(Archived document, may contain errors)
336 March 2, 1984 I COMPARABLE WORTH PART 1 A THEORY WITH NO FACTS INTRODUCTION The Equal Pay Act of 1963 and the Civil Rights Act of 1964 required equal pay for equal work in most fields of employment. Beyond the law, nearly every American solidly endorses equal pay for equal work. Yet many women's groups allege that this does not go far enough to-end what they see as co ntinued widespre'ad discrimination. As evidence, they argue that the average woman with a full-time job earns only 62 percent of a male's earn1ngs.l This earnings differential has triggered a campaign for what is called comparable worth This means, accord i ng to its advo cates, equal pay for jobs requiring similar levels of training responsibility, and other employee characteristics--even if these jobs are in completely different fields. Determining what is and is not llcomparablell would be left to bureauc r ats or judges or government boards. What comparable worth supporters thus seek is to replace the existing market system of wage determination with a wage-setting mechanism that would define "fair wages for jobs of supposedly comparable value. In effect, l lcomparable worth would require the government to force employers to pay equal wages for unequal work. Advocates of comparable worth assert that women have been victims of sex discrimination because they have been forced into jobs that typically are underp aid relative to 'lcomparablell male- dominated jobs. is based on no solid facts or data personal, cultural, and market conditions that are critical in determining wages and account for wage differentials. amination of these factors reveals that the relati v ely lower earn But this oversimplified view of the labor market Closer ex It ignores the host of Robert Pear, "Earnings Gap is Narrowing Slightly for Women," The New York Times -9 October 3, 1983, p. B15 2 ings obtained by women actually reflect their own preferences and productivity-not systematic sex discrimination by society. If such discrimination did exist, nondiscriminating and profit-seeking firms would simply hire females for lower wages, thereby lowering production costs and enabling'them to bid b usi ness away from discriminating firms by charging lower prices. In other words, the business community would consist of two groups those who discriminate, pay more to hire males, and thereby erode their own profits; and those who hire only women at lowe r cost, thereby increasing their profits nation on the basis of sex have yet to explain why this has not happened. The fact is, there is a great deal of wage competition in the labor market, and wages tend to reflect a worker's produc tivity. Indeed, after adjusting.for relevant and measurable productivity differences between men and women, many researchers have been able to eliminate about 50 percent of the wage gap. Some studies have explained the entire earnings gap in measurable economic terms.2 And ev en those studies that do not attribute the whole wage differential to economic factors cannot assert that the differential is due to sex discrimination-only that re liable statistics on certain factors are difficult to assemble. Notes Emory University eco n omist Cotton Lindsay The evidence of wide and persistent wage disparities between working men and working women thus does not point'unequivocally to discrimination. Economic theory suggests little scope for discrimination in employment to produce wage eff e cts, and therefore observed wage differences are likely to originate in productivity and taste differences, not in discrimination Those who allege wage discrimi The assumption ,that differences must mean discrimination is not supported by analysis of the factors affecting wage levels. Supporters of comparable worth must provide a far more convincing hypothesis and body of evidence before the1 can expect their argu- ments to appeal to more than emotionalism SOME FACTS ABOUT COMPARABLE WORK Comparable worth advocates urge the creation of wage boards to determine fair wages for work of comparable value. They ignore For a summary of the empirical evidence, see Cotton M. Lindsay and Charles A. Shanor, "County of Washington v. Gunther: Economic and Legal Conside rations for Resolving Sex-Based Wage Discrimination Cases," Supreme Court Economic Review, vol 1, 1982. Ibid p. 221. This study is the first of a two-part series on the comparable worth issue. Part I1 examines the implications of the comparable worth do ctrine for the competitive market system 3 the existing mechanism for computing such comparability-the free labor market In a competitive labor market, wages are determined by the voluntary interaction of employers and employees. And com- petition on both sides of the bargaining table ensures that the wage rate ultimately will be based on a worker's productivity. Supporters of the comparable worth doctrine fail to understand this wage-setting process; they attribute--erroneously--earnings differences betwe e n the sexes to discriminati~n Many factors affect a worker's productivity. These include inherited characteristics and environmental factors: intelligence health, physical strength, dependability, and sthe ability to in teract socially. Other characterist i cs, such as work skills, on-the-job training, and job experience, are learned or acquired during the process of education and in the workplace prisingly, employers will normally be willing to pay a higher wage to workers with a greater endowment of such a t tributes In addition, the demand for particular services plays a sig nificant role in the wage determination process. For example many assume that secretaries are paid less than lawyers solely because the value of a secretary's output is worth less than t h at of a lawyer's taries is because of their relative scarcities. There are many more secretaries than there are lawyers. The productivity of any worker, and hence the wage, diminishes as the numbers in an occu- pation grow. If there were as many lawyers a s there are secretaries lawyers could obtain only a fraction of the fees they now do A wage differential between these two occupations will exist as long as people view the costs of becoming a lawyer (such as tuition and earnings foregone during schooling) as