91 0 Francine D. Blau and Lawrence M. Kabhi Tre Gender Poy Gap — 843 The Gender Pay Gap Have Women Gone as Far as They Can? After half a cenrury of stabiliey in dhe eaen- ings of women relative to men, there has been a subs tive earnings since the late 1970s. One of the things that make this developmen especially dramatic and significane is that the recent ‘changes contrast markedly with che relative stability of earlier years. ‘These yost-1980 earnings changes are also interesting because, when you compare women to their male counterparts, gains have been prevalent across a wide spectrum. So, for example, at frst much of the female gains were entered on younges women, but row, while the gains may be a bit lager for younger women, women of all ages have nar- rowed the pay gap with men, The same broad progres is visible when we look at the trends in the gender pay gap by education. Less-educated women have narrowed she pay gp with less-educated men and highly edu- cared women have narrowed the pay gap swith highly educated men, ‘The exrnings gains of women are particu- Jariy remarkable because they have occurred during a period when overall wage inequality was rising. That i, the difference in pay be tween workers with high wages and workers with low wages has widened considerably cover the past 25 years orso. And yet, women, a low paid group, have nonetheless been able to narrow the pay gap witha relatively higher paid group, men. tial increase in women’s zela- Francine D. Blau aad Lawrence M. Kaha. “The Academy of Management Perpetnes (2007 ender Pay Gaps: Have Women Gone at Far a6 They Cand" 21, pp. 7-23. ‘The foregoing supports our initial observa- sion char there has been impoctans, signifi cane progress for women. On the other hand, however, there is still a gender pay gap. ‘Women continue co eacn considerably less shan men om average. It is also true that con- vergence slowed noticeably in the 1990s after women had especially gained relative 1o men Jn the 1980s. Although there were some larger gains for women in the carly 2000s, the long-run significance ofthis recene expe rience is uncleat. With the evidence suggest- ing thar convergence has slowed in recent years, the possibility arises chat che narrow- ing of the gender pay gap will noc continue into the Future. Moreover, there is evidence that although discrimination against women in the labor marker has declined, some dis- cfimination dos still continue to exist, Trends in the Gender Pay Gap In this section we look in more detail at the trends in the relative wages of women. Figure 1 presents data drawn from published gov- ‘ernment statistics on femmale-to-male carn- ings ratios of fll-rime workers. We focus on full-time workers to adjust for gender differ- ences ip hours worked. This is important be- cause women are more likely than men to ‘work part-time, Ideally we would like a mea- sure of wages or an hourly rate of pay. Un fortunately, we da not have a simiar longBleu and Lawrence M. Ki Bat — Francine 8888 Earrings Ratio (Percent) eeeaaad $e 8 6 & 8 veer Bg ‘Figure 1, Female-to-Male Earnings Ratios of Full-Time Workers 1955-2005 data series for hourly wages. Thus, we Focus here on the earnings of fullxime workers. ‘The figure gives the gender earnings ratio for ‘sso data series available from published gow- cemenent statistics. Again, both pertain t0 the relative earnings of female and male fll-time workers. Their, ee annual earnings series, is based on znnual earnings data on workers who are employed year round as well as fulltime, The second, the weekly earnings series, is based on the earnings of full-time workers over the survey week, regardless how many weeks per year the individea! works. ‘The annual ccarnings series has been available for the longest time period, 1955 to 2003; the weekly ‘earnings series as been available for a some- what shorter period, 1967 to 2003. While the exacr figure forthe gender earn ings ratio differs a bic for the ewo series, they both tell the same story in terms of che twends, Until the late 1970s oF early 1980s there was a remarkable constancy in the ratio, at around GD%. There were some year~ to-year fluctuations, but the ratio hovered around the 60% level. Indeed, if there was any discernible cend, it was a decrease in the ratio between 1955 and 1960. Then, over the 1980s, we see a period of strong, sus- tained increase in the ratio, This rising rend prevailed chrough perhaps 1990 ar 1993, de- pending on the series, However, during the 1990s, the pace of convergence in both the annual and the weekly earnings series slowed and both series behaved more erratically. The pace of change picked up again in the eatly 2000s.‘Hiowever, as noted above, the long- run significance of this recent experience is unclear, Ie may signal a resumption of a strong, long-run trend towards convergence in male-female eaenings or may prove to be of only shore ducaios Abstracting foom the differential trends cover the various subperiads and focusing on the period since the late 1970s as a whole, the gains have been quite remarkable, especially viewed in cerms of the long constancy in the gender ratio that preceded this time. So, for ‘example, based on the weekly earnings series, the gender ratio rose from 61.3% in 1978 to 79.4% in 2003. Again, much of cis increase was accomplished in a relatively share period of time, with the ratio teaching 76.8% by 1993, Of course, the 77-7996 figute remains below earnings parity. Thus, clearly all sources of the pay differential berween men and women have nor been eradicated,How do we explain these earnings gains for women? To address this question as well as (© understand why women continue to ‘earn less than men, we aeed 10 first consider the basic Factors chat explain che gender pay _gap 10 begin with Economists’ Explanations for the Gender Pay Gap The Rolo of Qualifi Diseriminstion Economists point to a number of factors that could be important in explaining the lower zarnings of women compared to men, but traditionally have focused on two primary faccors. Following Juhn, Murphy, and Pierce (1991), we eall these “gender-specific” factors in that they relate specifically to differences between women and men, either in dheir qualifications or how they are reared. With regard to qualifications, the human capital model has been especially important in pointing out the potential role played by ed- Ucation and experience The gender gap in educational arainment was never particularly large in the Uniced Stages, The biggest difference historically was that, although women were more likely t0 graduate from high school than men, they were fess likely to go on 10 college and gradu- ae education, Moreover, men tended t0 con- cencrate in career-oriented fields of study such as engineering, law. medicine and busi ‘ness that led to relatively high earnings. These ‘educational differences have decreased quite a bic in recent yeats, especially a the college level whece womer are actually now over half of college students; women have also greatly increased their representation in traditionally male professional ficlds. ‘Thus gender differ- ences in education levels have never explained a large portion of the everall gender pay gap: ‘most recently, in some samples gender differ- ences in years of schooling favor women, ‘The qualification chat has proven to be quite Jmportans is work experience because tradi- jons and The Gender Pay Gap = O45 ‘onally women moved in and out of the labor -market based on family consideraions. Before World War 11, most women lef the labor mar- ker permanently when they got married and had children. In the immediate postwar pe- riod, a pattern arose whereby older married ‘women returned to the labor marker afer their -hildven were in school or grown. Aa even big ger change has occutted in the past 20 r0 30 years a increasing numbers of women, includ- ing married women, started saying, in the labor force fairly continuously even when they hhad small children at heme. Today, even the _ajrity of women with children a year or less in age are participating in the labor force. Nonetheless, on average, women have less work experience than men and shar dffecence in