Order Code 98-278 E
CRS Report for Congress
Received through the CRS Web
The Gender Wage Gap and Pay Equity:
Is Comparable Worth the Next Step?
Updated December 20, 2004
Specialist in Labor Economics
Domestic Social Policy Division
Congressional Research Service ˜ The Library of Congress
The Gender Wage Gap and Pay Equity:
Is Comparable Worth the Next Step?
Women’s participation in the labor market has undergone considerable change
in the last few decades, with more than half of all women now in the workforce. In
addition, legislation and regulations have, since the 1960s, prohibited discrimination
against women in employment and pay. Although women’s pay relative to men’s has
increased over time, so that today women typically earn 76-79 cents for every dollar
earned by men, the persistence of the gender wage gap in the face of these changes
has prompted concern in some quarters about the equity or fairness of the market’s
wage-setting process (hence the terms “pay equity” and “fair pay”).
Studies have estimated that perhaps one-half of the observed gender wage gap
can be justified by productivity differences, measured by work experience and
educational attainment for example. If women had the same human capital attributes
as men, they might earn about 80% as much as men. Job-related factors (e.g.,
occupation and industry of employment) affect relative earnings as well. Studies that
include both sets of characteristics estimate that they might explain two-thirds of the
observed wage gap. If women were like men in terms of their individual and job
attributes, they might earn about 90% as much as men. Some believe that the
unexplained portion of the wage gap represents discrimination; others, an inability
to accurately measure and include all factors that affect gender differences in pay.
“Comparable worth” supporters contend that corrective action is needed because
employment discrimination relegates women to different jobs than men and wage
discrimination causes women’s work to be “devalued,” that is, paid lower wages than
jobs predominantly employing men. A comparable worth policy would extend the
current mandate of equal pay for equal work to equal pay for equivalent work within
a firm. (It thus would not directly address sex segregation in employment by job,
occupation, industry, and firm size.) Under comparable worth, an employer would,
through such means as an unbiased job evaluation, determine those jobs that had
equal total scores for such job attributes as skill, effort, responsibility, and working
conditions. The employer would then raise the wages of jobs deemed underpaid
(e.g., jobs having wages below other jobs with the same total scores). In this manner,
workers would no longer incur a wage penalty for employment in traditionally female
The size of a worker’s paycheck would be unrelated to supply/demand
conditions in the labor market under a comparable worth policy. Some highly rated
jobs within a firm may not warrant a pay increase if there is an abundant supply of
workers in the labor market to perform them, however. In other words, legitimate
(nondiscriminatory) pay differentials can exist between jobs equally rated by an
evaluation. Critics regard the substitution of job evaluations for market conditions
to determine relative wages as a critical flaw of comparable worth: by eliminating
wage differentials between “equivalent” male- and female-dominated jobs, it could
increase unemployment of women in the short run as well as remove the strongest
motivation for women to overcome discrimination in the long run. This report will
be updated as warranted.
The Male-Female Wage Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Reasons Suggested for the Wage Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
The Human Capital Explanation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
The Discrimination Explanation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Sex Segregation by Occupation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Why Does Segregation Depress Women’s Relative Wages? . . . . . . . . 8
Some Portion of the Wage Gap Remains Unexplained . . . . . . . . . . . . . 9
Reasons Suggested for the Trend in the Wage Gap . . . . . . . . . . . . . . . . . . . . . . . 11
The 1980s: Rapid Improvement in Women’s Relative Wages . . . . . . . . . . 11
The 1990s: Slowed Relative Pay Gains for Women . . . . . . . . . . . . . . . . . . 13
Policy Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Explanations and Remedies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Comparable Worth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Congressional Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Occupational Desegregation as a Remedy to the Wage Gap? . . . . . . . 17
Job Evaluation as a Wage-Setting Mechanism . . . . . . . . . . . . . . . . . . 18
The Economic Effects of Implementing Comparable Worth . . . . . . . 21
Comparable Worth’s Potential Impact on the Wage Gap . . . . . . . . . . 22
Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
List of Tables
Table 1. Ratio of Female-to-Male Earnings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
The Gender Wage Gap and Pay Equity:
Is Comparable Worth the Next Step?
Women’s participation in the labor market has undergone considerable change
in the last few decades. Women comprised 47% of all workers in 2003, up from 33%
in 1960. More than one-half of the noninstitutional female population age 16 or
older has been in the paid work force since 1979, according to U.S. Bureau of Labor
Statistics’ data. The majority of today’s married women have forsaken the traditional
role of full-time child care givers for the dual role of unpaid homemakers and paid
Over the same period, legislation was enacted with the intent of mitigating labor
market discrimination against women among other groups. The Equal Pay Act of
1963, an amendment to the Fair Labor Standards Act, mandates equal pay for men
and women employed in the same or substantially same jobs in a firm. The Civil
Rights Act of 1964 prohibits discrimination in employment and compensation against
women and other protected classes of workers. E.O. 11246 also forbids labor market
discrimination and requires affirmative action for protected classes of workers at
federal contractors and subcontractors.
One constant over time, however, has been the lower wages of women
compared to men. The persistence of the gender wage gap in the face of workplace
antidiscrimination edicts and of significant changes in women’s labor force
participation has prompted concern in some quarters about the “equity” or “fairness”
of the market’s wage-setting process.
It is contended that corrective action, which goes beyond current law, is needed
because ongoing discrimination against women is distorting the allocation of human
resources and creating inefficiencies in the labor market. Others believe that the pay
gap reflects differences in freely chosen labor market qualifications between the
sexes that legitimately affect relative wages. They assert that as the work
expectations, schooling, and labor market experience of women and men continue
to converge, the wage gap will narrow further in the absence of additional
intervention, which could itself impede economic efficiency.
This report examines the trend in the male-female wage gap and the
explanations offered for its existence. Remedies proposed for the gender wage gap’s
amelioration are addressed, with an in-depth focus on the comparable worth approach
to achieving “pay equity” or “fair pay” between women and men.
The Male-Female Wage Gap
The issue termed pay equity or more recently, fair pay, originates from the
chronic fact that women as a group are paid less than men. In 1960, half of all
women employed year-round full-time (i.e., 50-52 weeks, at least 35 hours per week)
earned more than $3,257 and half earned less than that amount; in the same year, the
median annual earnings of men employed year-round full-time were $5,368. More
than 4 decades later, according to U.S. Bureau of the Census data for 2003, the
median earnings of women with a strong commitment to the workforce were $30,724
while those of men were a substantially greater $40,668.
It is often noted that even when comparisons are made between similar groups,
women still earn less than men. Women with a bachelor’s degree employed yearround full-time earned $47,910 in 2003, while similarly educated men earned an
average of $69,913. Male high school graduates were paid $38,331 on average, well
above the $27,956 paid to female high school graduates. Women typically earn less
than men of the same age, as well. The wage gap tends to widen as age increases:
according to Census Bureau data for 2003, female 15-24 year olds were paid 79% as
much as male 15-24 year olds; female 25-44 year olds earned 67% as much as males
in the same age group; and, female 45-64 year olds were paid 59% as much as male
45-64 year olds. Although these disparities between seemingly similar groups of
men and women sometimes are taken as proof of sex-based wage discrimination,
they have not been adjusted to reflect gender differences in all characteristics — such
as differences in school course work among those with the same educational
attainment1 — that can legitimately affect relative wages.
The size of the male-female wage gap has shrunk at a slow and uneven pace
over the years. (See Table 1.) In the 1960s and 1970s, women employed year-round
full-time typically earned less than 60 cents for every dollar earned by men. The gap
narrowed steadily during the 1980s, so that by the end of the decade women were
being paid about 70 cents on the dollar. According to the data series on annual
earnings of year-round full-time workers (columns 2 and 5) and on weekly earnings
of full-time workers (columns 3 and 6), the ratio of female-to-male wages fluctuated
erratically during the 1990s, and the ratios derived from the two series at times
moved in different directions. The extent of improvement in the gender wage gap
that occurred during the 1980s does not appear to have been sustained during the
1990s. More recently, the trend seems largely positive. And, in 2003, women
typically earned 76-79 cents for every dollar earned by men. Despite substantial
changes in women’s labor force participation over the 40-plus year period, however,
the observed or unadjusted wage gap has narrowed by just 15 percentage points.
A portion of the wage gap between male and female college graduates is due to gender
differences in majors which, in turn, affect the occupations in which they subsequently
work. Charles Brown and Mary Corcoran, “Sex-Based Differences in School Content and
the Male-Female Wage Gap,” Journal of Economic Literature, vol. 15, no. 3, part 1, July
1997; and June O’Neill, “The Gender Gap in Wages, Circa 2000,” American Economic
Review, vol. 93, no. 2, May 2003 (hereafter cited as O’Neill, The Gender Gap in Wages,
Table 1. Ratio of Female-to-Male Earnings
Source: U.S. Bureau of the Census data on year-round/full-time workers, and U.S. Bureau of Labor
Statistics data on full-time workers.
