Free Markets, Free Choices:
Women in the Workforce


December 1995

Executive Summary

American women now find themselves at the center of the debate over race- and gender-based preference policies. Advocates of these programs need women's support to ensure the survival of the status quo. It is only with women voters on their side, with roughly 53 percent of the electorate, that defenders of special preferences can steer their golden cow into safe pastures. Since government maintains programs for women, preference advocates hope women will be more likely to support these programs.

Advocates of preferences rely on two tactics to align women with their cause. The first is the tactic of fear-mongering -- its machinations are the wage gap and the glass ceiling. In 1959, activists asserted that for every dollar a man earned, a woman earned a mere 59 cents. They maintained this was eo ipso proof of gender discrimination. By 1995, that gap has narrowed, but the average American woman still only takes home about 72 cents for every dollar the average man puts in his wallet. The pay gap's legacy is the "glass ceiling," a conveniently less concrete complaint about the barriers to success facing women at the highest echelons of corporate America.

That women are successful today paradoxically leads us to the second, and more pernicious, supposition: that women owe their increased success in the marketplace to programs that give them preference. While women still earn a fraction of their brothers' salaries, this fraction would be even less were it not for the leaders of the feminist movement, or so these leaders would have women believe.

This briefing exposes the first proposition -- that wage and attainment differentials between the sexes are due to discrimination -- for the red herring it is. Analyzing census data and summarizing economic labor studies, this briefing shows that the gaps reflect not discrimination but different levels and fields of education, different career choices and marriage. When women and men with the same levels of education, field of education, and workforce experience are compared, the gap virtually disappears.

The second contention, that women owe their professional success over the last thirty years to preference programs stewarded by feminist advocates, can't be proven one way or the other. Two facts are beyond dispute. First, over the last 30 years, women increased their representation both at the upper levels of corporate America and in this country's professional schools. Second, to varying degrees over the same period, programs that accord preference to women have been in place in America's institutions. But a correlation does not prove cause.

Chronicling the proliferation of women-owned businesses and the marked increase of women enrolled in professional schools, this briefing argues that the necessary condition for much of women's success is affirmative action in its original sense -- equality of opportunity. Furthermore, there is scant reason to believe that women will not continue to succeed in an environment in which they are simply accorded the same rights and opportunities as men, without special preferences.



Underpinning this briefing is the fundamental idea that all people should be accorded the same rights, equally protected under the law, in a free society. This implies certain things. First, humans, regardless of gender, are self-directing entities, ends in and of themselves rather than the means to a social planner's conception of a perfect and just society. Second, equality of results can neither be justly pursued nor practically attained. Given that people have different talents and make different choices, equality of rights and equality of outcomes are mutually exclusive. Pursuing equality of outcomes -- which necessitates assigning people unequal rights -- is anathema to the principles of a liberal democracy.

This briefing quantifies income and education across gender lines in order to refute the most blatant misrepresentations circulating in the popular press. But it does so with hesitation, recognizing that in so doing it flirts with the very premise it seeks to reject -- that in a discrimination-free world, women and men would be equally represented in all fields and in each of society's institutions. In a free society of self-directing individuals there are a myriad of reasons, many having to do with the division of labor within the family, why women and men may not choose the same fields of study and ultimately their professions in equal proportions. In fact, given the world we live in, it is the preference advocates' premise that must be justified: that all deviations from perfect gender equality in the marketplace derive from discrimination.

Defenders of preference programs claim that women owe their success in the marketplace to their programs. Illustrating the blurring of truth and propaganda is a guest editorial in The New York Times claiming that affirmative action, "for all its imperfections, has made a major difference in the lives of women and minority groups and has helped us achieve the constitutional commitment to the ideal of equality and fairness."1

But advocates must be reminded: correlation does not establish cause. Ice cream sales and rattlesnake bites increase at the same time of year, but few would argue that rattlesnakes are responsible for Dreyer's success. Since preference programs have been in place for the past thirty years, it is impossible to say with certainty how women would have fared in their absence. Advocates want women to believe that few gains would have been made in the absence of preferences -- dependence breeds loyalty. But no quantitative scale exists to measure the impact of these programs on women's lives. There is no way to factor out for root changes in social attitudes or for the impact of technology. Therefore, it is impossible to prove definitively that American women would or would not be where they are today without special preference.