more than that required to become a secretary these costs began to narrow, some would divert their energies from becoming secretaries to becoming lawyers. This employment shift would push down the wages of lawyers and raise those of secretaries, until t he wage and cost differential again returned to a more acceptable equilibrium Not sur But another reason lawyers earn more than secre I I I If the differential between Emory economist Lindsay summarizes how the marketplace pre vents employers from arbitra r ily determining wages The wage setting process is the result of two conditions, neither of which the employer controls. First, wages are limited from above by the worker's productivity in the job. Profit' considerations prevent the firm from paying the wo r ker more than he or she is worth to it. Second, supply considerations prevent an employer from paying to workers of a given productivity a wage that For a theoretical discussion of discrimination, see Gary Becker, The Economics of Discrimination (Chicago 1957 and Lindsay and Shanor, op. cit. University of Chicago Press, 4 makes working for that employer less attractive than working for other employers. To do so invites these workers to seek employment elsewhere.6 Thus, the market process has a built-in mec hanism to elimi nate discrimination: the profit motive. If there were differences in pay between the sexes based on discrimination rather than on differences in productivity, this situation could not be long sus tained. The reason: firms that discriminate d against women would face higher production costs than firms that did not discriminate against women. In the long run, discriminating employers would pay a financial penalty and would either change their ways or be driven out of the market motivated emplo y ers would pass up an opportunity to hire equally productive women for 40 cents less on the dollar-the difference in labor costs attributed by many feminists to discrimination just to indulge an irrational preference for hiring men It is inconceivable that profit WHY WOMEN EARN LESS If discrimination does not explain the earnings differential between men and women, what factors do? Women Entering the Work Force One factor is the very large recent influx of women into the labor force. The female labor force participation rate grew from 34 percent for those aged 16-64 in 1950 to 52 percent in 1980.8. A high proportion of these entrant s, of course, have had to start at entry-level positions, which predictably pay less than the average. The surprising aspect of this phenomenon, in fact, is that the wage gap has not widened as the female sector of the workforce has ballooned. That it has not widened is because women as a group are now moving into higher paying jobs. Once the surge of women into the labor force levels off, women will no longer be disproportionately represented in entry-level positions and their aggregate earnings relative to men will increase. Hours on the Job A major flaw in measuring alleged male-female wage differen tials lies in comparing the total annual earnings for men is com Ibid. This would be true of firms operating in the private sector. Discrimina tion may exis t in the public sector because it is not subject to competi tive forces. It would therefore be desirable to interject greater competition in the wage-setting process in the public sector, preferably by contracting out or privatizing many of the functions of government. U.S. Bureau of Census, American Women: Three Decades of Change, Special Demographic Analyses, CDS-80-8 (Washington, D.C U.S. Government Print ing Office, August 1983 p. 15 5 pared to the total annual earnings for women. Yet statistically wom en are less likely than men to work the full year. As a result the difference in earnings is greater than would be the case if quarterly, weekly, or hourly rate data were used for comparison. For example, in the second quarter of 1983, the median earnings for women were 66 percent of the median earnings for men.g correcting the data for this one factor reduces the alleged gap by 4 percentage points Simply Role Differentiation According to Stanford University economist Victor Fuchs role differentiation is t he principal explanation of male- female wage disparity.1 This role differentiation between the sexes begins in childhood, he says, and is eventually reflected in the marketplace through differences in labor force attachment occupational choice, schooling , location, time spent in market employment, and other factors that determine wages. The roles of men and women in the family unit traditionally have differed-the husband normally having primary responsibility for providing financial support, and the wife having primary responsibility for child rearing and household upkeep. Thus, if a married woman does take a job, it is generally one that is compatible with her household responsibilities and her husband's preferred place of work. As Fuchs points out, it i s women's own preference, not sex discrimination, that leads to their relatively lower earnings. Feminists may deplore the choice of role made by most women, but they can hardly blame employers for it. Geographic Mobility1 A key reason married women have f ewer opportunities than men Husbands generally earn more than their wives to accept the best jobs available to them arises from their re- stricted mobility because they work longer hours and have made greater investments in their jobs in terms of schoolin g, training, and experience. If both spouses are searching for work, it is unlikely that the most desirable jobs for each will be in the same area. So one has to forego his or her job preference. Since the husband is the primary earner in most families, an d the wife manages the household, the wife will generally accommodate to the husband's choice if the family is to maximize its total income. Married women must also search for employment close to home because of their family responsibilities, which gives them a smaller pool of jobs from which to choose. As a result, many women earn less than they otherwise could, not because of employer discrimination, but Pear, op. cit. Victor Fuchs Recent Trends and Long-Run Prospects for Female Earnings ,It American Economic Review, May 1974, pp. 236-2 42. See Robert. H. Frank Why Women Earn Less: The Theory and Estimation of lo l1 Differential Overqualification American Economic Review, June 1978 pp. 360-373. 