Note: The wage gap based on annual data is wider than the wage gap based on weekly data because
women generally work fewer weeks and hours per week than men. In addition, the annual data include
self-employed workers who have larger earnings differences by gender than the wage and salary
workers covered by the weekly series. Regardless of the interval, the gender wage gap would be wider
if all workers were compared because relatively more women than men work part-time or part-year
n.a. = not available
a. Based on median annual earnings of all workers age 15 or older (14 or older before 1980)
employed year-round full-time (i.e., 50-52 weeks in a year and at least 35 hours in a week),
including the self-employed. Before 1989, earnings covered civilian workers only.
b. Based on median weekly earnings of wage and salary workers age 16 or older employed full-time.
Reasons Suggested for the Wage Gap
The persistence of the gender wage gap has led to a search for explanations.
Basically, two schools of thought have developed. The human capital explanation
has a supply-side focus, that is, it looks at the personal characteristics of working
women and men. The sex-segregation-in-the-workplace or discrimination
explanation has a demand-side focus, that is, it looks at the characteristics of the jobs
in which women and men typically work. The human capital model focuses on the
voluntary choices made by women; the discrimination model, on the restrictions
faced by women.
The Human Capital Explanation
One school of thought asserts that women earn less than men because of the
division of labor within the family which results in productivity differences between
the sexes.2 Anticipated family responsibilities are believed to influence women’s
decisions about the amount and kind of investment in education/training (e.g.,
number of years of schooling, subject matter of course work, and general versus firmspecific training) as well as the length and pattern of time devoted to market work
(i.e., total years and timing of work experience, intermittent/continuous participation,
and part-time/part-year or full-time/full-year schedules). Because women expect to
spend fewer years than men in the labor force, it is less profitable for women to
invest in market-oriented skills. According to the human capital explanation, then,
women’s smaller human capital investment lowers their productivity, and hence their
earnings, relative to men’s.
The discontinuous employment history of women is thought to further depress
their relative wages.3 Because skills deteriorate with prolonged non-use, women’s
wages upon reentering the labor force initially are lower than their wages had been
when they exited the labor force to bear and care for their children. It is argued that
women who anticipate moving in and out of the labor force choose to enter
occupations having the least earnings penalty for intermittent employment, that is,
occupations having the lowest rate of skill depreciation. Therefore, women more so
than men would prefer employment in jobs in which wage growth is not closely
Jacob Mincer and Solomon Polachek, “Family Investments in Human Capital: Earnings
of Women,” Journal of Political Economy, vol. 82, part 2, Mar./Apr. 1974; and, Jacob
Mincer and Solomon Polachek, “Women’s Earnings Re-examined,” Journal of Human
Resources, vol. 13, no. 1, winter 1978.
Solomon Polachek, “Occupational Segregation among Women: Theory, Evidence, and a
Prognosis,” in Cynthia B. Lloyd, Emily S. Andrews, and Curtis L. Gilroy (eds.) Women in the
Labor Market (New York: Columbia University Press, 1979); Solomon Polachek,
“Occupational Self-Selection: A Human Capital Approach to Sex Differences in Occupational
Structure,” Review of Economics and Statistics, vol. 63, no. 1, 1981; Jacob Mincer and Haim
Ofek, “Interrupted Work Careers: Depreciation and Restoration,” Journal of Human
Resources, vol. 17, no. 1, winter 1982; Moon-Kak Kim and Solomon Polachek, “Panel
Estimates of Male-Female Earnings Functions,” Journal of Human Resources, vol. 29, no. 2,
spring 1994; and Audrey Light and Manuelita Ureta, “Early Career Work Experience and
Gender Differentials,” Journal of Labor Economics, vol. 13, no. 1, Jan. 1995.
linked to skill accumulation generally and firm-specific training particularly. As a
result of this voluntary choice,4 women who do not expect to work continuously
would want jobs that offer a smaller reward for additional work experience (i.e., jobs
with a flatter experience-earnings profile) than would continuous workers. Human
capital theory thus not only attempts to explain the relatively lower wages of women,
but also the difference in the occupational distributions of men and women.
The division of labor within the family could have a more direct impact on
women’s wages than through its effect on human capital accumulation. Married
women continue to be largely responsible for child care and other energy-intensive
household responsibilities. Because they consequently expend more effort than men
on these family duties, it arguably reduces the effort that women can put into market
work. If wages and market work intensity are positively related, then married
women’s wages will be lower than those of married men — even for workers having
the same human capital endowments.5 Some analysts also believe that household
duties affect married women’s occupational preferences by encouraging them to seek
less-demanding jobs which allow them to economize on the energy they expend on
Researchers have tried to explain the existence of the wage gap by estimating
the proportion that is due to differences in the average amounts of human capital
accumulated by female and male workers. The residual or unexplained portion of the
pay differential is due to differences in the rates of return (reward) to working women
and men with the same productivity-related characteristics. Because productivity
rarely can be directly observed, commonly used proxies include amount and quality
of formal education, on-the-job training, hours of work, job tenure, and amount and
continuity of labor market experience.
The human capital explanation assumes that individuals’ preferences/choices are
voluntary. It does not consider the possibility of pre-labor market (societal) discrimination.
For example, social norms/customs may constrain women’s ideas about appropriate careers
(e.g., becoming a nurse rather than a doctor). For more information on the relationship
between socialization and occupational choice, see Margaret Mooney Marini and Mary C.
Brinton, “Sex Typing in Occupational Socialization,” in Barbara F. Reskin (ed.) Sex
Segregation in the Workplace: Trends, Explanations, Remedies (Washington, D.C.: National
Academy Press, 1984). In addition, the access of workers to firm-specific on-the-job
training may be constrained by the organizations that employ them; thus, the amount of
specific human capital accumulation may not entirely reflect individual preferences. For
more information on this point, see Don Tomaskovic-Devey and Sheryl Skaggs, “Sex
Segregation, Labor Process Organization, and Gender Earnings Inequality,” American
Journal of Sociology, vol. 108, no. 1, July 2002.
According to one analysis, gender differences in personal and job characteristics explain
27-30% of the wage gap when housework was not considered; when the gender difference
in housework was taken into account, about 38% of the wage gap could be explained. Joni
Hersch and Leslie S. Stratton, “Housework, Fixed Effects, and Wages of Married Workers,”
Journal of Human Resources, vol. 32, no. 2, spring 1997.
Gary S. Becker, “Human Capital, Effort, and the Sexual Division of Labor,” Journal of
Labor Economics, vol. 3, no. 1, Jan. 1985 supplement.
While, to some, the unexplained portion of the gender pay gap indicates the
existence of sex-based wage discrimination, to others it indicates the limits of
knowledge. For the residual to solely reflect discrimination — that is, all earnings
differences between the sexes not accounted for by differences in their human capital
attributes — then all relevant variables must be included in the empirical studies and
the included variables must be measured accurately. However, productivity
characteristics may be omitted from studies because they are difficult or impossible
to measure (e.g., motivation) or are not included in a particular data set (e.g., field of
specialization in school), and they may be imprecisely measured (e.g., use of
potential rather than actual work experience).
Based on its review of seven empirical studies, the National Academy of
Sciences found that less than one-half of the wage gap between the sexes could be
explained by human capital variables alone.7 While this finding lends credence to
some researchers questioning of the ability of the skill depreciation hypothesis to
account for the wage gap over the long-run and to explain the different occupational
distributions of women and men,8 it also reflects the inherent difficulty of accurately
measuring all productivity-related characteristics. In addition, the finding gives
support to claims that factors other than productivity affect wages, including the
presence and strength of unions, the industry of employment, and the size of firms.
The Discrimination Explanation
In contrast to human capital theorists, other researchers look to job-related
variables as justification for the existence of the wage gap. Some focus particularly
on the relationship between sex segregation in the workplace and women’s
comparatively low wages. Segregation encompasses the clustering of women and
men in different occupational groups, in different occupations within these larger
groups, in different jobs within occupations, and in different industries or firms
performing the same jobs.
Sex Segregation by Occupation. Gender integration of occupations would
be expected to somewhat narrow the pay gap, given the relatively greater penalty
women are estimated to experience from the low wage-high percent female
Donald J. Treiman and Heidi I. Hartmann (eds.) Women, Work, and Wages: Equal Pay
for Jobs of Equal Value (Washington, D.C.:, National Academy Press, 1981). (Hereafter
cited as Treiman and Hartman, Women, Work, and Wages.)
Paula England, “The Failure of Human Capital Theory to Explain Occupational Sex
Segregation,” Journal of Human Resources, vol. 17, no. 3, summer 1982; John M. Abowd
and Mark R. Killingsworth, “Sex, Discrimination, Atrophy, and the Male-Female Wage
Differential,” Industrial Relations, vol. 22, no. 3, fall 1983; Mary Corcoran, Greg J. Duncan,
and Michael Ponza, “Work Experience, Job Segregation, and Wages,” in Reskin, Sex
Segregation in the Workplace; and Allison J Wellington, “Changes in the Male/Female
Wage Gap, 1976-1985,” Journal of Human Resources, vol. 28, no. 2, spring 1993.
