Women and the paradox of economic inequality in the twentieth-century.

Throughout American history, male/female has defined an enduring binary embodied in access to jobs, income, and wealth. Women's economic history shows how for centuries sex has inscribed a durable inequality into the structure of American labor markets that civil and political rights have moderated but not removed. This economic experience of women reflects the paradox of inequality in America: the coexistence of structural inequality with individual and group mobility. Women, like African Americans, have gained what T.H. Marshall labeled civil and political citizenship. No longer are they legally disenfranchised, and discrimination on account of race and gender is against the law. They have also increased their social citizenship, as represented by access to jobs and education, and women, in particular, benefit from many programs of the welfare state. Yet, they remain unequal. On the whole, they earn less than men, end up in occupational ghettos, bump up against glass ceilings, and find themselves, in relation to men, as poor as ever.

The process of internal differentiation characteristic of the history of groups defined by sex, race, or ethnicity provides the key to understanding how the paradox of inequality works. As a group, women, like blacks, experienced mobility that disrupted the processes which had systematically excluded them from access to jobs and income. But their assaults on durable categorical inequalities had their limits. Men/whites re-drew category boundaries, retaining their hold on economic advantage while women/blacks assimilated into and reproduced existing economic and occupational hierarchies among themselves. In the United States, group mobility did not challenge structural inequality--instead, it reinforced it. (1) This process of differentiation--the key to the paradox of inequality--is very important to understand. It is one of the primary ways that inequality works in America, and it poses serious issues and dilemmas for public policy.

In this article, we use the paradox of inequality to illustrate one of the mechanisms through which inequality has been, and continues to be, reproduced in modern American history. With data on occupations of the whole labor force over the century and on earnings from 1940 onward, we summarize some well known information and extend it with a fresh analysis of census microdata from 1900 to 2000 drawn from the University of Minnesota's IPUMS database. (2)

Of course, many kinds of inequality--in domestic, social, and political spheres--have shaped women's experiences. Crucial as these are, they are not, with a few exceptions, the focus of this article, which concentrates on labor market inequality and women's market work. (3) Market work is crucial because exclusion from the market has been the first source of women's economic inequality. Occupation, after all, has been the principal gateway to income. Because only a tiny percentage of people have had investments or other sources of capital with which to support themselves, most people without market work have been dependent on family and friends or on charity or the state. (4)

Even within the universe of market work, inequality is an idea with multiple dimensions. Its relations with sex, or for that matter, with race, ethnicity, or age, require examination from different vantage points--a point not appreciated or observed in most analyses of inequality. Here, then, we add a note to the methodological or conceptual study of inequality as well as to its substance. That is, we proceed by analyzing the relations among sex and inequality through four lenses: (1) participation--the share of women who worked; (2) distribution--the kind of jobs women held; (3) rewards--the relative income they received; and (4) differentiation--the distance among women on scales of occupation and earnings. We end with a brief case history of occupation and income in the banking and credit industry, which encapsulates the major trends in the history of women's inequality in the twentieth-century.

Participation

The first question is, how many women worked? As simple as it sounds, the question is fraught with ambiguity and controversy. There is, first, the problem of defining "work" for women. Limiting work to employment for pay in the labor market excludes many of women's historic economic activities, such as labor on family farms, taking in boarders and lodgers, or domestic manufacture. (5) It is clear that the addition of domestic work alters trends in women's labor force participation derived only from the census's occupational statistics. With keeping boarders considered as work, married women's labor force participation did not change very much between the beginning of the twentieth century and World War II. But measured by occupation alone, it nearly tripled. (Figure 1) (6) There is no right answer to the question these numbers pose: what is the correct way to calculate the labor force rates for women? Rather, the answer depends on the purpose. In this article, the goal is to study inequality. Therefore, the focus will be on market work, that is, work in the wage labor market, which is what the rates derived from occupation measure.

Of course, participation in market work is not an unambiguous measure. Women may have entered the wage labor market not because they wanted to but because they had to--only through wage-labor could the wives or widows of many workers assure their families adequate food, shelter, and clothing. (7) Whether women have worked on account of necessity or choice, they have found themselves excluded from the labor market, either entirely or from its best opportunities. Throughout the twentieth century, the history of women's relation to market work has been characterized by their segregation into a small fraction of jobs and by the prohibition of most married women from any labor force participation at all. Women have attempted to reconstitute labor markets along less gendered lines by attempting to break through these exclusionary practices and dismantle barriers, and their assault on labor market inequalities composes a major strand in their modern history. From whichever perspective one views women's relation to the wage-labor market, however, the theme remains inequality--whether the inequality of women forced by necessity into low-wage, dead-end jobs or excluded from participation in the labor market and access to its best rewards.