qualifcations is quantitatively impartane in ‘explsining the gender pay gap. Jacob Mincer and Solomon Polachek (1974) have done especially imporaanc work in high- lighting che role of labor market experience in explaining the gender pay gap. Given the tr ditional division of abor by gender in the fam ily, women tend co accurnulate less labor waar ket experience than men. Further, because women anticipate shorter and more discontin~ uuous work lives, they have lower incentives invest in market-oriented formal education and on-the-job taining. Their resulting. smaller human capital investments lower theie carnings relative to those of men. An addi- tional way in which the traditional division of labor may disadvantage women is that the longer hoars women spend on housework may also decrease the effort they pu into their mar- kee jobs compared to men, controlling for fhours worked, and hence alse reduce theit pro- ductivity and wages (Becker, 1985), ‘To the extent that women choose occupa- tions for which on-the-job training is less im- portant, gender differences in occupations are also expected. Women may especially zvoid jobs requiring large investments in skills that ate unique to 2 particular enterprise, because ie returns to such investments are reaped only as long as one remains with that employer. At846 - Francine D. Blau and Lawrence M. Ka the same time, employers may be reluctant ta hire women for such jobs because the firm bears some of the costs of such firm-spectic training, and fears nor getting a Full recurn on that investment. However, even controlling for experience and whatever other qualifications ean readily ‘be measured, there tends co be a pay differ ence between men and women that is not ex- plained and is potentially due to discrimina- tion. Gary Becker (1971; 1st ed., 1957) has ‘been especially instrumental in developing analyses of labor marker discrimination. Al though he was looking ar differences be- ‘ween blacks and whites, the idea of preju dice and is negative consequences are readily transferable to women versus men. Becker conceptualized discriminatory preferences as the desire to maintain social distance from the discriminaced group. It may at first seem todd ro hypothesize thar men would not like to associate with women on the job when they generally live together-vith women in families. However, the issue here may be more ore of socially appropriate roles than of the desire to maintain social distance, as Becker postulated was the case with race. Standard models in economics suggest dis crimination can arise in a variery of ways. In Becker's model, disc discriminatory tscesof employers, co-workers, ‘or customers. Alternatively, in models of sta- tistical discrimination,” differences in the ‘eeatiment of men and women arise fom av- erage differences herween the two groups in the expected value of productivity (or in the reliability with which productivity may be predicted), whieh may lead empleyers to dis- criminate on the bass of chat average (see for ‘example, Aigner 8 Cain, 1977). Finally dis. criminatory exclusion of women from ‘snale” jobs can resule in an excess supply of labor in “Female” occupations, depressing wages these for otherwise equally productive workers, as in Bergmana’s (1974) “overcrowding” model. The typical approach to analyzing che sources of the gender pay gap is to estimate nination is due «© the ‘wage regressions specifying the relationship ‘beeween wages and productivity-relared ehae- actetisies for men and women. The gendet pay gap may then be statistically decomposed ro two components: one due to gender dif ferences in measured characteristics, and the other “unexplained” and potsntially due to discrimination. Such empirical studies pro- vide evidence consistent with both human capital differences and labor market dicrimi- ination in explaining the gender pay gap. However, any approach that relies om a sta- tiscical residual will be open to question a8 0 whether all the necessary explanatory variables were included in the regression. For example, ven if measured human capital characteristics can explain only 2 portion of the wage gap be- tween men and women, itis possible that un- measured group differences in qualifications may explain pare of the residual. If men are more highly endowed with respect to these ‘omitted variables then we would overestimate discrimination. Alternatively, if some of the factors controlled for in such negressions—ike cccupation and tenure with the employer— themselves reflect the impact of diserimin sion, then discrimination will be underest rated. Moreover, if women face barriers to ‘entry into certain occupations, they may have higher unmeasured productivity than men in the same jobs. This factor would also suggest an underestimate of discrimination if we con- trolled for occupation * Using the residual from 2 regression to es- timate the effects of discrimination wil also run into teouble if feedback effecs are im- portant. Even small initial discriminatory differences in wages may cumulate to large ‘ones as men and women make decisions abour human capital investmenes and ime allocation in the market and the home on che basis of these wage differentials Results of statistical scudies of the gender pay gap may nonetheless be instructive. Rep- resentative findings Stom analyses of this ype may be illustrated by resales from a recent paper of ours (Blau 8 Kahn, 2006). Usingdata from the Panel Study of Income Dy- namics (PSID), which contaias information com actual labor marker experience fora large, nationally representative sample, we found a wage differential between male and female full-time workers in 1998 of 20%. ‘The re- strition to full-time workers is designed to focus on male and female workers who are as similar as possible.* ‘The impact of gender differences in char- aceeristics on the male-female wage differen- tial is shown in Table 1. The variables con- sidered include indicators of “human capital,” that is, those relating co education and experience, a well as measures of occu- Pation, industry and union starus. (Race is alo included as a control variable, but its ef- fect is small since the proportion of each race ‘group in the full-time sample is about the same for men and women ‘As would be expected, women's lesser amount of labor marker experience Is found ro be a significant decerminane of the gender ‘wage diferensal, explaining 11% of che gen- der gap in wages. Ths reflects a 3.5 year differ «nce in fulltime experience between men and ‘women, which, chough smaller than in previ« ‘ous years, i ila substancial factor explaining the wage gap. Interestingly, women in this sub sample are found to have higher edacational acainment than men, which fas indicated by the negative sign in the table) works co ower the gender wage gap by 79. Putting this somewhat diferent, gender differences in ed- ucational anainment do noc help to explain the gender wage gap, but rather work sightiy in the opposite direction, While in the popula tion asa whole, men’s educational atainment is sll somewhat higher than women's, when we focus on a sub-sample of the population which is nor only employed, bur employed full ime, women have a sight edge. Finally, gender differences in occupation and indusry are substantial and help to ex- plain a considerable portion of the gender wage gap. Men are more likely to be in blue- collar jobs and ta work in mining, construc- ‘The Geeder Pay Ban = 147 ‘Table 1. Contribution to the Wage Differential Berween Men and Women of Differences in Measured Characteristics, 1998 Characeries Percent Explained Educational arainment 67 Labor force expericace 105 Race 24 Occupational eaegory v4 Industry category 219 Union sas 35 Unexplained aud “Td 00s ‘Wage differential (94) 203 “Sour: Calaated rom data presexed in Blau and Kahn ‘2000, tion, or durable manufacturing; they are also more likely to be in unionized employment ‘Women are more likely to be in clerical or professional jobs and co work in the service Industry. Taken together, these variables ex- plain 53% of the gender wage gap—27% for ‘occupation, 22% for industry, and an addi- tional 4% for union status Although chese findings