the reason for the segregation, e.g., due to gender differences in
socialization, or because of labor market discrimination. The wage gap occurs if the demand for female
employees is small relative to the supply of female workers. The overcrowding model is consistent with
the evidence that, ceteris paribus,
earnings tend to be less in female-dominated occupations than in male-dominated
occupations. That men who work in
female-dominated occupations earn less, on average, than men who do not, is
consistent with the crowding model. Why
these men work for lower wages in female dominated professions is not clear –
are they trading wages for some nonpecuniary benefit, do they lack information
about alternative opportunities, or are they just unlucky? The important point is that the low wages in
these jobs result from the many women who crowd into these jobs.
Figure 7.4 assumes that men and women are perfect
substitutes for each other, and they are equally productive. This model also assumes that there are no
nonpecuniary differences in the relative attractiveness of F or M jobs. If there were nonpecuniary differences in
the two types of jobs, e.g., M is riskier than F, pay differences could be
caused by a compensating wage differential.
In other words, because M is riskier than F, wages will be higher at M
than at F to compensate for the higher degree of risk.
If there is no discrimination in the market, in
equilibrium, demand for women and men will be Df and Dm,
and supply of women and men will be Sf0
and Sm0. Because men and women are equally qualified
for both types of jobs, F and M, employers are indifferent between hiring men
and women. If the wage in one market is
higher than the wage in the other market, workers in the lower paying jobs will
migrate to the higher paying jobs. This
will increase in the supply of workers in the higher paying market, and
decrease the supply of workers in the lower paying markets. Wages will fluctuate until wages in both
markets are the same. Thus, the
nondiscriminatory equilibrium occurs when men and women earn the same wage
rate, W0.
When there is market discrimination against women,
or, if women choose to concentrate in certain types of jobs, a gender-based pay
differential will develop. For example,
overcrowding results in the supply of women and men becoming Sfd and Smd. The
exclusion of women from M jobs means that women are crowded into F jobs, which
reduces women's wages, and raises men's wages relative to the nondiscriminatory
equilibrium.
The point is that overcrowding could result in a
wage differential. (It may not if
supply equals demand at the same wage in female-dominated professions as it
does in male-dominated professions.)
The wage differential is a direct result of whatever it is that reduces
women's labor market mobility relative to men's. The key here is that crowding is due to something that inhibits
women's occupational mobility.
Overcrowding may also affect women's labor market
productivity. Remember that the more of
an input you use, the smaller is its marginal (physical) product. The less of an input you use, the larger is
its marginal product. If labor is
relatively cheap in female-dominated professions, employers will use more of it
and/or less capital. If labor is
relatively expensive in male-dominated professions, employers will use less of
it, and/or more capital. Thus, (at the
margin) the last woman hired in the female-dominated profession is less
productive than the last man hired in the male-dominated profession. Women become relatively less productive 1)
because of crowding and 2) because they work with less capital.
Institutional
Models. When economists talk about “institutions”
they are referring to the government, unions, or large corporations. Institutional models of labor market
discrimination emphasize that labor markets may not be as flexible as the
simple competitive market assumes.
Rigidities are introduced by the institutional arrangements found in
many firms, and by the barriers to competition introduced by monopolies –
either monopolies in the product market or unions in the labor market.
1. Internal Labor Markets. Firms sometimes create internal labor markets in which applicants are hired for
entry-level jobs, and all other jobs are filled internally. This system of hiring and promotion can
substitute for the difficult task of screening job applicants based on
unobservable (at the time of hire) characteristics. Workers are hired to perform jobs with low levels of responsibility,
and then are observed during a probationary period to determine their actual
productivity. One can construct
compensation schemes that are designed to motivate workers in internal labor
markets. These schemes require an
individual's efforts to be monitored over long periods of time so even if the
probability of detecting shirking in any one year is small, anyone who shirks
will eventually be observed.
2.