[FIGURE 1 OMITTED]

The increased participation of married women in market work constitutes the most important and dramatic development in the history of women's employment in the twentieth century. (Figure 1) In 1900, about 6 percent of married women were in the paid labor force; by 1990, the figure had multiplied nearly 10 times to 59 percent, where it remained in 2000. The first great burst of women into market work happened in the two decades after 1940. In those years a great barrier to married women's market work--the prohibition against employing married women in many white-collar industries and in teaching--began to crumble in the face of labor shortages and the demands of educated women. (8) Nonetheless, in 1960 paid work was still a minority experience for married women--3 of 10 were in the workforce. By 1980, the proportion had nearly reached a majority; by 1990 it was the norm. (9) Single women, of course, always were in the workforce more often than those who had married, but the difference between single and married women narrowed over the century. (10)

The intersection of history and experience becomes even more vivid with women's labor force participation considered by age cohorts. Only a relatively small fraction of married women born before the start of the 20th century--in terms of the data here the pre-World War I cohorts--participated in market work at any age. Among women born in 1915 and 1925--mothers of the baby boom--the situation had changed markedly. Many more of them worked, and their labor force participation increased until their late 40s or early 50s. At age 25, 20 percent of married women born in 1925 had entered market work--a fraction that swelled to 42 percent when they were 35 years old and 60 percent at age 45, when for the most part their children had left school. It was among these women born between the cusp of World War I and the middle of the Great Depression--women with more formal education than any previous cohorts--that the combination of marriage with market work at some point became an expected practice. Among married women born later in the century, market work started at younger ages, or, more accurately, did not end with marriage and childbirth. It was, in fact, nearly universal. Among the post-World War II cohort, 38 percent of 25 year old married women born in 1945, 59 percent born in 1955, and 70 percent born in 1965 worked. These were the first cohorts of women to combine motherhood of young children with paid employment. (11)

Education played a key part in the surge of married women into the workforce partly because of the increased number of jobs that demanded advanced education--health care, which, below the level of physician, women dominate, is one example--and partly because educated women sought work outside the home commensurate with their training and skills. Throughout the twentieth-century, in fact, women in the labor force had stayed in school longer than men. (12) More than half the working women born in the 1910s had at least a high school education (12 years of schooling), a fraction only reached by men born in later decades. These distinctions persisted: women workers born in the 1960s, for instance, had attended college more often than men. (Figure 2) (13) Clearly, the supply of educated women demanding employment had become a powerful force driving women's increased participation in market work. Demographic shifts also impelled women into the workforce. Later marriages, more divorces, and rising numbers of single-parents gave women new reasons to support themselves both before and during marriage. Important, too, was consumption. Women of the baby boom cohorts, raised in the prosperous 1960s and early 1970s, expected a rising standard of living. But with male wages stagnating, few men could provide it, and women's earnings became the only road to increased family prosperity--indeed, in many cases the only road to family economic survival. (14)

[FIGURE 2 OMITTED]

As a result, the American labor force looked very different at the beginning, middle, and end of the century. In 1940, only about one-quarter of the labor force consisted of women. By 2000, the proportion had reached 46 percent. In 1940, 48 percent of the female labor force consisted of women who had never been married; in 2000, that fraction had dropped to about one-quarter.

Generalizations about labor force participation rates gloss over the differences among women divided by race, ethnicity, and geography. [We discuss some of these elsewhere]. Here, however, one of them cannot be entirely put aside: the different experiences of black and white married women. Married black women, as is well known, always engaged in market work more than married white women. (15) But, in the course of the twentieth century, the differences among cohorts of white and black married women narrowed sharply as married white women entered the labor force. Among women born in 1895, 25 year old married black women were about 5 times more likely than their white counterparts to work. Among those born in 1935, the difference had dropped to 13 percent and for those born in 1975 it was 7 percent.