suggest chae gen der differences in work-related characteristics are important, they also indicate shat qualif- scations are only part of the story. The pro- portion of the wage differential that isnot ex- plained by chese types of productivity-telared characteristics inchudes the impact of labor market diserimination, although as men- tioned above, the residual may also include the effects of gender differences in unmea- sured productivity levels oF non-wage aspects of jobs. In this ease, 41% of che gender cannot be explained even wher gender di ferences in education, experience, industries ‘occupations, and union status are taken into account, We can consider the results of this -scudy somewhat differently by focusing on the gender wage ratio. The actual (unad- justed”) gender wage ratio is 80%; chat is, women's wages are, on average, 80% of men's wages. IF women had the same human eapi- tal characteristics (thar is, education and ex- perience), racial composition, industry and ‘occupational distribution, and union coverageas men, the “adjusted” ratio would rise to 91% of men’s wages. Thus, while measured characteristics are imporcant, women still carn less than similar men even when all measured characteristics are taken into ac count. And, as we suggested above, including ‘controls For occupation, industry, and union status may be questionable o the extent thar they may be influenced by discrimination. Nonetheless, the residual gap, however measured, may well reflect fuetors apaet from discrimination. One thac has received par ular attention recently isthe impact of ch dren on women's wages, since evidence of a negative elect of children on wages has been ‘obzained, ever in analyses which control for labor market experience (Waldfogel, 1998). ‘The reason may be that, in the past, having a child often meant that a woman withdrew from the labor foree fora substantial period, breaking her tie to her employer and forgo- ing the returns to any firm-specific training she might have acquired, as well as any ‘wards for having made an especially good job match. Given the sharp increase in the Labor force participation af women with young children chat has eecurred since the 1960s, this factor may have been of growing impor- tance in influencing the aggregate gender gap. However, the greater availability of parental leave, legally mandated in the United Staces since 1993, may well mitigare the effect of this factor on more recent co- hots. Indeed, Waldfogel finds that che nega tive effect of children on wages is substan- tially reduced for mothers who have ‘maternity leave coverage Some studies of discrimination have taken dlifferenc approaches €o the question, thus avoiding some of the problems of craitional analyses. Fest, wo studies have applied ra ditional econometric techniques to especially homogeneous groups and employed exten sive controls for qualifications, thus mini- mizing che effecr of gender differences in un- measured productivity characteristics. Wood, Corcoran, and Courant (1993) studied grad- uates of the University of Michigan Law School classes of 1972-1975, 15 yeats after ‘graduation. The gap in pay berween women and men was relatively small ar the wutset of their careers, but 15 years lter, women grad uuates earned only 60% as mach as men Some of this difference rected choices that workers had made, including the propensity ‘of women lawyers to work shorter hours Bus even controlling for current hours worked, as well as an extensive list of worker su sand other covariates, including family status, race, location, grades while in law school, and detailed work history daca, such a5 years practiced law, months of part: time work, and type and size of employer, a male advantage of 13% remained. In a simi lar vein, Weinberger (1998) examined wage differences among recent college graduates in 1985, Her conttols included narrowly de- fined college major, college grade poine aver age, and specific educational institution at ended. She found an unexplained pay gap of 10 to 15% beeween men and women. ‘A second set of studies used an experimen- ‘al approach, Neumark (1996) anaiyzed the results of a hiring “audit” in whick male and female pseudojob seckers were given similar résumés and sent to apply for jobs waiting on tables at the same ser of Philadelphia restau rant. In high-priced restaurants, a female ap- plicane’s probability of getting an interview ‘was 40 percentage points lower than a male's and her probability of gesting an offer was 50, percentage points lower, A second study ex- amined the impact ofthe adoption of “blind auditions by symphony orchestra in which a screen is used 10 conceal the identity of the candidate (Geldin & Rouse, 2000). The -sereen substantially increased the probability that a woman would advance out of pretimi- nary rounds and be the winner in the final round, The switch to blind auditions was found 20 explain 259% of the ineresse in the percentage Female in the top five symphony orchestras in che Uniced States, fom less than 59% of all musicians in 1970 10 25% in 1996.‘Third, several recenc studies have exam- ined predictions of Beckers (1971) discrimi- nation model. Becker and others have pointed ouc thae competitive forces should reduce of eliminate discrimination in the long run because the least discriminatory firms, which hire more lower-priced female labor, would have lower costs of production and should drive the more discriminatory firms out of business. For this reason, Becker suggested char discrimination would be more severe in firms or sectors that are shielded t0 some extent from competitive pressures. Consistent with this reasoning, Hellerstein, Neumark, and ‘Troske (2002) found chat, among, planes with high levels of product market power, those employing relatively more women were more profitable. Ina sim- ilar vein, Black and Strahan (2007) report that, with the deregulation of the banking induscry beginning in the mid-1970s, the gender pay gap in banking declined as men’s wages fell by considerably more than women’s (12% vs. 3%). This suggests that during the period of regulation, banks shared the rents fostered by regulation. primarily with men. Iewas thus men who lost the most in the shife o deregulation. And, Black and Brainerd (2004) find that increasing vulnera- bility to iaternational trade reduced apparent gender wage discrimination in concentrated industries, again as predicted by Becker's (1971) model. Finally, additional evidence on discrimina- tion comes from court cases. A number of temploymene practices which explicidly dis- criminated against women used to be quite prevalent; inchuding marriage bars restricting the employment of married women (Goldin, 1990), and the intentional segregation of ‘men and women into separate job categories with associated separate and lower pay scales for women (e.g. Bowe v. Colgare-Palmolive Co., 416 F2d 71) [7eh Cir. 1969]: UE v. ‘Westinghouse Flectric Co., 631 F2d 1094 [Bed Cit, 1980)), While many such overt practices have receded, recent court cases ‘The Gender Pay Gop ~ 849 suggest chac employment practices still exist which produce discriminatory outcomes for For example, in 1994, Lucky Stores, a major grocery chain, agreed to a settlement of $107 million after Jadge Marilyn Hall Patel found chat “sex discrimination was the standard operating procedure at Lacky wich respect to placement, promotion, movement to full-time positions, and the allocation of additional hours” (Stender v. Lucky Stores, Inc. 803 F. Supp. 259; [N.D. Cal. 1992]: King 1997). And, in 2000, the U.S. Infer- mation Agency agreed to pay $508 milion to settle a ease in which the Voice of America rejected women who applied for high-paying positions in che communications field. A lawyer representing the plaintiffs said chat the women were told things like, “These jobs are only for men,” or “We're looking for a male voice” (FEDHR, 2000). & final exam- pie is the 1990 case againse Price Water- ‘house, a major accounting frm, in which the only woman considered for a partnership was denied, even though, of the 88 candi- dates for partner, she had brought in the most business. Her colleagues criticized her for being “overbearing, ‘macho’ and abrasive and said she would have a better chance of tmaking partner if she would wear makeup and jewelry, and walk, talk and dress ‘more femininely.”” The Couze found that Price ‘Waterhouse maintained a partnership eval- uation system that “permitted negative sex- ually stereotyped comments to influence partnership selection” (BNA, (990; Lewin, 1990). Oftentimes, economists serve as expert ‘witnesses in court cases alleging discrimina- tion. Their analyses, when publicly available, provide a window into discriminatory prac- tices that still exist to some extent in the labor marker, although there is of course likely to be disagreement between experts employed by each side it the type of e dence that is relevant o¢ in the interpretation of the evidence. Far example, the Lucky‘Stores case cited above generated an interest- ing exchange summarized in Taylor (2001).6 Labor economist John Pencavel estifed for the plaintiffs, che women who brought che suit, He found chat women a Lucky earned beeen 76 percent and 82 percent as much as Lucky's male workers earned. Pencavel found that women were regularly placed in jobs chat paid less than jobs given male coworkers, although there was no significant difference berween the education and expec ence of the workers. There was fle differ cence im the wages of the male ard female workers within cach type of job: but some jobs paid more chan others and wore: hap pened tobe assigned tothe lower-paying obs Joan Haworth, another labor economist, vas an expert witness for the defendant, Lucky Stores. She reported survey evidence showing that Lucky’ assignment of women and men todifferent jobs reflected differences in the work preferences of men and women. “Thus, Lucky jusifed its job asignmenss by arguing chat there was a gender difference in awiudes roward werk, Lacy argued thar is employment policies were based on observed dliferences in the carer aspirations of male and female employecs. For example, one manager ac Lucky testified that women were more interested in cash register work and men ‘were more interested in oor work ‘As we noted above, Judge Marilyn Hall Patel decided the case in favor of the plain- tis, With respect to the evidence cited above, she wrote: “The court finds the defen- dane’ explanation that the statistical dispari- ties between men and women at Lucky are caused by differences in the work interests of ‘men and women to be unpersuasive.” An teresting aspect of tis case fs ehat both agreed that male and female employees re- ce-ved equal pay for equal work and that the pay differential was asociated with pay dif- ferences across occupations. They differed, sowever, aver the source of the occupational differences, the choices of women vs, dis- ccimination, This disagreement mirrers the aleernative explanations economists offer in general for wage and occupational difference between men and women: differences in qualifications based on the choices men and women make versus discrimination which limits the opportunities and pay of women ‘compared to men. ‘Some additional evidence supporting dis- crimination as a source of che type of occu- pational differences cited above is provided bya recent study of ighe years of data from an unidentified regional grocery chain on ‘gender differences in job citles and wage races (Ransom & Oaxaca, 2005). As in the case of the Pencavel analysis summarized above, Ransom and Oaxaca find » partern of ‘gender differences in inital job assignment and upward mobility within the Rem that “generally penalized women, even when the analysis accountled] for individuals’ charac- teristics” (p. 219). While one might again dispute the reason for these differences, the authors found thar job segregation of women and men was dramatically lower in the period afier che company lost a diserim- ination suit (1984) end reached a settlement (1986) in which ie initiated affirmative ac- tion policies, This implies thae it was possi- ble to find women interested in higher-level jobs, leading one to doubt thae such segeega- Xion was entirely voluncary. These eases emphasize the role of occupa- tioral segregation by sex within firms in pro- ducing pay diffe women, Pencavel explicitly notes that there was lite difference in pay berween men and women in che same job. It is worth noting that economists and sociologists who have examined this issue across a wider range of firms have tended to come to a similar con- clusion: pay differences between men and ‘women in the same narrowly-defined oecu- pational categories within che same firm tend to be small (Blau, 1977; Groshen, 1991; Pe- tersen & Morgan, 1995: and Bayard, Heller nces between men andstein, Neumark, 8 Troske, 1999), However, even when men and women are in the same occupation, they tend to be segregated by firm, and such establishment segregation con- tebutes substantially to the gender pay gap. The Role of Wage Structure In earlier work, building on a framework suggested by Chinhui Juhn, Kevin Murphy, and Brooks Pierce (1991), we point out that there is another factor that needs to be con- sidered when analyzing gender differences in pay, and thac is what we call wage structure (Blu & Kahn, 1996 and 1997). We define wage structure as being the marker returns ro skills and the rewards for employment in particular sectors ofthe economy. Marker re- ‘ts to skills denote the premiums the mar- ket determines for being a more experienced worker of a more highly educated workes, etc. Rewards for employment in particular sectors of the economy refer to the fact that, for example, unionized workers tend to earn ‘more than comparable nonunionized work- crs or workers in some industries—durable goods, manufaccuring for exampl—may carn more than similarly-qualified workers in other industries, say services. In addition, considerable research suggests chat predomi- nantly Female occupations pay less, even con- trolling for measured personal characteristics of workers and a variety of characteristics of occupations, although the imerpreation of such results remains in some dispute? We distinguish wage steucture from gen der-specific factors because the idea is thar these are the returns to skills or the rewards for working in a particuler industry or occu- pation regardless of whether you are male oF female. Why should wage structure affect the gender pay gap? To see how, lets chink a bie more about the two fictors we discussed ear- lier gender differences in qualifications and labor discrimination, Suppose women do have less experience, on average, than men do. Then, the higher the return to experience the lrger the gender pay gap will market ‘iho Gondor Pay Gap ~ 851 be. Or, suppose that jobs staffed primarily by women do pay less than jobs staffed primar- ily by men. Then, the higher the premium for being in a male occupation the larger che gender pay gap will be. This is interesting because these markee re- turns have in fact varied over time. In the last 25 yeats or so, the marker returns co skis, like those acquired with work experience, hhave increased. So this isa factor that, taken alone, would have worked to increase the gender pay gap. The rewards co being in male occupations and industries have in- ceased a8 well, and that factor, caken alone, would have increased the pay gap as well. So, one question that we have raised in our re- search is: How have women been able co sue- cessfully swim against the tide of rising re- curns so skills and rising rewards to being in particular industries and occupations? That hhow have they managed to narrow the pay ifferencial with men in the face of the ad- vere trends in wage structure that have worked against chem? Before looking at the results of our re- search addressing these questions, le’s con- sider the issue of why the rerurns to skills have been increasing, There is a fairly broad consensus among economists (though not complete unanimity) that within countries like the United States, one of the main rei- sons thae the returns co skills have been rising is thar the demand by employers for skilled workers has been rising relative ro the de- mand for unskilled workers. Why has this coccurted? There are at lease