(Hereafter cited as Wellington, Changes in the Male/Female Wage Gap.)
relationship.9 And, indeed, the gradual narrowing of the gender pay gap has occurred
over a period when the occupational distributions of women and men slowly have
become more similar.
In the 1960s, occupational segregation of the sexes declined somewhat, due
partly to the movement of men into female-intensive jobs (e.g., elementary school
teachers, librarians, and social workers).10 In the 1970s, gender segregation
decreased to a greater extent, due primarily to both men and women entering neutral
occupations.11 Women also increased their presence in a few male-dominated
occupations during the 1970s, with the greatest inroads being made in the rapidly
expanding occupational groups of executive, administrative, and managerial workers
as well as professional workers. (The pattern of and changes in sex segregation by
occupation are sensitive to the percentages that researchers utilize to define femaledominated, male-dominated and neutral occupations.)12
The occupational distributions of women and men have kept converging over
the ensuing decades, but at a progressively slower pace.13 The Index of Segregation
is the share of women (men) who would have to change jobs for their occupational
employment distributions to be identical. The index stood at about two-thirds in each
census year through 1970. It then dropped substantially, from 67.7 in 1970 to 59.3
in 1980. The index (at 52.0) showed a further convergence by 1990 in the
occupational employment patterns of women and men. Based upon Current
Population Survey data, as opposed to decennial Census data, the index in 1990 was
56.4. Its level of 53.9 in 1997 reflected a continuing, albeit slowed, reduction in
occupational sex segregation.
(It should be noted that national data underlie the above-described trend in
occupational segregation. Most workers and organizations function in local labor
markets, and the degree of occupational segregation by gender may vary from one
area to another. Research suggests that the size of the wage penalty for working in
female-dominated jobs is related to the degree of occupational segregation that exists
in a given local labor market.14)
Stephanie Boraas and William M. Rodgers III, “How Does Gender Play a Role in the
Earnings Gap? An Update,” Monthly Labor Review, Mar. 2003.
Cynthia H. Chertos, Lois Haignere, and Ronnie J. Steinberg, Occupational Segregation
and Its Impact on Working Women (New York: Center for Women in Government, 1982).
Nancy F. Rytina and Suzanne M. Bianchi, “Occupational Reclassification and Changes
in Distribution by Gender,” Monthly Labor Review, Mar. 1984. Note: Neutral occupations
were defined as having between 21% and 59% female employment; male-dominated
occupations were defined as employing 20% or fewer women; and, female-dominated
occupations as employing 60% or more women.
Barbara H. Wooton, “Gender Differences in Occupational Employment,” Monthly Labor
Review, Apr. 1997.
Francine D. Blau and Lawrence M. Kahn, “Gender Differences in Pay,” Journal of
Economic Perspectives, vol. 14, no. 4, fall 2000.
Philip N. Cohen and Matt L. Huffman, “Occupational Segregation and the Devaluation
of Women’s Work Across U.S. Labor Markets,” Social Forces, vol. 81, no. 3, Mar. 2003;
Why Does Segregation Depress Women’s Relative Wages? The
crowding hypothesis offers one explanation of why occupational segregation is
associated with lower earnings for women relative to men.15 According to this
theory, women are excluded from many jobs due to the discriminatory tastes of
employers, male employees, or customers. Because women have access to only a
limited number of occupations, the supply of labor to those occupations increases
which, in turn, reduces the capital-to-labor ratio. As a result, the productivity and
wages of both women and men in the crowded, female-intensive occupations are
depressed. Similarly, productivity and wages in male-dominated occupations are
higher than they would otherwise be the case because the supply of labor to them is
restricted. Discrimination, it is asserted, prevents labor mobility between the two sets
of occupations which would equalize wages between male and female workers with
the same human capital endowments.
The dual labor market theory offers another explanation for the relationship
between occupational segregation and the wage gap.16 According to this model, the
labor market has both primary and secondary sectors. The former is comprised of
jobs having opportunities for advancement, offering good wages and working
conditions, and providing job security. The latter is comprised of dead-end jobs, with
low wages, poor working conditions, and substantial employee turnover. Because
primary sector jobs require firms to invest in employee training, wage differentials
can arise from this market segmentation. Primary sector employers will want to hire
stable workers so that they have some assurance of recouping their training costs.
Since women historically have tended to move in and out of the labor force more
often than men, risk-adverse employers who rely on their impressions of the relative
job turnover of all men and women when deciding about hiring individual men and
women (i.e., statistical discrimination) would prefer men over women for primary
sector jobs. The outcome, some analysts claim, is that women are relegated to the
secondary, low-paying sector of the labor market.
It is suggested that limiting women to poor job opportunities could make their
alleged high turnover a self-fulfilling prophecy because the jobs provide women
little reason to stay with any one employer.17 Labor market discrimination could have
other feedback effects “if it discourages women from making human capital
investments, weakens their attachment to the labor force, and provides economic
incentives for the family to place priority on the husband’s career.”18 As a
and Philip N. Cohen and Matt L. Huffman, “Individuals, Jobs, and Labor Markets: The
Devaluation of Women’s Work,” American Sociological Review, vol. 68, no. 3, June 2003.
Barbara R. Bergmann, “Occupational Segregation, Wages and Profit When Employers
Discriminate by Race or Sex,” Eastern Economic Journal, vol. 1, nos. 2 & 3, Apr. and July
Peter B. Doeringer and Michael J. Piore, Internal Labor Markets and Manpower Analysis
(Lexington, MA.: D.C. Heath and Company, 1971).
Barbara R. Bergmann, The Economic Emergence of Women, NY, Basic Books Inc., 1986.
(Hereafter cited as Bergmann, The Economic Emergence of Women.)
Francine D. Blau, Marianne A. Ferrer, and Anne E. Winkler, The Economics of Women,
consequence of discrimination’s feedback effects on human capital variables, then,
empirical studies that use these variables to explain the wage gap could understate
discrimination’s impact. Rather than the gender wage gap being due to either
women’s choices or women’s constraints, the explanation likely is a mutually
reinforcing combination of human capital differences between the sexes and
discrimination against working women.
Some Portion of the Wage Gap Remains Unexplained. Studies have
attributed the wage gap’s existence to characteristics of both workers and jobs.
Typical job characteristics include occupational prestige, supervisory status, union
status, industry, occupation, and percent female. In those studies that adjust for
differences in the employment distributions of men and women at the detailed
occupational level, more of the gap is explained than in studies that adjust across
major occupational groups.19 This was the case among the seven empirical analyses
reviewed by the National Academy of Sciences which took into account job as well
as worker characteristics: all but one of the studies that used detailed occupational
classifications could explain between 30% and 71% of the male-female pay
The results of empirical studies reviewed by the National Academy of Sciences
and other researchers indicate that the addition of job-related variables to human
capital variables increases the ability to account for differences between the wages
paid to women and men.21 In one study, for example, taking human capital variables
Men, and Work (Upper Saddle River, NJ: Prentice Hall, 1998), p. 214. (Hereafter cited as
Blau, Ferrer, and Winkler, The Economics of Women, Men, and Work.)
The greater explanatory power of studies that use more detailed occupational
classifications is due to the fact that employment and earnings differences between men and
women can be quite large even within occupational groups. For example, U.S. Bureau of
Labor Statistics’ data for 2003 show that within the major occupational group of
professionals, women were 82% of elementary and middle school teachers while men were
93% of electrical and electronics engineers; among full-time wage and salary workers,
female elementary and middle school teachers had median weekly earnings of $757 and
male electrical and electronics engineers had median weekly earnings of $1,348. These
intra-occupational differences are obscured in studies that use highly aggregated
Treiman and Hartmann, Women, Work, and Wages.
Other reviews of empirical work on the male-female wage gap are included in Cynthia B.
Lloyd and Beth T. Niemi, The Economics of Sex Differentials (New York: Columbia
University Press, 1979); Naresh C. Agarwal, “Pay Discrimination: Evidence, Policies, and
Issues,” in Harish C. Jain and Peter J. Sloane, Equal Employment: Race and Sex
Discrimination in the United States, Canada, and Britain (New York: Praeger Publishers,
1981); June O’Neill, “Earnings Differentials: Empirical Evidence and Causes,” in Gunther
Schmid and Renate Weitzel (eds.) Sex Discrimination and Equal Opportunity: The Labor
Market and Employment Policy (New York:, St. Martin’s Press, 1984); Steven L. Willborn,
A Comparable Worth Primer (Lexington, MA: D. C. Health and Company, 1986); Glen G.
Cain, “The Economic Analysis of Labor Market Discrimination: A Survey,” in Orley
Ashenfelter and Richard Layard (eds.), Handbook of Labor Economics, vol. 1 (Amsterdam:
into account left 67.1% of the gender pay gap unexplained while the addition of jobrelated variables reduced the unexplained portion to 38.0%.22 According to the
analysis, if women had the same human capital attributes as men, they would have
earned 80.5% as much as men, and if women had the same occupational, industrial,
and union characteristics as men, the adjusted ratio would have been 88.2%. Another
study that included both human capital and job-related variables was able to account
for somewhat more of the gender wage gap, but it still left about 20% of the pay
disparity unexplained.23 Generally, even those studies that incorporate both types of
variables leave a substantial portion of the wage gap unexplained.