In one way, however, the labor market participation of white and black women went in opposite directions. This was the experience of women-on-their-own (women with an absent spouse, formerly married women, or those who had never married). In 1950 and 1960, young white women-on-their-own began to work more than black women. (Figure 3) In fact, the labor force participation of black women-on-their-own began to fall until 1970 when it turned upward. In 2000, though, it still remained below the rate for white women. Why did this happen? It did not result from the shift of black women out of agriculture. For it happened among women regardless of whether they lived on farms. (16) Nor was it because the availability of welfare allowed some unmarried black mothers to withdraw from the labor force. (17) Rather, job opportunities for black women declined when domestic labor nearly disappeared as a mass occupation. At the same time, black women more often than white lacked educational qualifications for jobs that required advanced skills, and they may have met discrimination in hiring. The problem with the discrimination argument is that racism should have been a more powerful influence earlier in the century. However, when most of the jobs open to black women were in domestic service or agriculture, race worked in their favor: these were the jobs for which they were preferred. Conversely, limited education left a large number of them less ready for the clerical and, after 1980, managerial and professional jobs, into which women were moving with astonishing speed. (See below) Indeed, black working women born before the 1940s had received much less education than their white counterparts. The great majority of black women in the labor force born in the 1890s--82 percent--had spent 8 years or less in school compared to 50 percent of white working women. (After mid-century, educational attainment soared among both white and black women, with black women making spectacular gains that greatly reduced the difference between them and white women. But until late in the century, most adult black women had far less education than white women. (18)) Instead of moving into newly opened clerical and managerial jobs, many black women found employment in contingent work, or in jobs with poor working conditions. As a consequence, they left jobs more often or were laid off more frequently than white women. (19) Adding to their problems, a great many of the rising number of unmarried black mothers lacked child care and other supports necessary to hold a job. In fact, one researcher estimates that the proportion of unmarried mothers among African American women explains from 53 percent to 68 percent of the employment gap between black and white women. (20)

[FIGURE 3 OMITTED]

2. Distribution

With exclusion from market work an index of inequality, the situation of women--notably married women--improved dramatically during the twentieth-century. But all work is not equal. Thus, the question is, what kinds of jobs did women hold? Here the picture is mixed. One can view the glass as half empty or half full. Women's occupational history, in short, reflected the paradox of inequality--mobility coupled with persistent inequalities that resulted in differentiated experiences among women divided by race, ethnicity, and class.

Consider the major trends. First, work as a domestic servant--women's major occupation outside of agriculture early in the twentieth-century--dwindled to insignificance by the end of the twentieth century. Indeed, in the late 19th and early 20th centuries, the rise of clerical work dominated the story of women's entrance into new kinds of work. It was for a very long time the only new female occupation that did not represent an extension of women's traditional work--taking care of children, tending the sick, or minding the home. (21) The proportion of working women engaged in clerical work (using the 1950 census occupational classification) rose from 9 percent in 1910 to 35 percent in 1970 before falling back to 31 percent in 1990 and 30 percent in 2000 as women entered other forms of work. (Figure 4)

The increased importance of clerical employment highlights two trends in white-collar office work: first was its feminization. In 1910, 66 percent of clerical workers were men. By 1950, the proportion had dropped to 37 percent and in 2000, it was 24 percent. Three occupational categories--bookkeepers; stenographers, typists, and secretaries; and clerks--composed the bulk of most clerical jobs, and they feminized at different rates. In 1910, 38 percent of bookkeepers, 85 percent of stenographers and typists, and 18 percent of clerks were women. Women became a majority of bookkeepers between 1910 and 1920 and of clerks in the 1950s. By 2000, 96 percent of stenographers, 89 percent of bookkeepers, and 72 percent of clerks were women. One other branch of clerical work was feminized from the start: in 1910, 90 percent of telephone operators were women--a fraction that persisted throughout the century. (22) In part, but only in part, the shift of women into clerical work responded to changes in occupational structure--the growing prominence of office work in the economy. But women's clerical employment grew at nearly twice the rate of clerical work itself. And it outstripped overall job growth among women. (23) The important point is that the feminization of the office represented a "real" shift in women's work, not simply a response to compositional changes in workforce demography or occupational structure.