This being the case, myths still parade as truths in the current discourse on women and preference programs. Making an honest appraisal of the position of women in the American economy, this paper will dispel the two most prominent myths about women, their role in the economy, and preferential treatment. First, that there is a yawning wage and achievement gap between the sexes due to systematic discrimination, and second, that women owe the gains they have made over the last thirty years not to expanded equality of opportunity, but to preference programs.

Equal Protection Under the Law

Advocates of preference programs would have women believe that America's history is riddled with their exclusion. Therefore, as victims, they are entitled to reparations. As Christina Hoff Sommers points out in her book, , "The 'gender war' requires a constant flow of horror stories showing women that male perfidy and female humiliation are everywhere."2

But is this the case? Are American women the unwitting victims of an ever-dominant patriarchy?

Women's political equality in America has a long, steady history. The principles upon which this country was founded, the notion that "all men are created equal,"3 laid the framework for every American to pursue his or her own dreams. In keeping with this ideal, the Fourteenth Amendment of the U.S. Constitution established the principle of equal protection under the law for all American citizens. The Nineteenth Amendment of 1920 introduced gender to the Constitution by superseding state laws to specifically state that the right to vote could not be "abridged . . . on account of sex."4

Women's economic and political rights continued to grow apace in the twentieth century. To address the changing face of the workplace and the awkwardness with which it was initially received, Congress passed the Equal Pay Act in 1963. This act sought not to equalize wages across professions -- to mandate, for example, that a receptionist should earn the same salary as a biochemist -- but to prohibit wage differentials based solely on sex. The Equal Pay Act specifies that wage differentials made pursuant to a seniority system, a merit system, or a quantitative measure of production are justified and legal. Consistent with the principle of equality of opportunity, it offers protection under the law while recognizing that women are equal to the challenge of competition.

One year later, during the passage of the 1964 Civil Rights Act (hereafter referred to as "the Act,") sex was added to the list of protected groups whom the Act was meant to serve. The Act federally protects the constitutional rights of traditional victims of discrimination in this country. "Sex" is mentioned in Title VII on Equal Employment Opportunity, where the Act specifically delineates inappropriate considerations in hiring practices, tacking sex onto a list that includes race, religion, and color.

Title VII explicitly states that nothing contained in its language can be construed as requiring employers to "grant preferential treatment" to any individual or group "on account of an imbalance which may exist with respect to the total number or percentage of persons of any race, color, religion, sex, or national origin employed . . . in comparison with the total number or percentage of persons of such race, color, religion, sex, or national origin in any community. . .."5 In other words, the intent of the 1964 Civil Rights Act was equality of rights, not equality of results or representational parity.

This distinction between equality of rights and equality of results is critical to the current debate over preference policies. The Constitution promises "the blessings of liberty"6 for all this country's citizens. The 14th and 19th Amendments affirm that constitutional promise of citizenship, while both the Equal Pay Act of 1963 and the Civil Rights Act of 1964 take federal protection one step further, ensuring every American's equal right to compete in a free market. All of these laws are protections of opportunity, not outcome.

Since individuals have disparate tastes and talents, ensuring equality of opportunity ensures an inequality of outcomes. Equality of results cannot be pursued without destroying equality of opportunity or equal treatment under the law. If outcomes are to be made equal, or we are to achieve representational parity, some people simply must be treated "more equal than others."

It is disingenuous to suggest that discrimination no longer exists in America. But discrimination, or "the according of differential treatment to persons of an alien race or religion,"7 is necessarily perpetuated when we privilege group membership over individual achievement. The anti-discrimination laws in place must be vigorously enforced, but race- and gender-based preferences fuel the very fire they claim to extinguish.

Given the legal protection American women enjoy, how then, to explain the much decried gender wage gap or that fashionable metaphor of institutional discrimination, the glass ceiling? The answer is personal choice. Everyday, at home and at work, we make decisions that affect our lives in different ways. Before we look at the specific allegations of the gender pay gap and the glass ceiling, we must examine the impact of individual choices on a woman's career and earning power. This examination will expose the fundamentally flawed concepts of the wage gap and the glass ceiling.

B.A., M.A., Ph.D., and Mrs. -- The Effects of Personal Choice

The biggest pitfall in using statistics to measure discrimination is the assumption that at any given level of aggregation all variables not accounted for in the model are constant or equivalent across the groups of study. Dr. Thomas Sowell, a noted economist at the Hoover Institution who has written extensively on civil rights and discrimination, calls this the "residual fallacy" in his most recent book, The Vision of the Anointed: Self-Congratulation as a Basis for Social Policy.8 Sowell writes:

The residual fallacy is one of the grand non sequiturs of our time, as common in the highest courts of the land as on the political platform or in the media or academe. At the heart of the fallacy is the notion that you really can hold variables constant -- "controlling" the variables, as statisticians say -- in practice as well as in theory.9

When researchers assume that the variables they do not measure are constant across the groups of study, they attribute any residual differences in measurement to discrimination. In the real world, as Sowell notes, that which is assumed to be constant rarely is and, therefore, discrimination is often erroneously inferred. It is meaningless to compare the median wage of all women to that of all men; common sense tells us all the factors affecting wages could never be constant at such a broad level of denomination.