One of the job characteristics frequently included in empirical analyses, such as
those described above, is occupation. However, to the extent that the difference in
occupational distributions of women and men
partly reflects employment discrimination or unequal occupational access ... then
clearly this gender difference cannot legitimately be used to help “explain” the
gender wage gap.24
In other words, the result of including occupation as an explanatory variable is
underestimation of labor market discrimination to the degree that gender differences
in occupational distributions reflect discrimination rather than personal preferences.
The feedback effect of discrimination on other variables commonly included in
empirical analyses (e.g., job tenure and unionization) also would lead to
understatement of discrimination’s impact on women’s relative wages.25
Because neither discrimination nor productivity can be measured directly, it is
likely that debate will continue on precisely how much of the wage gap is due to each
of them. By extension, disagreement also will continue about whether remedial
action is needed; and, if so, what kind of action.
Elsevier Science Publishers, 1986); Francine D. Blau and Marianne A. Ferber,
“Discrimination: Empirical Evidence from the United States,” American Economic Review,
May 1987; and, Elaine Sorenson, “The Wage Effects of Occupational Sex Composition:
A Review and New Findings,” in M. Anne Hill and Mark R. Killingsworth (eds.)
Comparable Worth: Analyses and Evidence (Ithaca, NY: ILR Press, 1989).
Blau, Ferber, and Winkler, The Economics of Women, Men, and Work.
U.S. General Accounting Office, Women’s Earnings: Work Patterns Explain Difference
Between Men’s and Women’s Earnings, GAO-04-35, Oct. 2003.
Michael P. Kidd and Michael Shannon, “Does the Level of Occupational Aggregation
Affect Estimates of the Gender Wage Gap?” Industrial and Labor Relations Review, vol.
49, no. 2, Jan. 1996.
Morley Gunderson, “Male-Female Wage Differentials and Policy Responses,” Journal
of Economic Literature, vol. 27, Mar. 1989. (Hereafter cited as Gunderson, Male-Female
Wage Differentials and Policy Responses.)
Reasons Suggested for the Trend in the Wage Gap
As shown in Table 1, the wage gap has narrowed gradually and sporadically
over the years. Some assert that the intransigence of discrimination, despite
enforcement of equal pay and employment legislation, accounts for the slow and
fitful erosion of the wage gap. Others conclude that it is related to, among other
things, the timing of market skill convergence between the sexes and to non-genderspecific changes in the labor market.
During much of the post-World War II period, the human capital endowments
of women entering the labor force worked against contraction in the wage gap.26
Working men’s educational attainment increased more than working women’s as a
greater number of less educated than more educated women joined the labor force
through the 1970s. In addition, the average work experience of employed women
was held down by the entrance into the labor force of many women who had little or
no prior experience. It is thus argued that the lack of skill convergence between
working women and men during the 1970s kept the wage gap fairly constant, despite
decreases in occupational and industrial segregation over the period.27
The 1980s: Rapid Improvement in Women’s Relative Wages
The situation changed in the 1980s, when the work experience and schooling
of women increased relative to men. About 25% of the 1% per year reduction in the
wage gap between 1976 and 1989 can be ascribed to the increase in the amount of
women’s work experience compared to men’s, and another 35%-40% to the rise in
the monetary return (reward) to an additional year of experience for women
compared to men.28 The relative gain in the level and reward to women’s educational
attainment accounted for a significant but smaller share of the wage gap’s
contraction.29 In terms of schooling level, the incidence of college graduates among
female labor force participants grew more than among male participants.30 While the
financial payoff for an additional year of schooling increased for all workers, the
James P. Smith and Michael P. Ward, Women’s Wages and Work in the Twentieth Century
(Santa Monica, CA: The Rand Corporation, 1984).
Fields, Judith and Edward N. Wolff, “The Decline in Sex Segregation and the Wage Gap,
1970-1980,” Journal of Human Resources, vol. 26, no. 4, fall 1991.
O’Neill, June and Solomon Polachek, “Why the Gender Gap in Wages Narrowed in the
1980s,” Journal of Labor Economics, vol. 11, no. 1, part 1, Jan. 1993. (Hereafter cited as
O’Neill and Polachek, Why the Gender Gap in Wages Narrowed in the 1980s.)
Results from another study also suggest that the primary factor in the wage gap’s
narrowing was increases in women’s average levels of work experience rather than of
educational attainment. Wellington, Changes in the Male/Female Wage Gap, 1976-1985.
Between 1992 and 2002, for example, the share of working women aged 25-64 with
bachelor’s degrees rose by almost 7 percentage points (from 25.0% to 31.5%) while the
share of comparably aged working men with bachelor’s degrees rose by a lesser 4
percentage points (from 27.5% to 31.7%). U.S. Bureau of Labor Statistics, Women in the
Labor Force: A Databook, Report 973, Feb. 2004.
greater increase among women may be related to changes in their college schooling
Another 20% of the decrease in the wage gap can be attributed to the decline in
wages of blue-collar workers — a predominantly male occupational group —
compared to white-collar workers.32 Thus, some non-gender-specific factors changed
the labor market in different ways for men and women during the 1980s. For
example, decreases in unionization and employment shifts by industry (e.g., away
from manufacturing) depressed men’s wages more than women’s, thereby narrowing
the earnings differential between the two.33 Increasingly unfavorable changes in the
wage structure for low-paid workers over the period also seemingly harmed men
more than women, with greater widening of male wage inequality and growing
returns to intellectual vis-a-vis physical skill (strength). One economist has argued
that intellectual skills are less equally distributed by gender than physical skills, and
as a result, growth in the relative value of intellectual skills will increase wage
inequality among men while simultaneously increasing the relative earnings of
women (who are assumed to be more intellectually than physically endowed
compared to men).34 Others similarly have advanced the notion that an inverse
relationship exists between trends in wage inequality among men and in the pay gap
between the sexes.35
The unexplained portion of the wage gap decreased during the 1980s as well.
Interpretations of this finding are ambiguous, however: the improvement in women’s
unmeasured labor market attributes or a decline in discrimination, or some
combination of the two, could have reduced the residual.36 As women’s relative level
of measured productivity characteristics increased over the decade (e.g., work
experience and occupational upgrading), it is possible that unaccounted for personal
characteristics (e.g., work expectations and job motivation) also improved among
women compared to men. As a result of women’s increased commitment to the labor
force and enhancement of their job skills during the 1980s, it is possible that
employers’ rationale for statistical discrimination may have diminished as well.
In a study of young workers with college educations, it was estimated that changes in
college major and grades accounted for much of the decrease in the group’s gender wage
gap. While changes in the distribution of college majors among women contributed to the
narrowing of the pay differential, more of the narrowing may be explained by increasing
monetary rewards to women’s skills compared to men’s skills within a given major.
(College grades was the measure of skills used in the study.) Linda Datcher Loury, “The
Gender Earnings Gap Among College-Educated Workers,” Industrial and Labor Relations
Review, vol. 50, no. 4, July 1997.
O’Neill and Polachek, Why the Gender Pay Gap in Wages Narrowed in the 1980s.
Francine Blau and Lawrence M. Kahn, “Swimming Upstream: Trends in the Gender
Wage Differential in the 1980s,” Journal of Labor Economics, vol. 15, no. 1, part 1, Jan.
Finis Welch, “Growth in Women’s Relative Wages and in Inequality Among Men: One
Phenomenon or Two?” American Economic Review, vol. 90, no. 2, May 2000.
Nicole M. Fortin and Thomas Lemieux, “Are Women’s Wage Gains Men’s Losses? A
Distributional Test,” American Economic Review, vol. 90, no. 2, May 2000.
Blau, Trends in the Well-Being of American Women, 1970-1995.
The 1990s: Slowed Relative Pay Gains for Women
Although the adjusted wage gap narrowed further during the 1990s, it did so to
a lesser extent that during the prior decade. According to one study, when human
capital variables alone are taken into account, the ratio of women’s compared to
men’s wages rose from 70.8% in 1979 to 81.9% in 1989, and then remained basically
unchanged in 1998 (at 81.2%). When the specification was expanded to include jobrelated factors, the ratios were higher but the trend was the same: 81.6% in 1979, and
91.0% in both 1989 and 1998.37
Differences in work experience by gender decreased less in the 1990s than in
the 1980s, but it appears that relative increases in women’s educational attainment
played a much larger role in shrinking the pay gap during the 1990s than during the
preceding decade. Consequently, work experience and schooling were estimated to
have had a similar impact over the 20-year period when considered jointly.