[FIGURE 4 OMITTED]

Second, clerical work no longer served as the first step on the ladder into management. It became, instead, a largely permanent status identified not by decision-making but, rather, by the exercise of routinized skills. (24) In the 1880s, technological change--the widespread adoption of the typewriter which had been patented in 1868--transformed clerical work. Other new technologies, however, soon followed: "The cash register was patented in 1879, dictating machines and stereotypes in the 1880s, the mimeograph machine in 1890, the Hollerith machine which was used to tabulate the 1890 census, a full-listing adding machine ... in 1892, and the first front strike typewriter in 1895." In the 1910s, "a second great wave of office machine invention took place." A list of the machines commonly found in American offices in 1919 included 30 items, none of which had been available before the 1880s. Machines facilitated the standardization of tasks. "By the early years of the twentieth century clerical work, which had been described in 1871 as requiring 'knowledge of languages, skills in accounts, familiarity with even minute details of business, energy, promptitude, tact, delicacy of perception' had become, for the majority of office workers, routine and mechanical." (25)

However, for many office workers clerical work prior to the introduction of machines also was dull and routine, and most of the men who worked as clerks spent the bulk of their time copying correspondence and other documents. The main skill they needed was good penmanship. In the long run, office machines may have routinized clerical work, but in their early years, the skills demanded by the typewriter and adding machines were new and challenging. Thus, women did not take over clerical work because it became less skilled. More to the point was the conjunction of rising demand--the sheer increase in the number of office jobs to be filled in the late nineteenth and early twentieth century--with the availability of young women who wanted work. Young women, increasingly high school graduates, wanted work commensurate with their educations. Clerical wages were good compared to other available jobs; unemployment was low; working conditions usually were clean and safe, and the status was relatively high. (26) At the same time, young men educationally qualified for clerical jobs increasingly went on to college or sought jobs in management while employers, who could absorb only a fraction of male clerks into managerial positions, worried about what to do with them in the long run and looked for workers whose tenure on the job would be temporary. Young women, therefore, were ideal: they were cheap, available, and educated, and they understood that they would leave when they married. (27) However, argues sociologist Samuel Cohn, they did not make equal inroads into the offices of all industries. Industries where the cost of labor was low compared to the cost of capital--where clerical workers composed a small proportion of the workforce--felt less pressure to feminize their office staffs than did industries where the cost of clerical work ate into profits.

Throughout the twentieth-century, relatively few women worked in skilled manual crafts or as unskilled laborers. Semi-skilled factory work as operatives was, however, increasingly important through mid-century. (28) Self-employment also remained a minor theme in women's work experience. In 1920, 12 percent of working women were self-employed. After an increase in 1940 and 1950--probably the result of unusual conditions during the Great Depression and War--self-employment settled back to around 4 percent until the 1980s: it rose to 6 percent in 1990 and then increased to 7 percent by 2000. The stability in the fraction of self-employed women, however, masks changes in the kinds of work they performed. The first theme is differentiation: self-employed women entered a much broader array of occupations throughout the century. The second theme is the changing nature of work. In 1910, two-thirds of self-employed women were domestics of one sort or another. (29) By mid-century, domestic proprietorship had nearly disappeared, its place taken by other forms of business--for the most part, small retail concerns (46 percent) and beauty parlors (10 percent). (30) By the last decades of the century, women who supplied independent business services and temporary workers (defined as independent contractors) had risen to prominence among self-employed women. Owners of other businesses joined by independently-employed teachers of one sort or another also remained important components. (31)

Unlike self-employment, jobs in management became much more common. Between 1970 and 1990, women finally made inroads into managerial positions. "The increasing representation of women among the ranks of managers," writes sociologist Jerry Jacobs, represents "the most dramatic shift in the sex composition of an occupation since clerical work became a female-dominated field in the late 19th century." (32) Unlike the feminization of clerical work, the movement of women into managerial and professional work was facilitated by the state: the Equal Pay Act of 1963, Title VII of the 1964 Civil Rights Act and the formation of the Equal Employment Opportunities Commission, and other measures mandating affirmative action. (33) Equal Employment measures and Affirmative Action proved especially important in increasing the proportion of black women in public jobs. (34) The proportion of women working as managers multiplied roughly 3 times, from 3 percent to 9 percent across these decades. (Figure 4) As a result, women began to close the gap with men. (35) The share of managers who were men declined from 86 percent in 1950 to 61 percent in 2000. Between 1960 and 2000, the share of women among professionals (excluding teachers) increased as well--from 35 percent to 51 percent. (Women always dominated teaching: throughout the century their fraction of teachers fluctuated between 72 and 84 percent.) Thus, the share of women who were either managers or professionals more than doubled from 12 percent in 1970 to 28 percent in 2000. (Figure 4) Together with women in teaching, clerical, and sales jobs, they made up about 72 percent of the female labor force compared to about 40 percent in 1920.