While it is certainly easier to deal in aggregates, it is not accurate. The data can be corrected in a variety of ways to provide a more realistic picture of the role of women in the marketplace. Education and marriage bear significantly on a woman's career; they shape her opportunities and are among her considerations when making commitments of time and energy. An examination of the impact of marriage and of education follows.

Educational Attainment and Field: When a B.A. does not equal a B.A.

Education is an area in which the "residual fallacy" can be especially misleading. Figure I compares annual earnings by highest degree earned between men and women. This would appear to control somewhat for differences in achievement, but it indicates that men still earn consistently more than women. It then becomes tempting for advocacy groups to assign the residual disparity to discrimination. But a moment of consideration yields several other factors that would affect earnings in addition to the level of degree an individual has earned: age, profession, continuous years in workforce, seniority, and field of study.

Figure I: Average Earnings of Women Relative to Men By Highest Degree Earned, 1994

Women do earn less than men at every educational attainment level, but when we examine the degrees they hold, we find that it is not discrimination but educational and career choices that account for the earnings' differentials.

In 1992, more than a third of the bachelor's degrees earned by women were in the fields of communication, education, English literature, health Professions, and the visual and performing arts. In this same year, less than 4 percent of women who earned degrees did so in engineering and mathematics and less than 1 percent did so in physical sciences. Notably, the largest category of bachelor's degrees earned by women was in business, at 20 percent, which bodes well for the future of women in the market.10 In the same year, 26 percent of men who earned bachelor's degrees did so in business and 13 percent did so in engineering.11

The contrasts become more striking when we move from bachelor's to advanced degrees. In 1992, 37 percent of master's degrees and 27 percent of doctorates earned by women were in education while 34 percent of master's degrees earned by men were in business and 22 percent of doctorates earned by men were in engineering.12 In other words, while one out of four women who earned Ph.Ds did so in education -- a field in which a person with an advanced degree earns a mean monthly income of $3,048 -- one out of five men who earned doctorates did so in engineering, a field in which an individual with an advanced degree earns a mean monthly income of $4,049.13

In 1992 women earned:

75 percent of the master's degrees and doctorates conferred in education,
70 percent of the master's degrees and doctorates conferred in public administration,
65 percent of the master's degrees and doctorates conferred in English literature, and
63 percent of the master's degrees and doctorates conferred in ethnic and cultural studies.14
In this same year men earned:

86 percent of the master's degrees and doctorates conferred in engineering,
75 percent of the master's degrees and doctorates conferred in physical sciences and science technologies,
65 percent of the master's degrees and doctorates conferred in business management, administrative services and marketing, and
60 percent of the master's degrees and doctorates conferred in mathematics.15
Again, it must be recalled that the figures cited above are for one year -- a snap shot, but not an accurate representation of the overall labor pool. Actually, since women have increased their representation in such fields as business and engineering in recent years, these figures overstate the proportion of women available for work in these fields, as they should be combined with lower percentages from previous years.

Pulling the camera back to look at aggregate achievement in different fields of study, we find:

As of 1990, 28 percent of women who held a bachelor's degree or higher earned that degree in education, while 24 percent of men had degrees in business management and 16 percent had degrees in engineering.16
In 1990, women held 81 percent of bachelor's and advanced degrees in nursing, pharmacy and health technologies, 72 percent of bachelor's and advanced degrees in education, and 61 percent of bachelor's and advanced degrees in liberal arts and humanities.17
In this same year, males held 91 percent of bachelor's and advanced degrees in engineering, 74 percent of bachelor's and advanced degrees in physical and earth sciences, 70 percent of bachelor's and advanced degrees in business management, and 67 percent of bachelor's and advanced degrees in mathematics and statistics.18
Figures II through IX take popular and high-yield fields of study, disaggregate them by gender distribution, and provide the mean monthly income of careers associated with those fields.

Mean Monthly Income and Gender Breakdown of Those Who Held Bachelor's Degrees or Higher in 1990 in Selected Fields of Study


Click on the highlighted text for a graphical presentation of data.