More favorable changes for women in both occupational upgrading and
deunionization occurred in the 1980s compared to the 1990s. Women moved into
managerial and professional occupations and out of clerical and service occupations
at about the same rate in both decades. Women’s representation in blue-collar
occupations did not expand to the same degree in the 1990s, however, as men more
often lost/left those jobs during the 1980s. Similarly, gender differences in the pace
of deunionization were larger in the 1980s than in the following decade.
But, unexplained gender differences in wages appear to have been the main
contributor to the slowed convergence in women’s and men’s earnings during the
1990s as against the 1980s. One unmeasured characteristic that might have
contributed to the slowdown is hours spent performing housework: as previously
discussed, it has been argued that housework decreases the amount of effort put into
one’s job; and, the gender difference in this measure decreased much more quickly
in the 1980s than the 1990s. If employers perceived that women had made a greater
commitment to the labor force, more in the former than the latter decade, it could
have prompted a reduction in statistical discrimination against women, again, more
substantially in the 1980s as compared to the 1990s. In addition, the occupational
upgrading women experienced in the 1980s put more of them in positions that might
have been subject to “the glass ceiling,” which could have impeded their wage
advancement in the following decade. It has further been suggested that shifts in
labor demand associated with, for example, technological change (e.g.,
computerization) were considerably more beneficial for women in the 1980s than in
Francine D. Blau and Lawrence M. Kahn, The U.S. Gender Pay Gap in the 1990s:
Slowing Convergence, NBER Working Paper 10853, Oct. 2004.
Policy responses to the gender wage gap depend upon where the roots of the pay
differential are thought to lie. Their effectiveness in raising the wages of women
relative to men depends, in part, upon how much of the gap really is associated with
each causal factor.
The explanations and remedies presented below are not necessarily mutually
exclusive. For example, both socialization and employment discrimination might
affect women’s occupational distributions and wages; and consequently, multiple
remedies would be appropriate.
Explanations and Remedies
Some think that even before women enter the labor market their career
aspirations are shaped by societal factors which individuals, such as parents or peers,
and institutions, such as schools or the media, inculcate in youngsters.38 Proponents
of this viewpoint would support measures directed at schools (e.g., Title IX of the
1972 Education Amendments, 1976 Vocational Education Amendment, and
Women’s Educational Equity Act) and at acquainting girls with role models
employed in a broad range of jobs (e.g., Ms. Foundation for Women initiated “Take
Our Daughters to Work Day” in 1995).
Others conclude that the wage gap is due to measured and unmeasured human
capital differences between the sexes. They would argue that additional government
intervention in the labor market is unnecessary because ongoing relative
improvements in women’s productivity characteristics will cause the wage gap to
shrink further. With women continuing to opt for market work as a major lifetime
activity, human capital theory suggests that women will increasingly make
occupational choices that differ from those they made in the past which should, in
turn, raise their relative wages. But, as nondiscriminatory (legitimate) reasons for the
gender pay disparity “are unlikely to change radically in the near future unless the
roles of women and men in the home become more nearly identical ... an unadjusted
gender gap may be with us for quite a while.”39
Still others believe that labor market discrimination, in part manifested through
sex segregation in the workplace, is responsible for the portion of the wage gap that
cannot be explained by gender differences in productivity-related characteristics.
Policy responses that these observers would support range from improved
enforcement of antidiscrimination laws and regulations by the Equal Employment
Opportunity Commission (EEOC) and the Office of Federal Contract Compliance
Programs,40 to government dissemination of information about and provision of
This has been referred to as societal or pre-labor market discrimination.
O’Neill, The Gender Gap in Wages, Circa 2000, p. 314.
One study (Kimberly Bayard, Judith Hellerstein, David Neumark, and Kenneth Troske,
“New Evidence on Sex Segregation and Sex Differences in Wages from Match EmployeeEmployer Data,” Journal of Labor Economics, vol. 21, no. 4, Oct. 2003) estimated that
training in comparatively high-paying nontraditional jobs for women,41 to employers
paying their female and male employees in comparable jobs the same wages.
The last approach has commonly been referred to as comparable worth.
Comparable worth would extend the current mandate that an employer pay equal
wages to workers in the same or substantially same jobs in a firm, to an employer
paying equal wages to workers in equivalent jobs in a firm. Equivalent jobs typically
have been defined as those whose performance involves commensurate skill (e.g.,
education and training requirements), effort, responsibility, and working conditions.
The idea motivating comparable worth is that the size of a worker’s paycheck
should be related to job content and not related to the predominant sex of employees
in an occupation. Comparable worth proponents argue that some jobs are
undervalued — that is, pay relatively low wages — because they are largely held by
women. Some also believe that “the kinds of skills traditionally exercised by women
[e.g., nurturing social skills] are valued less in wage determination than are
traditionally male skills [e.g., physical or supervisory skills].”42 Supporters of this
viewpoint thus have concluded that market-set wages are tainted by gender-based
Rather than continuing to rely on supply and demand conditions in the labor
market to determine wages, comparable worth advocates have proposed that a single
job evaluation study of all jobs or key positions in a firm be conducted so they can
be compared with each other in terms of such attributes as skill, effort, responsibility,
and working conditions. Employers would then raise the wages of workers in all jobs
or in female-dominated jobs deemed to be underpaid on the basis of the evaluation
(i.e., jobs having wages below other jobs with the same total scores on the attributes
included in the evaluation).
The comparable worth approach to wage determination has made the most
headway in state and local governments. Public sector unions often have played a
large role in organizations making comparable worth pay adjustments.
about half of the gender wage gap is due to women having lower wages than men in
narrowly defined occupations in the same firms, which suggests that enforcement of the
Equal Pay Act might be effective in raising women’s relative earnings.
In terms of nontraditional employment policy, P.L. 102-530 (Women in Apprenticeship
and Nontraditional Occupations Act, WANTO) provides technical assistance to employers
and labor unions to promote women’s employment in apprenticeable and other
nontraditional occupations. P.L. 102-235 (Nontraditional Employment for Women Act),
which was in effect through FY1995, authorized the use of the Job Training Partnership
Act’s discretionary funds to develop demonstration programs to help women enter highpaying occupations where they were underrepresented.
Paula England, Comparable Worth: Theories and Evidence (New York: Walter de
Gruyter, 1992), p. 40. (Hereafter cited as England, Comparable Worth.)
The usual function of unions to bargain with management over wages has
prompted some to suggest that the collective bargaining process may be a more
effective and practical means of narrowing the gender pay gap than relying on
legislative or judicial action.43 However, unions might be reluctant to press for
comparable worth pay raises if they thought it would jeopardize membership
solidarity. This could occur, for example, if workers in jobs determined to be
overpaid through an evaluation were downgraded (i.e., had their pay cut) or redcircled (i.e., had their wage frozen at its current level or raised at a slower rate than
that of underpaid workers). In addition, private sector unions have shown much less
interest in the issue than have public sector unions. And, only about 15% of all
workers are represented by labor unions.
Congressional Action. Comparable worth was considered by the 98th, 99th,
and 100th Congresses. The primary focus of legislation introduced during the 1980s
was the wage-setting practices of the federal government. While there was
considerable debate on the issue, no legislation was enacted. (See Appendix for a
legislative history covering this period.)
Comparable worth as a remedy for gender pay differentials continued to
generate a great deal of controversy in the 1990s. When the 102nd Congress
subsequently took up civil rights legislation (P.L. 102-166, the Civil Rights Act of
1991), the version considered by the House (H.R. 1) aroused dissent partly because
some Members believed Section 102 (Pay Equity Technical Assistance) to be a
stalking horse for eventually allowing comparable worth claims under Title VII of
the Civil Rights Act of 1964, that is, allowing a job evaluation’s finding of unequal
pay for equally rated female- and male-dominated jobs to be used as evidence in a
trial to prove the presence of sex-based wage discrimination.44
The decade also marked a shift in the focus of comparable worth, which
continues today.45 Rather than seeking to apply the approach to the wage-setting
practices of the federal government alone, proponents have tried to extend
comparable worth to all employers subject to the Fair Labor Standards Act (FLSA)
— predominantly firms in the private sector. The Fair Pay Act, which has been
introduced in each Congress since the 103rd, would require that employees within an
FLSA-covered firm who work in dissimilar but equivalent jobs — when viewed as
a composite of skill, effort, responsibility, and working conditions — be paid equal
wages regardless of sex, race, or national origin. The bill thus would expand the
standard in the Equal Pay Act of 1963 (EPA, an amendment to the FLSA), which
makes it illegal for covered employers to pay different wages to men and women who
Marvin J. Levine, “Comparable Worth in the 1980s: Will Collective Bargaining Supplant
Legislative Initiatives and Judicial Interpretations?” Labor Law Journal, vol. 38, no. 6, June
1987; and Robert H. Cohen, “Pay Equity: A Child of the 80s Grows Up,” Fordham Law
Review, vol. 63, Mar. 1995.
For a discussion of comparable worth claims in the courts see Sandra J. Libeson,
“Reviving the Comparable Worth Debate in the United States: A Look Toward the
European Community,” Comparative Labor Law Journal, spring 1995.