Dual Labor Market. The labor market is divided
into two sectors:
A. Primary Sector
· High levels of firm-specific OJT leads to high wage jobs
· Good working conditions
· Opportunities for advancement
· Low labor turnover
B. Secondary Sector
· Firm-specific skills are not important resulting in low wage jobs
· Bad working conditions
· Unstable employment
· Few opportunities for advancement.
· High labor turnover
Mobility barriers exist between the two
sections. The main mobility barrier is
statistical discrimination. That is,
primary sector jobs are rationed based on employers' perceptions of group averages. They tend to go to white males rather than
women and minority males. Why? Because employers believe that white males
are more likely than individuals from other groups to possess the desired
unobservable characteristics, e.g., a low mobility propensity, trainability,
trustworthiness, etc. The direction of
the causation between the probability of having the desired characteristic, and
having a secondary sector job is not clear – is it nature or nurture? Being trapped in the secondary sector and working
at low-wage dead-end jobs may cause individuals to engage in frequent turnover,
tardiness, etc.
Feedback
Effects. Feedback effects occur when labor market
discrimination or unequal treatment of women in labor markets may adversely
affect women's own decisions and behavior.
Figure 7.4 shows the two-way relationship between
pre-market experiences and labor market outcomes. Discrimination against women in the labor market reinforces
traditional gender roles in the family.
At the same time, adherence to traditional roles by women provides a
rationale for labor market discrimination.
A decrease in labor market discrimination will have feedback effects as
the equalization of market incentives between men and women induce additional
changes in women's supply-side behavior – their human capital investments. As more women enter previously
male-dominated fields, the larger number of female role models for younger
women may induce further increases in the availability of women for such
jobs. Thus, demand-side policies can be
expected to play an important role in sustaining a process of cumulative change
in women's economic status.
The Government and Equal
Employment Opportunity. Legislation to
end Discrimination.
Federal
Programs to End Discrimination. The
federal government has enforced two sets of rules in an attempt to eliminate
market discrimination:
1. A nondiscrimination requirement imposed on
almost all employers; and
2. The requirement that Federal contractors
engage in affirmative action.
The Equal Pay Act of 1963. This act effectively:
1)
Overturned
the protective legislation; and
2)
Outlawed
separate pay scales for men and women using similar skills and performing work
under the same conditions.
Prior to the 1960s, sex discrimination was officially
sanctioned through protective labor laws.
Protective labor laws limited the total number of hours that women could
work, prohibited them from working at night, lifting heavy objects, and working
during pregnancy. Not all states placed
all of these restrictions on women, but the effect of these laws was to limit
the access of women to many jobs.
The
Equal Pay Act of 1963 was deficient as an antidiscrimination tool, because it
"said nothing about equal opportunity in hiring and promotion." If there is prejudice (from any source)
against women, employers will treat the female employees as if they were less
productive or more costly to hire than the equally productive males. The market response is for female wages to
be less than male wages otherwise women will not be able to compete with men
for jobs. Essentially the law
suppressed a market mechanism that helped women have access to jobs. The act failed to acknowledge that for labor
market discrimination to be eliminated, legislation must require both
equal pay and equal opportunities in hiring and promotions for people of
comparable productivity.
Title
VII of the Civil Rights Act of 1964. Some defects in the Equal Pay Act of 1963 were
corrected the next year. Title VII made
it unlawful for any employer "to refuse to hire or to discharge any
individual, or otherwise to discriminate against any individual with respect to
his compensation, terms, condition, or privileges of employment, because of
such individual's race, color, religion, sex or national origin."
The
new law also addressed Union practices.
Historically, it had been difficult for racial minorities to obtain
admission into certain craft unions.
This exclusion denied minorities access to both the skills acquired
through union apprenticeship programs, as well as the employment opportunities
dispensed through union hiring halls.
Title VII made it unlawful for any labor organization to exclude
individuals from membership, to segregate membership, to refuse to refer for
employment, or to discriminate in admission to the apprenticeship programs on
the basis of race, color, religion, sex or national origin.