two reasons. The cone that that we would pur the most weight cm is technological change. The information and telecommunications revoluyion has ‘worked to pur more of a premiun on skill ac Tease chus far There are other scenarios possi bile, buc thus far it has increased the demand for skilled workers compared to less skilled workers. The other reason—we would put Jess weight on icaltiough ic has also played a role Today, less skilled workers in the United States are t0 international erade.os — mncine D, Blau and Lewrence M. Kahn some extent competing against less skilled workers from around the world; many of them are available at much lower wages. Fac- tors in addition to demand shifts that appear to have alo played a role are—a decline in the union movement since unions tend to push for more egalitarian pay structures, the falling real value the miaimum wage (ad- justed for inflation, che minimum wage is ac- tually lower coday than i¢ was in the 1970s), an influx of unskilled immigrants, and a de- crease in the sate of growth of college-eda- cated workers While rising returns to skills may be hy- pothesized to widen the gender pay gap. all clse equal, it is possible thse the demand Shifis discussed abose may have favored ‘women relative co men in certain ways, and thus coneribuced 00 a decrease in the unex- plained gender pay gap (Blau & Kahn, 1997; ‘Welch, 2000). Technological change is be- lieved 0 have caused within-industry de- mand shifts that Favored white collar workers in general (Berman, Bound, & G: 1994), Given the traditional male predomi- nance in blue-collar jobs. cis shife might be expected to benefit women relative to men. Similzely, to the extent thac the spread of computer technology is an importane source of recent technological change, the observa- tion that women are more likely than men to use computers at work suggests char women as a group may have benefited from shifis in demand associated wich computerization (Aucor, Karz, & Krueger, 1998; Weinberg, 2000). Diffusion of computers likely also benefits women because computers restruc- ture work in ways that de-emphasize physical srength (Weinberg, 2000). Explaining the Trends: The 1980s Returning to the rrends in the gender pay gap—how do we explain them? To answer this question, we summarize results from Blau and Kahn (1997 and 2006). Using daca from the PSID (ve reported on some ‘of our results above), we analyzed women's wage gains over the 1980s (1979-1989), which, as we saw in Figure 1, was a period of exceptionally rapid closing of the gender wage gap. We found that higher rewards skills did indeed retard wage convergence ducing this period buc this was more than offser by improvements in gender-specific faces OF particular importance was the decline in the experience difference between men and women: the gender gap in fulltime ex- perience fell from 7.5 t0 4.6 years over this period, Shifts in major occupations played a significant role too, as the employment of women as professionals and managers rose felative ro ments, while thei relative employ- ment in clerical and service jobs fell Women's wages also increased tclative v0 men's because of deunionization (the decline cof unions), Deunionization had a larger neg- ative impact on male than female workers because men, who have tradieonally been ‘more likely chan women to be unionized, ex- perienced a larger decrease in unionization than women, Another factor that worked to increase the gender pay ratio substantially was a decrease in the “unexplained” portion of the gender differential—thar is, a decline in the pay difference becween men and women with the same measured characteris- tics (ie, experience, education, occupation, industry, and unio statu) ‘Taken rogether, changes in qualifications ‘andl in the unexplained gap worked to in- crease the gender wage ratio substantially: ‘Working in the opposite direction, however, ‘were changes in wage structure (or returns to characteristics) that favored men over ‘women during this period. OF parccular im- portance were a rise in the return to expe cence (Since woren have less of it) and in- cereases in returns co employment in industries where men are more highly repre- scared. These shifis in labor market returns by themselves would have reduced the gen- der ratio substantially. Thus, in onder for the ‘wage gap to decline, che factors favorably af-Fecting women's wages had to be large enough to more than offset the impact of un- favorable shies in returns. This was indeed the case, so that che gender pay gap did de- over the 1980s. Can we say anything abour the reasons for the decline in the unexplained gender wage gap that occurred over the 1980s? Such a shift may reflect a decline in labor market discrimination againse women, but also an upgrading of women’s unmeasured Jabor market skills, a shift in labor markee demand favoring women over men, oF changes in the composition of the labor force due to che pattern of labor force en tries or exits. Indeed all of these factors may wall have played a role, and all appear cred ible during this period, First, since women improved theit relative level of measured skills, as shown by the nar- rowing of the gap in fulltime job experi and in occupational differences between men and women, ic is plausible thar they also en- hanced their relative level of unmeasured skills, For example, women's increasing fabor force actachment may have encouraged them to acquire more on-the-job tsining or en- couraged their employers to offer them more training, Evidence also indicates thar gender differences in college major, which have been ly related to the gender wage gap among, college graduates (Brown & Coreo- ran, 1997), decreased over the 1970s and 1980s (Blau, Ferber, & Winkler, 2002}; the marketability of women’s education has probably improved. The male-female differ- ence in SAT math scores has also been de- lining, filling from 46 points in 1977 to 35 points in 1996 (Blau, Ferber, & Winkler, 2002), which cou'd be another sign of im- proved quality oF women's education Second, the aigument that discrimination against women declined in the 1980s may seem less credible than that their unmea- sured human capital characteristics im proved, since the federal government sealed back its ansidiscrimination enforcement ef- ‘The Gonder Pay Gap — 958 fore during the 1980s (Leonard, *989). However, as women increased their ccm: ment to the labor force and improved their job skills, the rationale for statistical discrim- ination against chem diminisheds thus ic is plausible thar this type of discrimination de- creased. Further, in the presence of eedback effects, employers’ revised views can generate additional increases in women's wages by raising women’ returns to investments in job qualifications and skills "To the extent that such qualifications are not fully controlled for in the statistical analysis used 10 explain the change in the gender wage gap. this may also help to account for the decline in the “unexplained” gap. Another posible reason for a decline in discrimination against women is that changes in social attitudes have made such discriminatory tastes in creasingly less acceptable. Third, the underlying labor market de. mand shifts that widened wage inequality «ner the 1980s may have favored women rel- ative to men in certain ways, and thus may have also contributed 0 adecrease in the un~ explained gender gap. Overall, manufactur- ing employment declined. In addition, there is some evidence that technological change produced within-industry demand shifts thar favored white-collar relative to blue-collar workers in general. As noced above, given the traditional male predominance in blue-collar jobs, this shife might be expected to benef women relative to men, as would increased computer use. Finally, another factor contributing to the considerable narrowing of the “unexplained” gender wage gap in the [980s appears co be favorable shifts in the composition of the fe- male labor foree. Specifcally, we found chat, controlling for the meisured characteristics mentioned earlier, the women who entered the labor force over this period tended ta be those with relatively high (unmeasured) skills. This improved the quality of the fe- male labor force and hus contributed to the narrowing of the gender wage gap.B54 — Francine D and Lawrence M. Kab 80 1970 19901985 Figure 2. ‘Trends in Female and Male Labor Force Participation Rates, 1965. 