For current legislative activity, see CRS Report RL30902, Pay Equity Legislation in the
108th Congress, by Charles V. Dale and Linda Levine.
hold jobs that require substantially the same skill, effort, and responsibility and that
are undertaken in similar working conditions. Acceptable wage differentials between
equivalent jobs would continue to be those based on a seniority or merit system, or
a system linking pay to the quantity or quality of output. However, the bill would no
longer allow “any factor other than sex” to be a legitimate reason for wage
differences as it currently is under the EPA. The EPA prohibition against achieving
its equal pay standard by reducing the wage of any employee would remain in force.
In contrast, the Paycheck Fairness Act, which also has been introduced in several
Congresses, would have the U.S. Department of Labor issue job evaluation
guidelines based on objective criteria (e.g., education, skill, and decision-making
responsibility) for voluntary use by employers.
Occupational Desegregation as a Remedy to the Wage Gap? The
comparable worth approach to wage determination is endorsed by those who believe
that the prohibition of gender discrimination in the Equal Pay Act of 1963, Title VII
of the Civil Rights Act of 1964, and E.O. 11246 is not sufficient to reduce the malefemale pay differential because of the intransigence of sex segregation in the
workplace. They point out that, even if one thought workplace integration were an
adequate remedy to the wage gap, it would take an extremely long time to achieve
pay equity given the very slowly growing similarity in women’s and men’s
occupational distributions. (See earlier discussion in this report on this point.)
Comparable worth adherents further observe that occupational desegregation as
a remediation approach has its own drawbacks.
Since the influx of women continues to be largest in those occupations where it
began, some initially male jobs have now “tipped” and become
disproportionately female. Thus, further increases in the proportion of women
in these jobs increase rather than decrease segregation.46
Because occupational desegregation and job desegregation are not necessarily the
same, proponents also contend that the increased occupational similarities between
the sexes may overstate integration’s potential impact on relative wages. Within the
occupation of bus drivers, for example, women tend to hold part-time low-paying
jobs driving school buses while men tend to hold relatively high-paying jobs driving
buses for city transit systems. It thus is asserted that the growing presence of women
in nontraditional occupations may have unexpected outcomes — namely,
resegregation as well as “ghettoization,” with women and men having the same
occupational titles but working in different specialties — which limit its ability to
reduce the wage gap.47
Comparable worth advocates point out that occupational desegregation does not
address directly the low wages of workers in female-dominated occupations.48 They
question whether it is reasonable to expect numerous women in the middle or near
England, Comparable Worth, p. 16.
Barbara F. Reskin and Patricia A. Roos, Job Queues, Gender Queues: Explaining Women’s
Inroads into Male Occupations (Philadelphia, PA: Temple University Press, 1990).
Bergmann, The Economic Emergence of Women.
the end of their worklives to pay twice: after having paid the first time through lower
wages induced by occupational crowding, they would have to pay a second time
through costs associated with changing jobs (i.e., forgone wages and benefits like
pensions as well as educational/retraining expenses). A comparable worth policy
would raise the wages of workers in traditionally female occupations who are unable
or unwilling to change jobs.
Critics of comparable worth respond that the wage gap has in fact been
narrowing because the productivity-related characteristics of working women have
come to more closely resemble those of men. Some observers claim that
employment discrimination rather than wage discrimination accounts for women’s
depressed earnings: if more nontraditional jobs were open to women, the wage gap
would shrink (1) because female entrants to these occupations would receive higher
pay and (2) because the supply of labor to traditionally female jobs would decrease
thereby raising the wages of workers in those positions. According to both these
perspectives, comparable worth is not directed at the real issues — namely, gender
differences in human capital qualifications, and discrimination in hiring and
promotions. While comparable worth opponents consequently might support more
active enforcement of current antidiscrimination laws and regulations, they see no
need for further government intervention in the market’s wage-setting process.
Job Evaluation as a Wage-Setting Mechanism. Supporters of
comparable worth believe that discrimination results in a misallocation of labor by
steering equally productive women and men into different jobs and paying women
relatively lower wages. Firms operating in such a labor market produce more of the
goods/services that use “underpaid” female workers and less of the goods/services
that use “overpaid” male workers than they would if the market were free of gender
bias. Comparable worth advocates therefore reason that action is needed
to correct the market wage to remove the distortions caused by discrimination,
[that is, to] bring wages closer to what they would be in a non-discriminating
market, thus increasing efficiency and productivity.49
According to this viewpoint, it is discrimination rather than comparable worth policy
that interferes with the optimal functioning of the labor market. Job evaluation is the
mechanism that comparable worth advocates usually propose to use for both
redefining the status hierarchy of jobs as it relates to relative wages50 and for
adjusting market wages.
Janice Peterson, “The Challenge of Comparable Worth: An Institutionalist View,”
Journal of Economic Issues, vol. 24, no. 2, June 1990, p. 607.
It has been suggested that the amount of controversy over comparable worth relates to its
being about more than wages. In the view of some individuals, comparable worth
“challenges basic cultural assumptions about the relative value of the activities of different
groups in society.” It also “would redistribute not only economic resources, but also labor
market power to women workers.” Ronnie J. Steinberg, “A Want of Harmony: Perspectives
on Wage Discrimination and Comparable Worth,” pp. 24-25, in Helen Remick (ed.)
Comparable Worth and Wage Discrimination: Technical Possibilities and Political
Realities (Philadelphia, PA: Temple University Press, 1984).
This is the fundamental problem with comparable worth according to its critics,
namely, the substitution of job evaluation for market conditions in the wagedetermination process.
Wages differ not only because different jobs require different levels of skill,
effort, or responsibility, but also because of the inherent scarcity of certain skills
and talents relative to society’s demand for such skills and talents.51
Some jobs may have a high “value-in-use” as measured by the inputs of skill, effort,
and responsibility, but they may not warrant a high wage (i.e., “value-in-exchange”
for labor services) if there is a plentiful supply of workers to perform them.52
Because of varying occupational supply/demand conditions, legitimate (nondiscriminatory) pay differentials can exist between jobs that evaluations conclude are
of equal value to employers.
Moreover, opponents assert that a requirement to equalize wages of jobs having
the same total scores on evaluations would not allow the market to efficiently allocate
labor across occupations. If sewing machine operators’ wages were raised based on
job evaluation findings, for example, the higher wages would prompt more workers
than otherwise to enter the occupation. Because the demand for sewing machine
operators did not change, however, firms would be unwilling to hire these new
entrants and unemployment would increase. The wage determined by job evaluation
would have sent an incorrect signal to workers, with misallocation of labor resources
the result. Further, if demand for sewing machine operators were to subsequently
increase, comparable worth would prevent firms from using higher wages to
effectively signal this to workers: if they wanted to offer higher wages to attract
more sewing machine operators, firms would also have to offer higher wages for
other jobs found to be of equal and higher value — even though there was an ample
supply of labor at the lower wage for the other jobs.
These inefficiencies, and the likelihood of different firms’ job evaluations
producing varying ratings for the same occupations, have led some critics to predict
additional government intervention once comparable worth becomes the law of the
land.53 It is suggested that one form of federal intervention might be the development
and administration of a national job evaluation system because the government would
be the only entity capable of reconciling (possibly discriminatory) differences
between firms in their evaluations’ results. Some sort of regulatory or adjudicatory
process could well be needed to, at the least, resolve disputes over the specifics of
individual firms’ evaluations and to decide when evaluations should be revised to
reflect changes over time in job content. Another form of intervention might be
John Raisian, Michael P. Ward, and Finish Welch, “Pay Equity and Comparable Worth,”
Contemporary Policy Issues, Apr. 1986.
Gunderson, Male-Female Wage Differentials and Policy Responses.
Michael Evan Gold. A Dialogue on Comparable Worth (Ithaca, NY: ILR Press, 1983);
and Walter Y. Oi, “Neglected Women and Other Implications of Comparable Worth,”
Contemporary Policy Issues, vol. 4, Apr. 1986. For a discussion of the varying outcomes
of different commercial job evaluations at one firm see E. Jane Arnault, Louis Gordon,
Douglas H. Joines, and G. Michael Phillips, “An Experimental Study of Job Evaluation and
Comparable Worth,” Industrial and Labor Relations Review, vol. 54, no. 4, July 2001.
government in the role of labor allocator: because wages could no longer be used to
attract workers to specific jobs under a comparable worth system, the government
might have to encourage or direct people into occupations in order to resolve
imbalances between supply and demand.
Comparable worth proponents respond that they have not endorsed a national
job evaluation system, and that it is not unusual for some economists and business
groups to make dire predictions about the effects of legislation which would increase
government regulation of the labor market. In addition, supporters claim that a
comparable worth policy is not intended to measure the inherent value of jobs, but
rather to eliminate the wage disparities between female-dominated and maledominated jobs that cannot be justified by differences in productivity requirements.
As to the use of job evaluations in setting wages, comparable worth adherents
contend that many firms have long used them and that wages set by market
conditions actually are incorporated in job evaluations.