6 truly reflect the risks and costs for the firm, aggressive com- panies (perhaps owned by women) could improve their profits by recruiting only women. Proponents of the discrimination theory have thus far failed to explain why this does not happen In addition to the effects of employees leaving voluntarily wages a re also related to the layoff rate in an industry. Indi viduals working in an industry characterized by frequent layoffs or employment insecurity are generally paid more than workers in similar industries with a lower layoff rate, since the demand side of the wage equation reflects the risk faced by the employee because of their own choice of role in the family. found that this may explain as much as one-quarter of the wage differential between the sexes. l2 One study has I Turnover Rates13 Women tend to m o ve in and out of the labor force more often than men do, and this reduces the employability--and hence wages of women as a group. High turnover rates among women require firms to hire and train new workers more frequently, adding to .total labor costs. It is only rational, therefore, for firms to seek workers with a low probability of leaving have to make decisions on the basis of past experience. employer believes a woman is more likely than a man to leave the firm--and statistics show this to be the case - -the woman is likely to be hired only if she accepts a lower wage than a man with identical credentials to compensate the firm for the additional risk. l4 With imperfect information about the potential productivity of a worker, employers If an l2 l3 Ibid. , p. 3 70. This estimate was for Standard Metropolitan Statistical Areas with populations under 250,000. See Elizabeth M. Landes, "Sex-Differences in Wages and Employment of the Specific Capital Hypothesis Economic Inquiry, October 1977, pp 523-538, and J ames F. Ragan and Sharon Smith The Impact of Differences in Turnover Rates and Male/Female Pay Differentials Journal of Human Resources, Summer 1981, pp. 343-365. Women are 11 times more likely to drop out of the labor force than men A Test l4 are 1984, p. 104. See Illinois Commission-on the Status of Women, Minority Report, 7 Since male-dominated jobs are more closely linked to fluctuations in the economy, part of the wage gap between the sexes can be ex- plained by the wage premium paid men to compensate them for the greater risk of job loss. One study has concluded that Ithe com- bined effect of training and turnover account for between 67 and 100 percent of the differential. If Education and Traininql Reflecting their higher turnover rates, women tend to invest less than their male counterparts in productivity-increasing edu- cation and training. And since women as a group are more likely to interrupt their care e r development-=for childbearing and house hold duties--they tend to have less job expe-rience and seniority than men. This is an important factor in explaining their rela- tively lower earnings A major reason that many women do not invest as much in their own "human capital" is precisely because they expect to be absent from the workforce for extended or frequent periods, and thus have a shorter "payoff'l period over which to recoup their investment in education and training. Consequently, they are less li k ely to study for highly specialized fields, such as the sciences, medi cine, and law, because the payoff period for these disciplines is longer. By contrast, many other female-dominated fields, such as teaching and nursing, have relatively short and inexp e nsive train ing periods for extended periods. The higher turnover rate associated with women means, therefore, that firms are generally less willing to provide women with costly training to upgrade their skills induces women to work in occupations where s p ecific training is less important, thereby minimizing the investment loss as they move from job to job. It means also that women are less likely to be found in high-paying jobs that require considerable training If the traditional family roles continue to change, and women play a more important and permanent role in the labor force, these earnings differentials can be expected to narrow. Women's atti tudes toward investing in their future are already showing signs of significant change. For example, over t h e last 30 years, women's college enrollments and completion rates have steadily approached those of men.17 Nearly half of all bachelor and master's degrees in 1979-1980 were awarded to women. Women are also making major inroads into traditionally male-dom i nated I Skills depreciate when workers are out of the labor force This 0 l5 l6 Landes, op. cit p. 532 See Jacob Mincer and Soloman Polachek, "Family Investments in Human Capi tal: Earnings of Women," Journal of Political Economy, March/April 1974 pp. 576- 608 12-13 l7 The following statistics are from U.S. Bureau of Census, op. cit., pp. 0 8 fields such as dentistry, medicine, law, engineering, mathematics l and science now go to women, compared with just 5 and 10 percent, respectively, in 19 70. As these t rends continue and women in these new disci- plines mature and represent a larger proportion of the profession, future pay gaps will narrow And about-20 percent of all law and medical degrees Career Choices Since women tend to select professions that allo w them to move in and out of the labor force easily, or to work part time, they are disproportionately-but understandably-represented in jobs that afford them this flexibility. Thus, even if their education and training equals or exceeds that of men doing s imi lar jobs, wages in these female-dominated professions are never- theless lower, due to the supply and demand factors of the occupation If women want higher earniags, then they must choose work where earnings do not generally reflect the higher turnove r and other factors associated with women. As Coleman Young, Mayor of Detroit, has explained If a painter makes more than a secretary then let more women be painters."18 Working Conditions Many wage comparisons overlook other aspects of employment. For exa m ple, women may trade off wages