Figure II: Business
Figure III: Education
Figure IV: Engineering
Figure V: English and Journalism
Figure VI: Liberal Arts and Humanities
Figure VII: Mathematics and Statistics
Figure VIII: Nursing, Pharmacy and Health Technologies
Figure IX: Physical and Earth Sciences


As Figures II through IX illustrate, the fields in which women and men choose to pursue their education vary widely in the salaries they command upon graduation. Given this reality, comparing incomes by the highest degree earned does not measure discrimination. Median income ratios like those in Figure I are politically useful, but factually vacuous. Unfortunately, they are often perceived as fact by people who should know better.

Earnings and Marriage

Marriage plays an ambiguous role in this numbers game. The effect of marriage on women's earnings can be politically treacherous ground on which to tread; in some circles, a discussion of the issue is tantamount to relegating women to domestic dependency. The fact is, married women tend to earn less themselves but improve their overall standard of living through shared income. In other words, marriage has an asymmetrical effect on earnings of women.

As Dr. Thomas Sowell notes in his 1985 book, Civil Rights: Rhetoric or Reality?, marriage has exactly the opposite effect on the earnings of men and women. "Marriage increases a man's rate of participation in the labor force compared to single men and reduces a woman's labor force participation rate compared to single women," writes Sowell.19

What Sowell described in 1985 still holds true today. As illustrated by Figure X and XI, from 1960 through 1993, married men participated in the labor force at higher rates than single men while the opposite is true for women. The largest participation gap, as illustrated by Figure XII, is between married men and married women. Although the gap has narrowed significantly over the last 33 years, in 1993, married men still participated in the labor force at a rate 30 percent greater than that of married women.

Figure X: Male Labor Force Participation Rates, 1960-1993

Figure XI: Female Labor Force Participation Rates, 1960-1993

Figure XII: Labor Force Participation Rates for Married Women and Married Men, 1960-1993

Labor force participation rates affect income levels in various ways. Perhaps the strongest connection lies among labor force participation rates, continuous years in the labor force, and seniority, all of which affect compensation levels.

Labor force participation rates correlate positively with continuous years in the workforce. This makes sense if you consider that at any given time the labor force participation rate is simply a snap-shot of those in the workforce at that moment. In other words, even in 1993 when nearly six out of ten married women were in the workforce, those six women could change from year to year and month to month.

Information from the U.S. Bureau of the Census bears this out. A 1992 U.S. Census study on the relationship between education level and earnings again found that at every education level "men have more months with work activity than do women."20 A 1984 U.S. Bureau of the Census study found that while only 1.6 percent of a man's work years were spent away from work, 14.7 percent of a woman's work years were spent away from work -- an eightfold difference.21 This prompted the U.S. Department of Labor (DOL) to conclude "women spend significantly more time away from work and are apparently unable to build the seniority that men achieve."22 The DOL also noted "turnover data for women and men have shown higher rates for women than for men."23

With less seniority, it is hardly surprising that women's earnings are lower than men's. Dr. Robert Topel, in a 1990 National Bureau of Economic Research working paper, reported that there is a "very strong connection between job seniority and wages in the typical employment relationship: other things held constant, 10 years of job seniority raises the wage of the typical worker by over 25 percent."24 While it is simply common sense that people who work more earn more, these inconvenient facts are often absent from "factual" presentations on discrimination. The good news is that reality and theory are congruent. The same Census study that showed women, on average, spend eight times as many years absent from the work force than men, showed that women with no interruptions earned nearly as much as men, even before accounting for educational levels and fields of study.

Sowell explains that the gap in income is not between women and men but between married women and everyone else, which is primarily the result of a voluntary domestic division of labor, not labor-force discrimination. Sowell further explains the effects of marriage on men's and women's income.

A married man's hours worked annually increase with the number of children. A married woman's hours decrease as the number of children increases. Married men work more and earn more than single men, while it is just the reverse with women. Married men with children work the most and earn the most, while married women with children work the least and earn the least. Altogether, married women living with their husbands average only 25 percent of the annual income of married men living with their wives.25

The overwhelming majority of American women -- 81 percent as of 1993 -- choose to marry.26 Some people might be troubled by the fact that even in an age of two-career families, women not only physically bear the children but also carry the majority of the burden in raising them. But in a free society there is little that can, or should, be done about the choices people make in the ordering of their lives. The good news is that there is vast evidence that women who choose to remain single, invest in education, and work long hours, have in the past and continue to fare about as well as men in the labor market, and apparently continue to do so. Again we find the residual fallacy; in this case, what advocacy groups measure as discrimination can largely be attributed to marriage and individual choice.