Depending on the type of system used, market wages have varying degrees of
influence on the weights (i.e., relative importance) accorded the job attributes in
evaluations. In fact, this is one of the problems with relying on job evaluations to
achieve comparable worth according to its supporters. They find this drawback to
be especially egregious at firms which conduct separate evaluations of female- and
male-dominated job families (e.g., clerical workers and blue-collar workers,
respectively). It is claimed that the use of multiple pay plans within a firm “is the
single most important reason that the common practice of job evaluation has done
little to close the sex gap in pay.”54 For this reason, comparable worth advocates not
only call for unbiased job evaluations but also for the same evaluation system to be
applied across all job families within an establishment.
The job evaluation tool also has been criticized for its complexity and
subjectivity.55 Even before an evaluation is begun, definitions of a firm (e.g., an
establishment within a multi-establishment enterprise with(out) regard to proximity
of other locations) and of female- as well as male-dominated jobs (e.g., composed
60% or 70% by one sex) must be made. After developing job descriptions for all
positions within a firm, the evaluators must select the job attributes or compensable
factors to be analyzed. Then, for each job included in the analysis, they must assign
a numerical value to each of the chosen attributes. In order to develop a total score
for each job, the evaluators also must determine weights for the individual factors.
For example, they must decide whether a job’s skill component is more important
(i.e., given more weight) than a job’s supervisory component. After having gone
through these several steps, a final decision must be made about how to relate scores
to wages and whose wages should be adjusted (e.g., should the wages of all
underpaid jobs be raised or just underpaid female-dominated jobs; should the wages
of overpaid jobs be reduced, frozen, or given smaller increases; and, what level
should wages be raised to — the average of all jobs or male-dominated jobs with the
same total points, or some proportion of the difference between female- and maledominated jobs with the same total points). Comparable worth opponents argue that
England, Comparable Worth, p. 219.
Raisian, Ward, and Welch, Pay Equity and Comparable Worth.
the decisions required at each stage of this complex procedure provide an opportunity
for the biases of evaluators and others to affect the outcome.
Supporters of comparable worth acknowledge the possibility of bias by noting
that job evaluations as typically conducted often have favored “men’s work” over
“women’s work.” Sources of bias against women in job evaluations include writing
job descriptions that omit some compensable factors typical of women’s jobs, giving
lower ratings to women’s jobs on the factors that are included, weighting factors
typical of men’s jobs (e.g., unsafe working environment) more heavily than those
typical of women’s jobs (e.g., finger dexterity), and using separate evaluations of job
families predominated by men and by women. Because of these practices,
comparable worth proponents want bias-free evaluations to be conducted. The
policy’s detractors suggest, however, that what actually could occur is the
substitution of one set of biases for another.
The Economic Effects of Implementing Comparable Worth. Critics
of comparable worth expect that its efforts to raise wages in female-dominated
occupations could harm women through disemployment. If employers must raise
wages of female-dominated jobs deemed to be underpaid by job evaluations,
employment in those occupations will decrease unless demand is totally inelastic
(i.e., demand does not change in response to a change in the wage). The higher labor
costs resulting from comparable worth pay raises could lead firms to replace workers
in previously underpaid positions with now cheaper labor-saving technology. In
addition, the resultant increase in production costs could prompt a decrease in output
and total employment, or an increase in prices to consumers. The outcome of
implementing comparable worth under this scenario would be both reduced
employment in traditionally female jobs and higher wages for the women still
employed in those fields.
Opponents of comparable worth also expect its implementation would cause the
wages of some women to fall. This is because some sectors or groups (e.g., small
firms or part-time workers) often are not covered when a law is enacted.56
Specifically with regard to a comparable worth policy, it is more likely that medium
and large firms rather than small firms (e.g., those with fewer than 100 employees)
would be covered because numerous different job titles are needed to conduct a job
evaluation and because the cost of using job evaluations decreases as firm size
increases.57 Women who either were displaced from or were unable to get jobs in the
part of the economy covered by comparable worth legislation would increase the
supply of labor in the uncovered sector, thereby putting downward pressure on their
own wages as well as the wages of other uncovered workers with whom they would
most directly compete (e.g., other women).
For example, the comparable worth policy in Ontario, Canada exempts private sector
firms with fewer than 10 employees. Among covered firms, those with fewer than 100
employees are treated somewhat differently than larger firms. U.S. General Accounting
Office, Pay Equity: Experiences of Canada and the Province of Ontario, GGD-94-27BR.
Richard S. Smith, “Comparable Worth: Limited Coverage and the Exacerbation of
Inequality,” Industrial and Labor Relations Review, vol. 41, no. 2, Jan. 1988.
The results of one analysis suggest that, if comparable worth were implemented
with a small firm exemption, it could well help higher-paid women (more often
employed by large firms) and harm lower-paid women (more often employed in
small firms). Moreover, it appears that “occupational segregation is highest in the
sector least likely to be the target of comparable worth policies and lowest in the
sector most likely to be covered.”58
Comparable worth advocates concede that there could be trade-offs for elevating
wages in traditionally female occupations. But, as this is true of many policy
initiatives, they do not consider it an adequate reason for maintaining the status quo.
How much comparable worth actually would cost depends, in part, on “how many
workers are displaced, how quickly, and what happens to them. At present, we
simply do not know how severe these problems would be.”59
Some believe that indirect benefits might flow from comparable worth’s direct
costs. Women no longer able to get traditionally female jobs after the upward
adjustment of the jobs’ wages might “leave the safe harbor of the female occupations
and consider competing for jobs heretofore filled with males,”60 while men might be
more attracted to the now higher paying, traditionally female jobs. The
implementation of comparable worth might thus encourage occupational
Comparable worth also might raise women’s aggregate
compensation as they expand their presence in traditionally higher paid, maledominated jobs and in now relatively better paid, female-dominated jobs.61
Comparable Worth’s Potential Impact on the Wage Gap. Comparable
worth is intended to eliminate pay differences between jobs within a firm that are
related to an occupation’s sex composition. Given its intrafirm focus, the policy will
not affect the portion of the gender wage gap that is due to other manifestations of
workplace segregation, including differences between the firms (e.g., large versus
small) or industries (e.g., manufacturing versus services) in which women and men
typically work.62 In order to estimate the potential effect on the wage gap of enacting
comparable worth legislation, then, studies should control for job characteristics as
well as measures of productivity in order to separate their impact from that of
Generally, the more variables included in empirical analyses of the causes of the
wage gap, the smaller the portion related to an occupation’s gender composition.
Ibid., p. 237.
Blau, Ferber, and Winkler, The Economics of Women, Men, and Work, p. 225.
Barbara R. Bergmann, “Does the Market for Women’s Labor Need Fixing?” Journal of
Economic Perspectives, vol. 3, no. 1, winter 1989, p. 57.
Peter F. Orazem and J. Peter Mattila, “Male-Female Supply to State Government Jobs and
Comparable Worth,” Journal of Labor Economics, vol. 15, no. 1, Jan. 1998.
According to one estimate (Judith Fields and Edward N. Wolff, “Interindustry Wage
Differentials and the Gender Wage Gap,” Industrial and Labor Relations Review, vol. 49,
no. 1, Oct. 1995), 31%-38% of the observed wage gap may be explained by differences in
women’s and men’s employment distribution by industry and by differences in their pattern
of interindustry wage differentials.
Studies that use different units of analysis (e.g., occupations or individuals) also
estimate varying magnitudes for the “wage penalty” caused by working in femaledominated rather than male-dominated occupations. Analyses that take into account
the industry of employment produce smaller estimates of the effect of occupational
segregation on the wage gap than do other studies. Results also can vary because of
differences in time periods analyzed. For these reasons, studies have produced a
wide range of estimates for the effect of an occupation’s gender composition on the
wage gap (0% to 42%, according to one review).63 Given the breadth of this range,
empirical research conducted to date provides little guidance as to the potential
impact of a comparable worth policy on the gender wage gap.
A more definitive answer about the potential magnitude of comparable worth’s
impact on the wage gap also is lacking based on studies of state and local
governments’ implementation of the policy.64 Analyses conducted before the policy
went into effect in five state governments and in San Jose, California predicted that
women’s relative wages would rise by 15%. The researchers assumed that, when
implemented, the policy would increase the pay in female-dominated jobs to the pay
in male-dominated jobs deemed equivalent through job evaluation. In reality,
however, most of the governments included in the two studies raised the wages of all
underpaid jobs which likely made the actual increase in women’s relative wages less
than the predicted increase.