for better working condi- tions, such as good hours and pleasant surroundings, particularly are characterized by less agreeable and more dangerous surroundings. indoor work. Many male-dominated professions, on the other hand , I Taxes Tax policy has discouraged women from entering paid employ- ment (particularly once they are married) and developing market able skills. Because the wife is usually a secondary earner, the first dollar she earns will effectively be taxed at her h usband's highest marginal tax rate. By lowering her after-tax reward for work, the progressive tax system creates an economic disincentive for women to pursue a demanding career. Social Security payroll taxes represent an additional disin- centive for wome n considering working or investing in activities that could raise their earnings potential. As the Social Security system is currently structured, a nonworking wife can receive a l8 l9 Cited in John H. Bunzel To Each According To Her Worth?" The Public In terest, Spring 1982, p. 84. See Michael J. Boskin, "The Effects of Government Expenditures and Taxes on Female Labor," American Economic Review, May 1974, pp. 251-256, and Paul McGouldrick Why Women Earn Less," Policy Review, Fall 1981, pp 63-76. 9 spouse' s benefit equal to 50 percent of her husband's own benefit. For each dollar in Social Security benefits she could earn herself therefore, a wife effectively loses a dollar of the spousels bene fit. Even if she were to earn a benefit greater than the spous e's benefit the taxbenefit ratio is still biased against her working. For example if she had the same earnings history as her husband her net extra benefits would only be half of those received by her husband, since she could have obtained half these bene fits without working at all Motivational Differences It appears that part of the difference between the earnings of the two sexes can also be attributed to their different goals in life. Explains Michael Levin, professor of philosophy at City College of N e w York Women, most especially married women, are less willing to work their way up the career ladder and tend to see their income as supplementing that of their husband, the bread winner. 112 O THE CASE OF SINGLE WOMEN those that If the wage gap really is caused by sex'discrimination, then 1 making this charge have to explain why studies have found single women earn almost the same wages as single men.21 The reason is mainly because they have similar employment charac teristics to'leave the labor force for household reasons, such as raising children or because her husband has changed jobs. So skill and experience become the dominant factors in wage setting single women nor single men, on the other hand, are likely to work as hard as married men, who often h a ve the added responsibility of supporting several dependents the single woman is less likely than the married woman I I I Neither In an examination of the wages of married and unmarried females, for instance, Hoover Institution economist Thomas Sowell not e s that unmarried men and women receive about the same in wages for the same job and same credentials. Married men tend to do better, he. says, because they are motivated to provide financial assistance for their spouses, who in turn do somehwat worse in t he market because of other duties. Sowell notes that, IISuch a 2o 21 Cited in Robert D. Hershey Women's Fight Shifts to Comparable Worth," The New York Times, November 1, 1983. See Thomas Sowell Affirmative Action Reconsidered The Public Interest Winter 1976, pp. 47-65; James Gwartney and Richard Stroup Measurement of Employment Discrimination According to Sex Southern Economic Journal 19 73. DD. 575-587: and Walter Block. "Economic Intervention,'Discrimina- I tion, and Unforseen Consequences in Discrimin ation, Affirmative Action and Equal Opportunity, edited by W. E. Block and M. A. Walker (Vancouver British Columbia: The Fraser Institute, 1981), pp. 103-125. 10 situation may not be just--but it does not result, however, from employer discrimination Il2 C ONCLUSION While the concept of equal pay for equal work is nearly uni- versally supported, the notion of I1comparable wortht1 is widely criticized--for good reason backed by no solid facts or data. Itcomparable wortht1 rests on an assumption that society s ystema tically discriminates against women on a massive scale, segre gating them into low-paying jobs simply because they are women. The fact that a disproportionate number of women work in rela tively low paid employment, however, does not prove discrimi n a- tion It rests on skimpy research and is Instead, the argument for The occupational and pay patterns of men and women can be explained without resorting to unsubstantiated claims of sex dis crimination. These patterns result from different cultural role s . The key factor is that women typically have spent about half as many years as men in !aid employment, choosing to devote more time to work in 'the home. Stemming from this underlying distinction, differences in education, training, seniority, experience , turn over, labor force participation patterns, working conditions, personal preferences, and general labor market conditions explain most, if not all, of the wage gap. If the structure of the traditional family continues to change, with younger women mov ing into fields traditionally held by men, the forces of supply and demand will narrow the earnings differential between the sexes. Adjusting to these changes in its unprejudiced and inevitable way, the market system will effect wage adjustment to the evo l ving role of women far more effectively than will the fundamentally flawed notions underpinning the doc trine of comparable worth I As Part I1 of this series shows, the Itcomparable worth alternative to the market system is likely to be devastating to the U.S. economy the law, wages will henceforth be set by judges, boards of Ilex perts and interest group pressure. When the market is distorted in this way, shortages and surpluses are inevitable, leading to ever louder demands for even more government contr ol of the labor market Should the doctrine be given the backing of Peter G. Germanis Schultz Fellow 23 Sowell, op. cit p. 56 See June O'Neill The 'Comparable Worth' Trap Wall Street Journal January 20, 1984.