Myths of Oppression -- The Arsenal of Gender Warfare,

With this proper conceptual framework, we are now prepared to analyze specific claims of widespread discrimination against women.

The Gender Wage Gap

The gender wage gap is simply a false construct, an example of statistical manipulation that casts women in the role of victims. The pay gap myth allows preference proponents to assert that, despite anti-discrimination laws, women are still denied equal pay in an inherently sexist labor market.

This trick is as old as the game of politics. In the 1960s, protesters wore 59-cent buttons, claiming that women were the victims of market discrimination because the average working woman earned 59 cents of every dollar annually earned by the average working man. Today, wage gap activists grudgingly concede that the aggregate numbers are closer to 72 cents for every dollar (the aggregate gender gap in wages declined, on average, about one percent per year since 197627), but either slogan, 59 cents or 72, is factually bankrupt.

In anticipation of public debate over special preference programs, the California National Organization for Women (CA NOW) compiled a hefty booklet this year, titled Guide to Women and Affirmative Action. In chart after chart, NOW paints a picture of systematic wage inequities. One section of the reader, "Statistics on Women and Work in the United States and California," illustrates the manipulation of information typical of this debate.

It begins with "In 18 out of 19 categories reported by the Bureau of Labor Statistics, women earn less than men with a range from 60 cents on the dollar to a high of 89 cents on the dollar."28 This sounds dramatic because NOW refuses to disaggregate, or break down, the wage profiles by age, experience, or even educational achievement and field.

Similarly, the National Committee on Pay Equity circulated a paper proclaiming that in 1993, women "earned 72 percent as much as men."29 They arrive at this number by dividing the median annual earnings for full-time working women by the median annual earnings for full-time working men. Different career paths aside, this gap statistic does not account for total or continuous time spent in the workplace, nor does this "factoid" account for educational attainment.

When all the factors that go into determining one's wage are taken into account, it becomes clear that simply dividing the aggregate median earnings of women by the aggregate median earnings of men reveals little about whether women are being discriminated against in the labor market. Employers, after all, do not hire averages, they hire people. As discussed above, among the chief factors that influence pay levels are educational level and field -- we wouldn't expect someone with a BA in sociology to earn as much as someone with a master's in Finance -- regardless of their respective genders.

Time spent working, particularly continuous years in the workforce, also bears directly on an individual's level of compensation. Irrespective of gender or field of work, it is hardly reasonable to expect someone who has taken a significant break in his career to earn the same as someone who has not. Similarly, someone who works 55 hour weeks will, in the absence of discrimination, earn more over time than someone who punches the clock at 40 hours. After all, someone who works 55 hours a week piles up 3.5 years of experience, at a 40 hour a week standard, every 2.6 years. Women still work fewer hours than men. In 1994, only 55 percent of women worked 40 or more hours a week compared to 75 percent of men. While 16 percent of men worked 55 or more hours a week that same year, only 6 percent of women did.30 Women cannot expect equivalent annual incomes for less time on the job. Looking at hourly earnings for women, we find they are almost 10 cents to the dollar closer to men's than the annual earnings' figure.31 Yet 72 cents, the annual figure, is widely cited as an accurate representation of wage differentials.

The gender pay gap virtually disappears when age, educational attainment, and continuous time spent in the work force are factored in as wage determinants. As early as 1971, never-married women in their thirties who had worked continuously earned slightly higher incomes than men of the same description.32 In the academic world, Dr. Sowell notes that "single women who received their Ph.D.s in the 1930s had by the 1950s become full professors slightly more often than male Ph.D.s as a whole."33 Sowell also found that never-married academic women earned slightly more money in 1968-1969 than men in the same situation.34 June O'Neill, economist and current Director of the Congressional Budget Office, has written extensively on gender wage differentials. A 1993 study in the Journal of Labor Economics, co-written with Solomon Polachek, confirms that a woman's success in the labor force, like a man's, can be attributed to education, seniority, and quality of experience. Among O'Neill and Polachek's findings is that the gap in earnings between men and women who hold Ph.D.s in economics is a mere five percent.35

Similarly, data from the National Longitudinal Survey of Youth show that among women and men ages 27-33 who have never had children, the earnings of women are close to 98 percent of men's earnings.36 Indeed, when women and men who have made similar life choices are compared, the gap virtually disappears.