Effectiveness.
1.
Title
VII was not retroactive. It only
applied to acts of discrimination that occurred after July 1, 1965.
2.
The
law permitted exceptions to the general requirement of nondiscrimination,
"where religion, sex, or nation origin is a bona fide occupation
qualification reasonably necessary to the normal operation of a business." In practice this has limited applications,
e.g., certain jobs in religious organizations, nursing homes where patients are
predominantly of one sex, etc.
3.
It
allowed wage and other employment differentials that resulted from "a bona
fide seniority system . . . provided that such differences are not the result
of an intention to discriminate."
4.
No
party subject to the statute is required to grant preferential treatment to any
group based on existing imbalances in the work place.
Title
VII applies to all employers involved in interstate commerce with at least 25
employees. The Equal Employment
Opportunity Commission (EEOC) enforces it.
EEOC has the authority to mediate complaints, encourage lawsuits by
private parties or the U.S. Attorney General, or (since 1972) bring suites
itself against employers that have violated the law.
Over
the years, the federal courts have fashioned two standards of discrimination
that may be applied when discriminatory employment practices are alleged:
1.
Disparate treatment occurs when 1) individuals are treated differently, e.g., paid
different wages or benefits, because of race, sex, color, religion, or national
origin; and if 2) it can be shown there was an intent to discriminate.
2.
Disparate impact. Under disparate impact,
personnel policies that appear to be gender or racially neutral but lead to
differential impacts by race and gender are prohibited unless they lead to job
performance. Under this approach, it is
the result not the motivation that matters. For example, job application forms may ask
about convictions but not about arrests.
(Minorities tend to have higher arrest rates.)
While
disparate treatment is the more obvious approach to defining discrimination, it
is not the definition that the courts have relied on the most frequently. The adoption of the disparate impact
standard by the courts as a standard of discrimination has mounted a
significant challenge to employer personnel screening devises. In recent years two difficult issues have
arisen in the application of the law – the treatment of seniority and
comparable worth.
Seniority. A seniority system has a strong potential for perpetrating the
effects of past discrimination. For
example, in many Southern firms, job segregation was accompanied by departmental
seniority arrangements, where seniority is time in the department and not time
with the firm (tenure). In the late
1960s, when companies sought to comply with Title VII, they a) moved women and
minority men from the low-wage jobs into higher-wage jobs in other departments;
and b) increased the hiring of women and minorities. Under both (a) and (b), women and minority men had relatively low
levels of seniority under departmental seniority systems. In the late 1970s when economy took a
downturn, firms began to lay off workers.
A disproportionate number of those who were laid off were women and
minority men. The irony is that in many
of these cases individuals with little departmental seniority had a lot of
plant seniority, and they might have retained their jobs if they had not been
promoted.
The
courts were in a quandary. "Bona
fide seniority systems" were protected under Title VII, but under the
disparate impact standard they were discriminatory. The lower courts took the position that a seniority system was
not bona fide if it discriminated, and thus, only the plant-wide seniority
systems were bona fide. The Supreme
Court reversed the decision (International Brotherhood of Teamsters v US
(1977)).
This
brings up the notion of fictional seniority, which is seniority retroactive to
the date the individuals would have been hired if the firm had not
discriminated. The Supreme Court has
ruled that frictional seniority is an appropriate remedy for individuals who
can show they are victims of past discrimination. It is not, however, appropriate to dismiss current employees as
part of the remedy of past discrimination (Franks v Bowman Transportation
(1976) and Fire Fighters Local 1784 v Stotts (1984)).
Comparable
worth. In the early 1970s, the American Federation
of State, County and Municipal Employees (AFSCME), the labor union representing
government employees in the state of Washington, began pressing a novel
idea. As a class, female state workers
were being paid less than their male colleagues were. And, although women and men didn't perform identical jobs, their
different jobs demanded similar qualifications. Since they were comparable, they were of comparable worth. In 1974 the state officials commissioned a study
by management consultant Norman Willis.