336 March 2, 1984 I COMPARABLE WORTH PART 1 A THEORY WITH NO FACTS INTRODUCTION The Equal Pay Act of 1963 and the Civil Rights Act of 1964 required equal pay for equal work in most fields of employment. Beyond the law, nearly every American solidly endorses equal pay for equal work. Yet many women's groups allege that this does not go far enough to-end what they see as co ntinued widespre'ad discrimination. As evidence, they argue that the average woman with a full-time job earns only 62 percent of a male's earn1ngs.l This earnings differential has triggered a campaign for what is called comparable worth This means, accord i ng to its advo cates, equal pay for jobs requiring similar levels of training responsibility, and other employee characteristics--even if these jobs are in completely different fields. Determining what is and is not llcomparablell would be left to bureauc r ats or judges or government boards. What comparable worth supporters thus seek is to replace the existing market system of wage determination with a wage-setting mechanism that would define "fair wages for jobs of supposedly comparable value. In effect, l lcomparable worth would require the government to force employers to pay equal wages for unequal work. Advocates of comparable worth assert that women have been victims of sex discrimination because they have been forced into jobs that typically are underp aid relative to 'lcomparablell male- dominated jobs. is based on no solid facts or data personal, cultural, and market conditions that are critical in determining wages and account for wage differentials. amination of these factors reveals that the relati v ely lower earn But this oversimplified view of the labor market Closer ex It ignores the host of Robert Pear, "Earnings Gap is Narrowing Slightly for Women," The New York Times -9 October 3, 1983, p. B15 2 ings obtained by women actually reflect their own preferences and productivity-not systematic sex discrimination by society. If such discrimination did exist, nondiscriminating and profit-seeking firms would simply hire females for lower wages, thereby lowering production costs and enabling'them to bid b usi ness away from discriminating firms by charging lower prices. In other words, the business community would consist of two groups those who discriminate, pay more to hire males, and thereby erode their own profits; and those who hire only women at lowe r cost, thereby increasing their profits nation on the basis of sex have yet to explain why this has not happened. The fact is, there is a great deal of wage competition in the labor market, and wages tend to reflect a worker's produc tivity. Indeed, after adjusting.for relevant and measurable productivity differences between men and women, many researchers have been able to eliminate about 50 percent of the wage gap. Some studies have explained the entire earnings gap in measurable economic terms.2 And ev en those studies that do not attribute the whole wage differential to economic factors cannot assert that the differential is due to sex discrimination-only that re liable statistics on certain factors are difficult to assemble. Notes Emory University eco n omist Cotton Lindsay The evidence of wide and persistent wage disparities between working men and working women thus does not point'unequivocally to discrimination. Economic theory suggests little scope for discrimination in employment to produce wage eff e cts, and therefore observed wage differences are likely to originate in productivity and taste differences, not in discrimination Those who allege wage discrimi The assumption ,that differences must mean discrimination is not supported by analysis of the factors affecting wage levels. Supporters of comparable worth must provide a far more convincing hypothesis and body of evidence before the1 can expect their argu- ments to appeal to more than emotionalism SOME FACTS ABOUT COMPARABLE WORK Comparable worth advocates urge the creation of wage boards to determine fair wages for work of comparable value. They ignore For a summary of the empirical evidence, see Cotton M. Lindsay and Charles A. Shanor, "County of Washington v. Gunther: Economic and Legal Conside rations for Resolving Sex-Based Wage Discrimination Cases," Supreme Court Economic Review, vol 1, 1982. Ibid p. 221. This study is the first of a two-part series on the comparable worth issue. Part I1 examines the implications of the comparable worth do ctrine for the competitive market system 3 the existing mechanism for computing such comparability-the free labor market In a competitive labor market, wages are determined by the voluntary interaction of employers and employees. And com- petition on both sides of the bargaining table ensures that the wage rate ultimately will be based on a worker's productivity. Supporters of the comparable worth doctrine fail to understand this wage-setting process; they attribute--erroneously--earnings differences betwe e n the sexes to discriminati~n Many factors affect a worker's productivity. These include inherited characteristics and environmental factors: intelligence health, physical strength, dependability, and sthe ability to in teract socially. Other characterist i cs, such as work skills, on-the-job training, and job experience, are learned or acquired during the process of education and in the workplace prisingly, employers will normally be willing to pay a higher wage to workers with a greater endowment of such a t tributes In addition, the demand for particular services plays a sig nificant role in the wage determination process. For example many assume that secretaries are paid less than lawyers solely because the value of a secretary's output is worth less than t h at of a lawyer's taries is because of their relative scarcities. There are many more secretaries than there are lawyers. The productivity of any worker, and hence the wage, diminishes as the numbers in an occu- pation grow. If there were as many lawyers a s there are secretaries lawyers could obtain only a fraction of the fees they now do A wage differential between these two occupations will exist as long as people view the costs of becoming a lawyer (such as tuition and earnings foregone during schooling) as more than that