2003, Explaining the Trends: The 1980s Why did convergence in female and male wages slow aver the 1990s? Again, drawing fon our previeus work (Blau & Kahn, 2006) ‘we may suggest some tentative answers. We found thar human capital trends cannot ac- count for the slowdown: women improved their relative human capital by about the same amount in both the 1980s and the 1994s In the 1980s this upgrading consisted of rising relative experience while in the 1990s ic consisted to a lesser extent of rising relative experience and to a greater extent of increasing educational atainmenc of women relative to men. Nor did changes in wage seructure in the 1990s have a more adverse effect on women than changes in the previ- ‘ous decade—in face the impact of changing wage structure was actually more negative for women in the 1980s, Slowing convergence in men’s and women’s occupations and de- ‘ce of unionization in the 1990s was found to account for some of the slowdown, but only a small portion We found that the major reason for the slowdown in wage convergence in the 1990s was the considerably smaller narrowing of the “unexplained” gender pay gap in the 1990s compared to the 1980s. Our reason- ing above suggests that ihis could be due to slower improvement in women's wameasured ‘qualifications relative to men's in the 1990s than in the 1980s: a smaller decline in dis- ccimination against women in the 1990s than in the 1980s; 0 less favorable demand shifts for women in the 1990s than in che 1980s. Each of these Factors appears to have played a rote in explaining the observed trends. In addition, comolling for measured characteristics, female labor force entrants ‘were less skilled during the 1990s, perhaps as a result of the entry of many relatively lowe skilled, female single-family heads, Indeed, differences between the two decades in such shiffs in labor force composition were found to explain as much as 25% of the apparent slowdown in convergence in the unexplained gender pay gap in che 1990s,As we noted above, women narrowed the experience gap at a slower pace in the 90s than they did in the 80s. Figure 2 shows the twends in male and female labor force partic- ipation thae underlie this development. The most striking trend shown in che figure is that che difference in the participation rates of men and women has narrowed considerably since the starting year, 1965, This is due t a slow steady decrease in male labor Force par- ticipation combined with a much sharper and dramatic inerease in female labor force participation, The decrease in male pacticipa- tion does not appear ta be due very much to changes in gender roles. Rather it primarily reflects the fact that men are retiring at eat~ lice ages and ae staying in school longer. An- other factor has been the weakering job mar- ket for less skilled men (uhn, 1992). ‘While che daca in the figure begin in the smid-1960s, the large increases in female par- ticipation in face date back to the 1940s, In- tercstingly, the trend towards rising female labor force participation was trong and con- sistent unsil about 1996, After that the line becomes noticeably faster. Women’s partci- pation increased a bit dhrough 1997, with no further increases thereafter. How do these pariciparion trends relate £0 the average experience levels of women work- ers? Unfortunately, it is noe possible co figure this out just by looking ec participation rates. “This is because the labor force participa rate of women ext increase for either of cwo reasons ora combination of bath: On the one hhand, participation may rise because a lor of new groups of women come into the labor market. Tais tends to lower the average expe- rience of women workers because there are & lot of new entrants, On the other hand, par- ‘icipation can increase because women stay in the labor force more consistently over a pe- fiod of time, rather than moving in and out, “This works to raise average experience levels of women workers. Research has shown thac during the 1970s the average experience of women did not in- The Gender Pay Gap — 055 crease because those two factors counter bal- anced each other. There were a lor oF new en- tants and a lot of women staying in more continuously, thus average experience e- mained about the same (Goldin, 1990). In the 1980s, though, the increase in the labor force participation of women was due to more of them remaining in the labor force more consistently. And, as we have seen, the average experience of women workers rose accordingly. This suggests that che flattening. of the crend in female labor force participa- tion shown in the chart caused the gender gap in experience to decline more slowly in the 1990s than in che 1980s. Before ‘caving the subject of the participa- tion trends, i is interesting to consider the larger significance. Viewed more broadly, what the tends show is an enormous change jn gender roles and a movement away from the traditional family of breadwinner hus- ‘band and homemaker wife oa family where both husband and wife work outside the hhome, although not necessarily giving equal weight co each of their careers. Ralph Smit £1979) aalled this process a “subtle revolu- tion.” The trends suggest that this subde rev- olucion, having accomplished a great deal may be slowing down now. Is it stopping? Not necessarily. But we have reached a sicua- tion where, looking at women in the prime working ages (that is, 25 t0 55), over three quarcers of them are in the labor force. This means that female labor force participation rates in the United States are very high, al- though still below the male rates of around 90% in this age group. So ic may not be sur- prising that, oF necessity, Future participation trends will be less dramatic than past trends, Prospects for the Future Although we readily acknowledge thae pre- dicting the future is a tricky business, we cautiously offer some thoughts on the prospects for the fucure, Whae will happen to the gender pay gap in che coming years? Re- cent developments make the answer to this. Blau and Lawrence M. Kahn question particularly uncertain, As we have seen, after a petiod of consistent and sus- tained narrowing of the gender pay gap over the 1980s, convergence became more firul in the 1990s. Perhaps what we saw in the 1990s was a mere pause: perhaps we were consolidating the veally massive changes that had occurred over the preceding 10 to 20 yyears—not just in the gender pay gap but also in women’s labor force participation and in che occupations in which they work: per- haps the next 20 years will show similar re- rnewed gains on all these fronts. That could very well be, Or we may have reached a point that we are going to stay at for quite a while, big change frome the past but not so mach change in the furore. It is even possible that under certain circumstances the gender pay gap could begin to widen, returning to levels of an earlier period, While we cannot choose among these options with certainty, it may be instructive to consider each of che Factors that we have identified as influencing the gender pay gap and consider the possible fur ‘ure course of each and its likely impact on the pay gap, One of the factors influencing the trends in the gender pay gap is overall crends in wage inequality. Rising wage inequality, co the extent chac ic results from increasing te turns 0 skills like work experience that ‘women have less of than men, on average, is expected to widen the gender pay gap, all ese equal. In this respect, it has been noted that wage inequality increased less during the 1990s than during the 1980s (Katz & Autor 1999). IF this tapering off in the trend to- wards rising inequality should continue inco the future, the negative effect of this factor ‘on the pace of convergence in the gender pay ‘gap will be small (On the other hand, to the extent chat ing wage inequality was due to demand shifts that favored women relative co men, it may bbe