The study examined 62 job classifications in which 70% or more of the
employees were women and 59 job classifications in which 70% or more of the
employees were men. The results from
this study indicated that in jobs of comparable worth, women were paid about
20% less than the men.
The
consultants created four criteria and allocated points to each one.
Knowledge
and skills 280
Mental
demands 140
Accountability 160
Working
conditions 20
A five-member committee scored each job within the
121 job classifications. For example if
a highway maintenance worker and a clerk typist had about the same total scores
their jobs were declared to be of comparable worth and therefore should have
equal pay. When the state government
balked at adopting the Norman Willis study, in 1982 the union went to the
District Court. The union claimed that
failure to pay equal salaries for jobs of comparable worth is a violation of
Title VII. In 1983, District Judge Jack
E. Tanner ruled in favor of the union.
He decided the state acted in bad faith when they commissioned the study
whose findings of pay discrimination were then not implemented. He awarded 15,500 state employees primarily
women back pay estimated at $800,000,000 to $1,000,000,000. In mid-September 1985 the U.S. 9th Circuit
Court of Appeals reversed the 1983 State of Washington ruling. Implementing a comparable worth policy is
difficult in terms of the operational difficulties of deriving job evaluation
procedures. In addition to the
operational difficulties, comparable worth also raises several policy
questions. Is it likely to be more
effective to use the law to break down occupational barriers caused by
discriminatory job segregation, or to develop a compensation scheme that may
leave job segregation unchanged? Will
raising the wages of traditional "female" occupations reduce women's
incentive to enter nontraditional occupations?
Will raising women's wages adversely affect their employment
levels?
The
empirical evidence. What are the direct effects
of gender-based labor market discrimination?
To answer this question you have to take as given any gender differences
in qualifications. (In other words, you
have to ignore the reason for any gender differences in labor market
qualifications, which begs the question: Are differences in qualifications due
to differences in men's and women's voluntary choices, or do they result from
labor market discrimination?) The book
authors posit two questions.
1)
Are
gender differences in labor market outcomes fully explained by gender
differences in qualifications, or (potential) productivity? [The answer is no.]
2)
If
not, how large is the unexplained portion of the gender differential? [It varies depending on the time period, the
sample, and the variables used in the analysis.]
Earnings
Differences.
There are a plethora of studies on the
male-female earnings gap. All of these
studies have found that that a substantial portion of the pay gap cannot be
explained by gender differences in qualifications.
Q. What do we mean by gender differences in qualifications?
Table
7.1 gives an example of gender differences in qualifications for a sample of
full-time workers who were 18 to 65 in 1989.
What does Table 7.1 show? The
most important difference in work-related qualities is that men have more full-time
work experience than do men, and as we have seen--more experience results in
higher pay. Women do have more
part-time experience, which doesn't pay as much as full-time experience. Also of importance are occupational and
industrial differences in men's and women's jobs. Women are also less likely than are men to be members of a
union.
Table
7.2 shows you the consequences of these differences in work-related
characteristics. Controlling for all of
these work-related characteristics reduces the amount of the gender-based wage
gap that can't be explained from 27.6 cents to about 10.5 cents. Figure 7.1 shows the same thing. That is, controlling for differences in
human capital characteristics reduces the part of the wage gap that we can't
explain.
Two
things affect the quality of the evidence presented in Table 7.2 and Figure
7.1. First, we don't have information
about all of the characteristics of individuals that affect their productivity. For example, we usually don't have any
information regarding motivation, dependability, etc. If men are more qualified than women with respect to information
that is excluded from the analysis, the extent of labor market discrimination
is likely to be overestimated. If women
are more qualified than men with respect to information that is excluded from
the analysis, the extent of labor market discrimination is likely to be
underestimated. For example, when I was
at the University of Alaska Fairbanks, I examined salary differences between
male and female faculty members. Among
the associate professors, the women were more likely to have taught at Tanana
Community College, which merged with UAF in 1987. As a result, they were less likely than the male associate
professors to have a Ph.D. and their base salaries were more likely to have
been determined by a collective bargaining agreement. Both things would misstate the extent of gender-based
discrimination associated with professorial rank at UAF.