required to become a secretary these costs began to narrow, some would divert their energies from becoming secretaries to becoming lawyers. This employment shift would push down the wages of lawyers and raise those of secretaries, until t he wage and cost differential again returned to a more acceptable equilibrium Not sur But another reason lawyers earn more than secre I I I If the differential between Emory economist Lindsay summarizes how the marketplace pre vents employers from arbitra r ily determining wages The wage setting process is the result of two conditions, neither of which the employer controls. First, wages are limited from above by the worker's productivity in the job. Profit' considerations prevent the firm from paying the wo r ker more than he or she is worth to it. Second, supply considerations prevent an employer from paying to workers of a given productivity a wage that For a theoretical discussion of discrimination, see Gary Becker, The Economics of Discrimination (Chicago 1957 and Lindsay and Shanor, op. cit. University of Chicago Press, 4 makes working for that employer less attractive than working for other employers. To do so invites these workers to seek employment elsewhere.6 Thus, the market process has a built-in mec hanism to elimi nate discrimination: the profit motive. If there were differences in pay between the sexes based on discrimination rather than on differences in productivity, this situation could not be long sus tained. The reason: firms that discriminate d against women would face higher production costs than firms that did not discriminate against women. In the long run, discriminating employers would pay a financial penalty and would either change their ways or be driven out of the market motivated emplo y ers would pass up an opportunity to hire equally productive women for 40 cents less on the dollar-the difference in labor costs attributed by many feminists to discrimination just to indulge an irrational preference for hiring men It is inconceivable that profit WHY WOMEN EARN LESS If discrimination does not explain the earnings differential between men and women, what factors do? Women Entering the Work Force One factor is the very large recent influx of women into the labor force. The female labor force participation rate grew from 34 percent for those aged 16-64 in 1950 to 52 percent in 1980.8. A high proportion of these entrant s, of course, have had to start at entry-level positions, which predictably pay less than the average. The surprising aspect of this phenomenon, in fact, is that the wage gap has not widened as the female sector of the workforce has ballooned. That it has not widened is because women as a group are now moving into higher paying jobs. Once the surge of women into the labor force levels off, women will no longer be disproportionately represented in entry-level positions and their aggregate earnings relative to men will increase. Hours on the Job A major flaw in measuring alleged male-female wage differen tials lies in comparing the total annual earnings for men is com Ibid. This would be true of firms operating in the private sector. Discrimina tion may exis t in the public sector because it is not subject to competi tive forces. It would therefore be desirable to interject greater competition in the wage-setting process in the public sector, preferably by contracting out or privatizing many of the functions of government. U.S. Bureau of Census, American Women: Three Decades of Change, Special Demographic Analyses, CDS-80-8 (Washington, D.C U.S. Government Print ing Office, August 1983 p. 15 5 pared to the total annual earnings for women. Yet statistically wom en are less likely than men to work the full year. As a result the difference in earnings is greater than would be the case if quarterly, weekly, or hourly rate data were used for comparison. For example, in the second quarter of 1983, the median earnings for women were 66 percent of the median earnings for men.g correcting the data for this one factor reduces the alleged gap by 4 percentage points Simply Role Differentiation According to Stanford University economist Victor Fuchs role differentiation is t he principal explanation of male- female wage disparity.1 This role differentiation between the sexes begins in childhood, he says, and is eventually reflected in the marketplace through differences in labor force attachment occupational choice, schooling , location, time spent in market employment, and other factors that determine wages. The roles of men and women in the family unit traditionally have differed-the husband normally having primary responsibility for providing financial support, and the wife having primary responsibility for child rearing and household upkeep. Thus, if a married woman does take a job, it is generally one that is compatible with her household responsibilities and her husband's preferred place of work. As Fuchs points out, it i s women's own preference, not sex discrimination, that leads to their relatively lower earnings. Feminists may deplore the choice of role made by most women, but they can hardly blame employers for it. Geographic Mobility1 A key reason married women have f ewer opportunities than men Husbands generally earn more than their wives to accept the best jobs available to them arises from their re- stricted mobility because they work longer hours and have made greater investments in their jobs in terms of schoolin g, training, and experience. If both spouses are searching for work, it is unlikely that the most desirable jobs for each will be in the same area. So one has to forego his or her job preference. Since the husband is the primary earner in most families, an d the wife manages the household, the wife will generally accommodate to the husband's choice if the family is to maximize its total income. Married women must also search for employment close to home because of their family responsibilities, which gives them a smaller pool of jobs from which to choose. As a result, many women earn less than they otherwise could, not because of employer discrimination, but Pear, op. cit. Victor Fuchs Recent Trends and Long-Run Prospects for Female Earnings ,It American Economic Review, May 1974, pp. 236-2 42. See Robert. H. Frank Why Women Earn Less: The Theory and Estimation of lo l1 Differential Overqualification American Economic Review, June 1978 pp. 360-373. 