hypothesized that such shifts, and the rel ative advaneage they may have given women relative 1o men, have also tapered ofFand are likely to be dampened in the fucure as well This is suggested by the face that the shed- dling of blue collar, manufacturing jobs was particularly pronounced in the 1980s. A closely related development, deunionization, which also disadvantaged men to a greater extent chan women, is likely to occur at a slower pace in the future as unionization rates in the privat sector have reached sing digits, giving litte scope for substantial fu ture declines, and public sector unionization remains relatively stable. While demand shifis favoring women may have slowed, so to0 has the growth ia the supply of women to the labor market. If this slower groweh supply continues into the furure, it may ‘mean that demand- and supply-side shifts offset each other and thus, taken together, do not have much effect on convergence in the sender pay gap. Although averall female labor force part ‘pation increased modesttyn the 1990s, wel- fare reforms and other government policies spurred an increase in employment among single mothers (see, for example, Meyer & Rosenbaum, 2001). The growth in participa- ‘ion among single heads, who cend on aver- age 1 be fess well educated than other women, could also have slowed wage conver- gence by shifting the composition of the fe- male labor force toward low-wage women. ‘We did indeed find some evidence consistent with this in our eadlier work (Blau. 8 Kaho, 2006), though this factor does not appear 0 be the main reason for the slowing conver- gence in the gender wage gap in che 1990s. ‘Thus it seems unlikely that a Further entry of single mothers inco the labor force in the fu- ‘ure will have a large impact on the agrregaze ‘gender pay gp. Moving toward the more traditional fac- tots of womer’s relative qualifications and the possibility of labor market diserimina- tion against them, there is lice reason to:ex- pect large changes here either. The flatteningof the growth in women's labor force pattic- ipation rates, if ic continues, suggests thar large increases in women's work experience and labor force commitment are unlikely a thougk this statement must be qualified somewhat since, as we have explained, trenas in the average experience of women cannot be inferred directly from changes in participation rates. Similariy, now that ‘women comprise the majority of college stu dents, further large gains ix the relative edu- cational attainment of women appear un- likely, though there is room for continued reductions in the gender differences in col- lege major and ar the graduate level in pro- fessional schools and Ph.D. programs in many fields Turning to labor market discrimination, ‘now that the most open and egregious forms of sex discrimination have been greatly re- duced ot climinated and discrimination in- creasingly becomes more subtle and possibly even unconscious, future large declines in discrimination in the labor market may be- come more difficule to attain, In addition, the decreases in statistical discrimination that we hypothesized as occurring in response to women’s increasing labor force attachment can be expected t0 slow as increases. in women's attachment also slow. However, there seems room for some further decrease in stastcal dseriminatior as the profound changes in gender roles that have already oc- curred continue 10 percolate through the labor marker and the lager sociery and as ad~ ditional changes continue to occur, albeit ata slower pace than in the past. And ie is likely chat even subsle barriers do change as women increasingly encer new areas and achieve suc- cess at higher levels, Putting this somewhar differenily, while the glass ceiling may not have broken completely, itis showing a lor of cracks and i likely to show more and more cxacks as time goes on “Taking all chese Factors into account, our best guess is that we are going to have fur- The Gender Pay Gap ~ 857 ther changes in the direction of conver- {gence, but most probably at a slower pace ‘Our own view is that one development that is extremely unlikely is that we will see a re- versal of the gains in relative wages and labor force participation women have expe- rienced over the past 25 to 30 years. We do hot expect a substantial widening of the male/female pay gap or labor Force partici- pation gap to occur. On the other hand, le precisely how mach narrowing we will sce in the future is an open quescion, the gender pay gap seems unlikely o vanish in the near cerm. For one thing, women continue to con- Front discrimination io the labor market, and, although its extent seems to be decreas- ing, it seems unlikely co be completely elimi- nated soon. Jn addition, at least some of the remaining pay gap is surely ted co the gender division of labor in the home, both directly through ics effect on women's labor force at- cachment and indirectly theough ies impact on the strength of statistical discrimination against women, Women still retain primary responsibility for housework and child care in most American families. However, this pattern has been changing as families spond to rising labor market apportunities for women that increase the opportunity cost of such arrangements. Further, policies chat facilitate the integration of work and family responsibilities, both voluntary and govern- rment-mandated, have become increasingly prevalent in recent years. Employers are likely to continue to expand such policies as they respond 10 the shifiing composition of the work force and a desire to retain employ- ces in whom they have made substantial in- vestments. In the longer run, the increasing availability of such policies will make ie easier for women to combine work and family, and also for men to cake on a greater share of household tasks. Finally, while our principal concern has been with the pay of women relative to men,58 ~ Francine D. Blau and Lawronce M. Kahn Table 2. Mean Earnings of Education Groups Relative to High School Graduates, 1974 and 2003 (96) 2003, Eaton ‘Men Women ‘Men Women High schook 1-3 years, 889 853 59 766 dyer, 100.0 100.0 100.0 100.0 College: 1S yous 1136 1126 128 uss Aor more yeas 155.0 12 23 " ‘Das rfc year ound, ile workers 18 years of age and er nv 2003, endian income fx yeu oF elle is computed 43 weighted 2ergeof the nadine for “ome calle, ne depe” and soa degree” Sowce 2003; PINCIM Table ofthe US. Cen Burra Caren Poplaon Survey, 2004 Annu Sci nd Esopomic Supplemens, fom percent go macrol032004peinclnew0. O00 le ?-35, fom www.ceasusgov/heincomelitne 35 hm Income Tables ~ People, trends in inequality among women show a deterioration in the relative economic status ‘of less educated women that is srikingly par- allel to similar trends in the labor marker for ‘men (see Table 2), These developments for less educated women serve to underscore the widening gap berween more and less skilled Americans of both sexes, as well as to em- phasize its broad dimensions notes 1, Becween 1980 and 1990, the average annual crease in the ratio was 1.14 percentage pois for annual earnings and -74 percentage points for weekly canings, while, herween [990 and 2000, it was only .16 percentage points for annual caring: and 42 percentage points for weekly earnings. Rel- ative earings growth in the early 2000 was more robust: beeween 2000 and 2003, the average annual increase in the ratio was .75 percentage points for and 1-14 percentage points for annual earni weekly earnings. 2, One shore term factor cal be the recession of 2001 and the relatively high enemploymene rates ‘hat lingered in its aermath. The demand for male workers tends to be more cyl sensitive than. thas for female workers due to ther greater concen tration in blue-collar jobs and durable manufseur. ing induseries '3f 3 likely one is unable wo comgleely con trol for nomwage job characteristics such as finge 1974: US. Cents Bure Historical benefits, safes jb securiy, den the residual may again noe give an accurite estimate of the extent of discrimination against worsen. We cannot say a pri- or what the effect of sich omissions is. On the one hhand, wo the exene thar men at likely to work in les safe or less secure jobs than women, such analy ‘cs may overestimate discrimination. On the ether hand, to the extent chat men have higher finge ber cic levels, an analysis of wage residuals will ander: state discrimination, To some degree, these nonwage characteristics ean be accounted for by controlling for industry and occupation, although 8 men: tioned, these controls may aso reflect exclusionary iting practices. 