Second,
if differences in work-related characteristics are due to pre-market
discrimination (e.g., barriers to education, etc.), or labor market
discrimination (e.g., barriers to specific occupations, OJT, etc.), this kind
of study will seriously underestimate the impact of discrimination. Other methods would have to be used to
ferret out the impact of this type of discrimination. Your book mentions evidence using the following non-traditional
(to economists) methods.
1)
Audit
studies (also mentioned in "Evidence of Discrimination in Employment:
Codes of Color, Codes of Gender") that are used to determine if hiring
decisions are subject to labor market discrimination. Audit studies use sets of equally qualified individuals who
differ only in their sex, or race, or age, etc.
2)
Counting
the number of employment discrimination cases in which employers have been
found guilty, or reached an out-of-court settlement with the plaintiffs. (Be a little careful with inferring much
from this. If you reduce the cost of
anything, you should get more of it. If
it is easier to sue, we should have more lawsuits. Second, it may be cheaper to settle than to go to trial.)
3)
Contemporary
public opinion surveys that show people believe women are less qualified in
general than are men, and/or women are discriminated against in the labor market.
Occupational
Differences. We've seen that women are more likely than
men to be concentrated in clerical and service jobs, and men are more likely
than women to work in higher paying jobs such as skilled craft workers. And, the more narrowly you define
occupations, the more occupational segregation by gender you're likely to
see. For example, the information in
Table 5.1 (p. 124) states that the segregation index was 41.9% in 1972, and
33.8% in 1995. Using more detailed
census data shows that the segregation index was 67.7% in 1970, 59.3% in 1980,
and 52% in 1990. Along the same lines,
even though the percentage of men and women in the professional category is
about the same, men are more likely than women to work in higher paying
professional specialties such as law, medicine and engineering. Women are more likely to be employed in the
lower paying professions such as public education.
1.
What
are the consequences of occupational segregation? The major consequence of occupational segregation is that it
contributes significantly to the wage gap.
Your book points out the results from a couple of studies that have
found controlling for differences in men's and women's occupations can account
for 8 percent to 14 to 23 percent (when occupations are narrowly defined) of
the wage gap. Moreover, even in the
same occupational category, women tend to be employed in low-wage firms and
industries, whereas men tend to be employed in higher-wage firms and
industries. An example, might be a
manager of a day care center versus the manager of an automotive plant. Both have the same occupation, manager, they
just work in different industries. Your
book states the results from a study that found that controlling for gender
differences in major industrial group and unionism explained an additional 27
percent of the wage gap. Further,
controlling for detailed occupation, industrial differences account for 12 to
17 percent of the pay gap among equally qualified male and female workers.
2.
What
are the causes of occupational segregation.
The causes of occupational segregation can be classified into
supply-side and demand-side factors.
Supply-side factors are
differences in men's and women's qualifications that occur because of
differences in men's and women's human capital investments. Some of these differences are the results of
voluntary choices. If women expect to
exit the labor market for a period of time, they will be more likely to invest
in the following type of education and skills.
1)
Those
that are useful in both the market and the home (e.g., nursing, etc.); and/or
2) Those that are not likely to
depreciate quickly or to become obsolete.
Other
decisions may be subject to feedback effects.
(It isn't possible to distinguish between the two in most economic
studies because the information that would allow that kind of distinction is
not available.)
Demand-side
effects represent differences in treatment.
In other words, if women are hired for specific types of jobs, and men
for others based solely on sex. It
isn't possible to sort out how much of the occupational segregation is due to
differences in choices (completely voluntary or not), and how much is due to
differences in treatment. Both,
however, are important.