6 truly reflect the risks and costs for the firm, aggressive com- panies (perhaps owned by women) could improve their profits by recruiting only women. Proponents of the discrimination theory have thus far failed to explain why this does not happen In addition to the effects of employees leaving voluntarily wages a re also related to the layoff rate in an industry. Indi viduals working in an industry characterized by frequent layoffs or employment insecurity are generally paid more than workers in similar industries with a lower layoff rate, since the demand side of the wage equation reflects the risk faced by the employee because of their own choice of role in the family. found that this may explain as much as one-quarter of the wage differential between the sexes. l2 One study has I Turnover Rates13 Women tend to m o ve in and out of the labor force more often than men do, and this reduces the employability--and hence wages of women as a group. High turnover rates among women require firms to hire and train new workers more frequently, adding to .total labor costs. It is only rational, therefore, for firms to seek workers with a low probability of leaving have to make decisions on the basis of past experience. employer believes a woman is more likely than a man to leave the firm--and statistics show this to be the case - -the woman is likely to be hired only if she accepts a lower wage than a man with identical credentials to compensate the firm for the additional risk. l4 With imperfect information about the potential productivity of a worker, employers If an l2 l3 Ibid. , p. 3 70. This estimate was for Standard Metropolitan Statistical Areas with populations under 250,000. See Elizabeth M. Landes, "Sex-Differences in Wages and Employment of the Specific Capital Hypothesis Economic Inquiry, October 1977, pp 523-538, and J ames F. Ragan and Sharon Smith The Impact of Differences in Turnover Rates and Male/Female Pay Differentials Journal of Human Resources, Summer 1981, pp. 343-365. Women are 11 times more likely to drop out of the labor force than men A Test l4 are 1984, p. 104. See Illinois Commission-on the Status of Women, Minority Report, 7 Since male-dominated jobs are more closely linked to fluctuations in the economy, part of the wage gap between the sexes can be ex- plained by the wage premium paid men to compensate them for the greater risk of job loss. One study has concluded that Ithe com- bined effect of training and turnover account for between 67 and 100 percent of the differential. If Education and Traininql Reflecting their higher turnover rates, women tend to invest less than their male counterparts in productivity-increasing edu- cation and training. And since women as a group are more likely to interrupt their care e r development-=for childbearing and house hold duties--they tend to have less job expe-rience and seniority than men. This is an important factor in explaining their rela- tively lower earnings A major reason that many women do not invest as much in their own "human capital" is precisely because they expect to be absent from the workforce for extended or frequent periods, and thus have a shorter "payoff'l period over which to recoup their investment in education and training. Consequently, they are less li k ely to study for highly specialized fields, such as the sciences, medi cine, and law, because the payoff period for these disciplines is longer. By contrast, many other female-dominated fields, such as teaching and nursing, have relatively short and inexp e nsive train ing periods for extended periods. The higher turnover rate associated with women means, therefore, that firms are generally less willing to provide women with costly training to upgrade their skills induces women to work in occupations where s p ecific training is less important, thereby minimizing the investment loss as they move from job to job. It means also that women are less likely to be found in high-paying jobs that require considerable training If the traditional family roles continue to change, and women play a more important and permanent role in the labor force, these earnings differentials can be expected to narrow. Women's atti tudes toward investing in their future are already showing signs of significant change. For example, over t h e last 30 years, women's college enrollments and completion rates have steadily approached those of men.17 Nearly half of all bachelor and master's degrees in 1979-1980 were awarded to women. Women are also making major inroads into traditionally male-dom i nated I Skills depreciate when workers are out of the labor force This 0 l5 l6 Landes, op. cit p. 532 See Jacob Mincer and Soloman Polachek, "Family Investments in Human Capi tal: Earnings of Women," Journal of Political Economy, March/April 1974 pp. 576- 608 12-13 l7 The following statistics are from U.S. Bureau of Census, op. cit., pp. 0 8 fields such as dentistry, medicine, law, engineering, mathematics l and science now go to women, compared with just 5 and 10 percent, respectively, in 19 70. As these t rends continue and women in these new disci- plines mature and represent a larger proportion of the profession, future pay gaps will narrow And about-20 percent of all law and medical degrees Career Choices Since women tend to select professions that allo w them to move in and out of the labor force easily, or to work part time, they are disproportionately-but understandably-represented in jobs that afford them this flexibility. Thus, even if their education and training equals or exceeds that of men doing s imi lar jobs, wages in these female-dominated professions are never- theless lower, due to the supply and demand factors of the occupation If women want higher earniags, then they must choose work where earnings do not generally reflect the higher turnove r and other factors associated with women. As Coleman Young, Mayor of Detroit, has explained If a painter makes more than a secretary then let more women be painters."18 Working Conditions Many wage comparisons overlook other aspects of employment. For exa m ple, women may trade off wages for better working