4In addition to gender differences in qualfica- tiens and the extene of discrimination, che gender ets ilferenal may alo be affected by the sl selection of women and men ine fal-sime employ ‘ment and, snore generally into the labor fore. In other words, those choosing to participate—or to work fulltime—inaydiller from those outside the labor force oF parr-ime workers in terms of both theis measured and nnmeasured characteristics. One possibly, for example, is tha labor force partci= panes area posiively selected group of those who have received higher wage fers. Similarly allxime workers may be more highly qualifed and more commistd to marker work. We in fact find char, at 2 point ia ime de gender pay gap i male if aly full-time workers are considered than if part-ximers ‘we examine the and nonparticipants ae include impact of changes in female and male selection into the labor force fr ends in che gender pay gup (sce bela). Other research thar has examined the caeings differential for white and black women has ‘ound hae, if seléselecion is no accounted for, the ‘ace differential is underestimaved: see Neal (2008) 5. The study cootrls for 19 occupations and 25 industries. 6. Thit quotation is fiom the Textbook Site for Principles of Mierecanumies, 3rd ed Additonal Ta fc, “Using Economics eo Explain Gender Pay Gaps. ac heap eollege Arco comfeconomice/taylorecon! Belmicrlscudens/add topia/ch02_genderpay-bnl, accesed June 28, 2002. The summary is bated on materials presented in Wests Federal Supplement (1993). 7. Se, for example, Sorensen (1990), Kilbourne, England, Farkas, Beron, and Weir (1994), and ‘Macpherton and Hirsch (1995), REFERENCES ‘Aignc.D. 8€ Cain, G. (1977). Satine theories of discrimination in labor market. Industrial and Labor Relations eviews 30, 75-187 ‘Autor, DHL, Kats, LE, & Keveger, AB, (1988). Computing incqualiy: Have compute changed the labor marker? Quarry Journa of ona 113, 1169-1214 Bayard, K, Helen, f, Neumark, D.&¢ Tose K. (1990), Newevdence on sex segregation and sex ference in wages from matched enpayee- conployer dacs. NBER Working Paper No. 7003, March Becker, GS, (1957). The economia ofdcrimintin, ed. Chicago: Univesity of Chicago Pres. 1971: Isted Becket, GS, (1985) Human capi effort and the ‘sexta division of labor. fourm of Labor Econan- ic, 3, S33-S58 Bergmann. B. (1974). Occupational segregation, ‘wags, and profit when employers discriminate by race or sex. Eastern conomi: Journal 1, 1-2 103-110, Berman, E Bound, j, & Griliches, Z. (1994) Changes in he demand of skilled labor within US, ranufsring industries Esidence from the zanial survey of manufacturing, Quits Jour tal of Economic. 109, 367-397 Black, SE, 8 Brainerd, E. 2004) The impact of ‘fohalscion on gender discrimination, Industrial Labor Relations Reviews 97, 540-559, lack, SE 8 Steahan, PE (2001). The division of ‘pols: Rent-sarng and dscrminaton in regu tated indasry. American Economic Review. 91, 814-851. The Gender Pay Gop — 988 Blau, ED. (1977), gual pay in be ofc Lexington, MA: Lexington Books. Blau, ED. (1998). Tends in the well-being of Amer- Jean women, 1970-1995, journal of Fmomic Lit centure, 36, 2-165. Blau, ED., Ferber, MA., & Winkler, AE, (2006). The economics of women men, and werk, Sih ed. “Upper Sale River, NJ Prentice-Hall, Blau, FD, & Kahn, LM. (1996). Wage structure and gender earnings differentials: An inte sional comparison. Ezonomica, 63, S29-862, Blau, FD., & Kahn, LLM. (1997). Swimming up- seream: Trends in the gender wage difletnal in the 1980s. Journal of Labor Economics 15, 1-82. Blau, FD., 8 Kahn, LLM. (2006). The US gender ‘ay gap in the 1990s: Slowing convergence, Jn dutrial Labor Relation: Review, 60, 45-66. Brown, C., & Corcoran, M. (1997). Sexbured dif. ferences in school content and the maleerale wage gap. Journal of Lalor Economies, US, 81-65, ‘Bureau of National Afbies (BNA). (1990). Daily ‘Labor Report, no. 235, December 6, pp. ALI-AL3 and Ft-FI0, Federal Human Resources Week (FEDHR), (2000, April 5). Government to pay $508 million for sex discrimination at U.S. Information Agency. Vo. 6p. 47 Goldin, C. (1990). Understanding the gender gap. [New Yorke Orford University Pres. Goldin, C. & Rouse, C. (2000). Orchestrating m= partiality: The impact of ‘blind’ auiions fe male musicians, Americon Economic Review, 9, 715-741 Geroshen, E:L, (199%), The structure of the fer male/male wage differential: I ie whe you at, what you do, oF where you work? Journal of Huraan ours, 26, 457-472 Hellerstein, J.K., Newmark, Di, & Troske, K. (2002). Market forces and sex discrimination. Journal of Human Reourcs, 37, 353-380, Juha, C. (1992). Decline af male labor marke par- ‘icipation: The role of declining matker opporau- nites, Quarterly Journal of Economics. 107. 7-21 Joho, Co, Murphy: KIM. 8 Piers, B. (1991), Ac counting for the showdown in black-white wage convergence, In M. Kosters (Ed. Worker and heir wages (pp. 107-143). Washington, DC ABI Pres. ‘8 Autor, D.H, (1999). Changes in the ‘wigestrcare and earnings inequality, In CO. Ashenlter & D. Card (Eds), Flandboak of labor860 — Francine D. Blau and Lawrence M. Kahn economics (vol. 3A, pp 1463-1955). Amscerdam: Elsevier Kilbourne, B.., England, B, Farkas, Gy Beron, K. 8 Wein, D. (1994). Returns to sk, eompensat- ing lifferendals, and gender bias: Eee of occu pational characteristics on che wages of white ‘women and men, American Journal of Sociology. 100, 689-719. King, R. (1987, April 27), Women taking action against many companies, The Timer Picayune Lewin, T (1990, May 16). Patneeship awarded to ‘woman in sex bias ease. The New Yrk Time, pp Al. AIZ. Lonard, J (1989). Women and affirmative action, Journal of Economic Penpecives, 3, 61-75. MacPherson, DA. & Hiesch, BLT. (1995). Wages tnd gender composition: Why do women's jobs pay les? Journal of Labor Economics, 13, 426-471 Myer, BD, & Rosenbaum, D.T. (2001), Welfare he cared income tax credit and the labor supply of single mothers. Quarterly Joural of Ecarmic, 116, 1063-1114 Mincer, Ju & Polachek, 5. (1974), Family inves ‘ments in human capital: Farings of women, Journal ofPoiical Econemy 82, ST6-SUB. Neal, D. (2004). The measured black-white wage ‘3p among, wemen is too smal, Journal of Poi ‘al Ecnouy 12, S\-S28, Neumark, D.M. (1996). Sex discrimination in rescauranc hiring: An audi sud. Quarterly Jour- ral of Econamis, 111, 915-941. Petersen, T, & Morgan, LA (1995). Separate and unequal Oceupation-establishment sex segrega tion and the gender wage gap. American Journal of Seioligy 10, 329-861 Ransom, M., & Oaxaca, RL. (2008), Incrafirn mo Bility and sex diffeences in pay. Industrial and Labor Relations Reviews, §8, 219-237. Smith, RLE. (1979), The movement of women ico the labor fore. In RLE, Smith (Ed.), The subtle revolsion: Women at work (pp. 1-29). Washing- ton, DC: Urban lnsccue, Sorensen, F. (1990). The crowding hypothesi and comparable worth issue. Journal of Human Re- sore, 25, 55-89. ‘Taylor, J.B. (2001). Principles of economics, 3nd Boston, MA: Houghton Milfs, Waldfogel, |. (1998). The fimily gap for young ‘women in the Unite Sates and Briain: Came rernity lee make a difference? Journal of Labor Economics, 16, 535-545. Welch, (2008), Growth in women’ relative wages and in inequalcy among men: One phenomenon tr evn? American Economic Review, 10, 444-449, Wout Federal Supplements (1993). Vol. 803, pp 1259-537, Se Paul, MN Wes Publishing Compa. Weinberg, B. (2000). Computer use and the de- ‘mand fr female workers, Jndutriel and Lebar Re lasiant Review, $3, 290-308. Weinberger, CJ. (198). Race and gender wage gaps Jin the market for recent college graduates. Inds rial Relation, 37, 67-84, Wood, R.G., Corcoran, ME, & Cowrane, P (1993), Pay differences among the highly pad: ‘The male-female earnings gap in lawyers salaries. Journal of Labor Economics, Vt, 417-441.