Other studies examined the effect of comparable worth on the wage gap after
the policy had been implemented.65 While Iowa’s comparable worth policy also
called for raising the wages of all (not just female-dominated) underpaid jobs, it
originally called for pay cuts as well. In light of union objections, however, a
compromise was reached which gave smaller wage increases to underpaid jobs in
return for no pay cuts. A study of the actual experience with comparable worth in
Iowa found, not surprisingly, that women’s relative pay rose by just 1.4%.66 In
Minnesota, by contrast, most of the comparable worth pay adjustments went to
women on the state’s payroll and particularly to workers in female-dominated jobs
(mainly women, by definition). One study estimated that the state government’s
comparable worth policy was responsible for almost a 12% increase in women’s
relative pay.67 Thus, the manner in which comparable worth is implemented can
Elaine Sorensen, Comparable Worth: Is It a Worthy Policy? (Princeton, NJ: Princeton
University Press, 1994). (Hereafter cited as Sorensen, Comparable Worth: Is It a Worthy
One study estimated that the degree of wage discrimination against female-dominated jobs
in Iowa state government was greatly overstated, however, due to measurement error in job
evaluation. Shih-Neng Chen, Peter F. Orazem, J. Peter Mattila, and Jeffrey J. Greig,
“Measurement Error in Job Evaluation and the Gender Wage Gap,” Economic Inquiry, vol.
37, no. 2, Apr. 1999.
In terms of comparable worth’s employment effect in Minnesota, it might have resulted
in somewhat reduced job growth: while women on the state’s payroll actually increased by
17.2% between 1981 and 1988, their numbers would have increased by 20% in the absence
substantially influence the policy’s actual effectiveness in narrowing the gender wage
Misgivings have arisen about how comparable worth has been implemented in
some instances, because of the influence this can have on the achievement of pay
equity. The “pay for points” approach would, in theory, adjust the wages of all jobs
based on their job evaluation score, with overpaid jobs having their pay cut and
underpaid jobs having their pay raised. Because of objections from (representatives
of) workers whose pay is to be cut, however, this approach typically is compromised
such that underpaid jobs receive smaller wage increases than they otherwise would
have and no one’s pay is reduced.
This compromise “pay for points” approach has three serious weaknesses: (1)
it does not achieve the basic purpose of comparable worth — to eliminate the
underpayment of “women’s work”; (2) it does not target pay adjustments to
female-dominated jobs; and (3) it is overly dependent on the job evaluation
system. This approach cannot achieve equal pay for comparable worth because
it relies on pay cuts to achieve this aim, but these cuts are never enacted. By
targeting all underpaid jobs for pay adjustments, this approach increases the cost
of comparable worth and undercuts the gains to female workers. Finally, it relies
primarily on the job evaluation system to determine salaries, a system that is
known to be subjective and arbitrary. A comparable worth policy only needs to
eliminate the variation in wages that is negatively correlated with the
“femaleness” of a job once job requirements are taken into account. This
approach tries to eliminate all wage variation once job requirements are
It appears that political considerations can erode comparable worth’s ability to raise
women’s relative wages because the interests of a broader constituency come into
As with other policies, employers may be able to avoid comparable worth’s
mandate and thereby both limit its cost to them as well as its effectiveness as a
remedy for the gender wage gap.70 If a comparable worth policy were not universally
applied, covered employers could contract out the functions of female-dominated
jobs to uncovered firms. Similarly, they might increase their use of part-time or
temporary workers if coverage did not extend to these groups. Possibly, jobs could
be redefined in order to downgrade female-dominated jobs by shifting their more
complex tasks to already higher paid employees (e.g., have managers assume the
of comparable worth (a difference of some 420 jobs); male employment would have risen
by 3.9% without the policy instead of 2.9% with the policy (a difference of about 160 jobs).
A number of other studies also have shown small employment effects. See Shulamit Kahn,
“Economic Implications of Public-Sector Comparable Worth: The Case of San Jose,
California,” Industrial Relations, vol. 31, no. 2, spring 1992.
Sorensen, Comparable Worth: Is It a Worthy Policy? pp. 131-132.
Lynda J. Ames, “Fixing Women’s Wages: The Effectiveness of Comparable Worth
Policies,” Industrial and Labor Relations Review, vol. 48, no. 4, July 1995.
Raisian, Ward, and Welch, Pay Equity and Comparable Worth.
word processing functions of their secretaries). Jobs might also be redefined in order
to raise the wages of particular incumbents whom employers wanted to retain.
Concluding Remarks. Whether the proponents of the human capital
explanation or the discrimination explanation are correct about the underlying
reasons for the gender wage gap, the same end result is expected according to
standard economic analysis — namely, depressed wages. Whether women limit their
fields of work in order to fulfill family responsibilities or discrimination confines
women to certain jobs, the market’s wage response is the same: the abundant supply
of women to certain jobs relative to employer demand will prompt firms to use more
women than otherwise and, in hiring each new worker, spread complementary inputs
(e.g., capital) more thinly; this, in turn, decreases the productivity of workers in
female-dominated jobs, and consequently, their earnings.
In neither the human capital nor the crowding model are the wages of
predominantly female jobs necessarily depressed because employers (un)consciously
decide that particular jobs should be devalued, that is, worth relatively low wages.
Rather, it is the presence of numerous qualified persons ready to fill those jobs that
removes the firm’s incentive to offer higher wages in order to attract more workers
or to keep from losing workers already on the payroll.
Regardless of the reason for pay differentials, they are economically desirable
from the perspective of an efficiently functioning labor market. Wage differentials
encourage workers, especially new entrants, to seek higher paying jobs;71 and, they
encourage firms to use more of the plentiful source of labor. If comparable worth
were mandated, however, both these incentives would disappear. With wage
differentials between “equivalent” female- and male-dominated jobs gone, the
strongest motivation for women to overcome discrimination in the long run would
be lost. In addition, female unemployment would increase in the short run without
wage differentials to encourage workers to exit from or not enter crowded,
traditionally female jobs. Comparable worth also would adversely affect the
allocation of labor among occupations in which wage differentials might not be due
While the advocates of comparable worth have a plausible and empirically
supportable case that discrimination is responsible for a substantial portion of the
male-female wage gap, the policy itself offers a limited solution (i.e., raising the
wages of workers in traditionally female jobs) which cuts two ways with respect to
gender equity: (1) increasing unemployment in female-dominated fields, and (2)
reducing women’s incentive to further integrate male-dominated jobs. These
economic costs may be worth incurring, but that is a judgment for policymakers to
Even if women are traditionally excluded from higher paying jobs, the incentive still
exists for them to try to obtain better positions. To the extent that women encounter
discrimination in this effort, it is of the type that current law exists to remedy.
H.R. 5680 passed the House in the 98th Congress but became stalled in the
Senate. Title I of the bill, the Federal Pay Equity and Management Improvement Act
of 1984, specified that a study of the federal classification systems be conducted to
see if they discriminated against women. S. 958, the Civil Service Amendments Act
of 1984, omitted such a study. No conference committee was appointed after passage
of S. 958. The Senate subsequently voted down an amendment to a continuing
resolution that would have adopted H.R. 5680. During consideration of the
amendment, a compromise was reached on a pay equity study. In return for passage
of a bill to reform the merit pay system and to reorganize the senior executive
service, it was agreed that the General Accounting Office (GAO, now the
Government Accountability Office) would report on methodologies for conducting
a pay equity study of the federal classification systems; hearings would be held on the
GAO report; and, legislation would be introduced to authorize a pay equity study
based on GAO’s findings.
GAO subsequently issued Options for Conducting a Pay Equity Study of
Federal Pay and Classification Systems (GAO/GGD-85-37). It discussed two
approaches for determining why some female federal employees earn less than male
federal employees: an economic analysis to measure and explain gender wage
differentials due to employee characteristics (e.g., educational attainment and
seniority); and a job content analysis to examine job characteristics (e.g., skill level
and working conditions). In GAO’s view, an effective pay equity study would use
both analyses. Then, during the 99th Congress, House and Senate committees held
hearings on the GAO report. The House Committee on Post Office and Civil Service
reported out H.R. 3008, the Federal Equitable Pay Practices Act of 1985. It would
have established a Commission on Equitable Pay Practices which was required to
contract with a private consultant, chosen from a list of experts drawn up by GAO,
to conduct a pay equity study employing both an economic analysis and a job
evaluation. The consultant’s study and Commission recommendations were then to
be forwarded to the President and the Congress. The House passed the bill as
amended and referred it to the Senate Committee on Governmental Affairs. In the
Senate, the Subcommittee on Civil Service held hearings on a similar bill, S. 519 (the
Federal Employee Anti-Sex-Discrimination in Compensation Act of 1985), and S.
5, the Pay Equity Act of 1985, was proposed as well.
H.R. 387, the Federal Equitable Pay Practices Act of 1988, was introduced
during the 100th Congress and was passed by the House. It was identical to the 99th
Congress’ H.R. 3008. S. 5, the Pay Equity Act of 1987, also was reintroduced but
in a substantially revised form. It was a broad bill requiring that the Equal
Employment Opportunity Commission develop guidelines for eliminating
discriminatory wage-setting practices in the federal government and among federal
contractors as well as furnish a report detailing its activities under the Equal Pay Act,
and a Commission on Compensation Equity to select a consultant who would
conduct a pay equity study of the government’s pay system. The study included in
S. 552, the Federal Employee Compensation Equity Act of 1987, was similar to S.
5’s study, with both calling for conducting an economic analysis and job evaluation.
The Senate took no action on this legislation.