condi- tions, such as good hours and pleasant surroundings, particularly are characterized by less agreeable and more dangerous surroundings. indoor work. Many male-dominated professions, on the other hand , I Taxes Tax policy has discouraged women from entering paid employ- ment (particularly once they are married) and developing market able skills. Because the wife is usually a secondary earner, the first dollar she earns will effectively be taxed at her h usband's highest marginal tax rate. By lowering her after-tax reward for work, the progressive tax system creates an economic disincentive for women to pursue a demanding career. Social Security payroll taxes represent an additional disin- centive for wome n considering working or investing in activities that could raise their earnings potential. As the Social Security system is currently structured, a nonworking wife can receive a l8 l9 Cited in John H. Bunzel To Each According To Her Worth?" The Public In terest, Spring 1982, p. 84. See Michael J. Boskin, "The Effects of Government Expenditures and Taxes on Female Labor," American Economic Review, May 1974, pp. 251-256, and Paul McGouldrick Why Women Earn Less," Policy Review, Fall 1981, pp 63-76. 9 spouse' s benefit equal to 50 percent of her husband's own benefit. For each dollar in Social Security benefits she could earn herself therefore, a wife effectively loses a dollar of the spousels bene fit. Even if she were to earn a benefit greater than the spous e's benefit the taxbenefit ratio is still biased against her working. For example if she had the same earnings history as her husband her net extra benefits would only be half of those received by her husband, since she could have obtained half these bene fits without working at all Motivational Differences It appears that part of the difference between the earnings of the two sexes can also be attributed to their different goals in life. Explains Michael Levin, professor of philosophy at City College of N e w York Women, most especially married women, are less willing to work their way up the career ladder and tend to see their income as supplementing that of their husband, the bread winner. 112 O THE CASE OF SINGLE WOMEN those that If the wage gap really is caused by sex'discrimination, then 1 making this charge have to explain why studies have found single women earn almost the same wages as single men.21 The reason is mainly because they have similar employment charac teristics to'leave the labor force for household reasons, such as raising children or because her husband has changed jobs. So skill and experience become the dominant factors in wage setting single women nor single men, on the other hand, are likely to work as hard as married men, who often h a ve the added responsibility of supporting several dependents the single woman is less likely than the married woman I I I Neither In an examination of the wages of married and unmarried females, for instance, Hoover Institution economist Thomas Sowell not e s that unmarried men and women receive about the same in wages for the same job and same credentials. Married men tend to do better, he. says, because they are motivated to provide financial assistance for their spouses, who in turn do somehwat worse in t he market because of other duties. Sowell notes that, IISuch a 2o 21 Cited in Robert D. Hershey Women's Fight Shifts to Comparable Worth," The New York Times, November 1, 1983. See Thomas Sowell Affirmative Action Reconsidered The Public Interest Winter 1976, pp. 47-65; James Gwartney and Richard Stroup Measurement of Employment Discrimination According to Sex Southern Economic Journal 19 73. DD. 575-587: and Walter Block. "Economic Intervention,'Discrimina- I tion, and Unforseen Consequences in Discrimin ation, Affirmative Action and Equal Opportunity, edited by W. E. Block and M. A. Walker (Vancouver British Columbia: The Fraser Institute, 1981), pp. 103-125. 10 situation may not be just--but it does not result, however, from employer discrimination Il2 C ONCLUSION While the concept of equal pay for equal work is nearly uni- versally supported, the notion of I1comparable wortht1 is widely criticized--for good reason backed by no solid facts or data. Itcomparable wortht1 rests on an assumption that society s ystema tically discriminates against women on a massive scale, segre gating them into low-paying jobs simply because they are women. The fact that a disproportionate number of women work in rela tively low paid employment, however, does not prove discrimi n a- tion It rests on skimpy research and is Instead, the argument for The occupational and pay patterns of men and women can be explained without resorting to unsubstantiated claims of sex dis crimination. These patterns result from different cultural role s . The key factor is that women typically have spent about half as many years as men in !aid employment, choosing to devote more time to work in 'the home. Stemming from this underlying distinction, differences in education, training, seniority, experience , turn over, labor force participation patterns, working conditions, personal preferences, and general labor market conditions explain most, if not all, of the wage gap. If the structure of the traditional family continues to change, with younger women mov ing into fields traditionally held by men, the forces of supply and demand will narrow the earnings differential between the sexes. Adjusting to these changes in its unprejudiced and inevitable way, the market system will effect wage adjustment to the evo l ving role of women far more effectively than will the fundamentally flawed notions underpinning the doc trine of comparable worth I As Part I1 of this series shows, the Itcomparable worth alternative to the market system is likely to be devastating to the U.S. economy the law, wages will henceforth be set by judges, boards of Ilex perts and interest group pressure. When the market is distorted in this way, shortages and surpluses are inevitable, leading to ever louder demands for even more government contr ol of the labor market Should the doctrine be given the backing of Peter G. Germanis Schultz Fellow 23 Sowell, op. cit p. 56 See June O'Neill The 'Comparable Worth' Trap Wall Street Journal January 20, 1984.