Among managers and professionals, however, jobs were not divided evenly by sex. Men still dominated the best work. After a study of jobs, authority, and earnings among managers, Reskin and Roos conclude:

   Women managers were concentrated near the bottom of chains of
   command; they tended to supervise workers of their own sex,
   consistent with conventions that women should not supervise men; they
   were substantially less likely than men to exercise decision-making
   authority; and their involvement in decision making was largely
   confined to offering input into decisions that men made. (36)

Even with teachers excluded, stratification by sex among professionals remains strong. Since early in the century, women have dominated the human services professions (nurses, librarians, social workers, and so on). (Figure 5) In 1910, they composed 81 percent and in 2000, 78 percent. At the same time, while their share of the most prestigious professions rose, they remained a minority, and they also remained a minority among scientific and technical professionals, increasing their share from 1 percent in 1910 to 20 percent in 2000. They did better, however, in some other categories, although, again, ones that offered less prestige, pay, and authority. Women remained close to parity with men among arts and entertainment professionals; they increased their representation in business service professions from 0 in 1910 to 58 percent in 2000 and among technicians from 27 percent in 1920 (the first year with data on the category available) to 57 percent in 2000. (Figure 5)

Two other points about the distribution of women in jobs are important. First, sharp as these figures for occupational segregation are, they underestimate the concentration of women in various kinds of work. The reason is that they rest on broad occupational categories rather than on specific jobs. Broken down by jobs, occupational segregation appears even more intense. A study by Tomaskovic-Devey illustrates the point. Using a large sample from North Carolina, he found that "the average man in North Carolina is in an occupation that is nationally 27% female, but he holds a job that is only 8% female. The average woman is in an occupation 65% female, but a job that is 88% female." (37) These figures have huge consequences, he shows, because of their impact on earnings. Two processes--the allocation of women to segregated jobs and the devaluation of jobs dominated by women--account for 75 percent of the gap between the earnings of men and women in his sample. The other point about this history of occupational segregation is more optimistic: the movement of women into new kinds of work represents real change, not just a response to structural shifts in the economy. Sociologist Suzanne M. Bianchi calculated indexes of occupational dissimilarity between men and women for 1970, 1980, and 1990 and decomposed changes into those accounted for by desegregation and by structural shifts. She found that 75 percent of the growing similarity in the work of women and men resulted from occupational desegregation and 25 percent from structural shifts. (38)

[FIGURE 5 OMITTED]

Given the persistence of occupational segregation, it is not surprising that a number of major industries remained highly segregated by sex. Others underwent striking change. Thus, mining, construction, logging, and work in steel mills and blast furnaces remained 80 percent to 90 percent male throughout the century as did employment in the railroad industry, trucking, taxi, water, and garbage industries and automobile repairing. Women, by contrast, remained at least 40 percent of conventionally female industries, mainly in apparel and textile manufacturing, personal and household service of one form of another, some varieties of retailing, education, and hospitals (as nurses). The largest list of industries, however, consists of the ones in which women gained a substantial presence during the century, increasing at least 20 percentage points as a share of the workforce. These included some forms of manufacturing, many forms of retailing, commerce, the professions, leisure and recreation, and public administration: in short, all types of industries except those requiring heavy labor or a few, such as truck driving, where men by one means or another managed to retain a strong hold or where employers felt less compelled to replace them with women.

Women made especially large inroads into the world of commerce, a result reflected in the number of them in clerical and, after 1970, managerial occupations. In banking and credit, the fraction rose from 10 percent to 48 percent at mid-century to 69 percent in 2000; in insurance from 25 percent to 44 percent to 65 percent. There were similar spikes in security and brokerage services and in real estate, advertising, and accounting, auditing, and bookkeeping. These trends should be interpreted cautiously. To be sure, women held many more managerial jobs in these industries, but they also made up the foot soldiers--the secretaries and clerks who performed routinized work in the new white collar factories. In 2000, for instance, 63 percent of women in the banking and credit industries held clerical jobs.

What explains the inroads made by women into so many occupations after 1970? The answer, Reskin and Roos argue, lies in the intersection of changes in labor queues--the attractiveness of women as workers to employers--and job queues--the attractiveness of jobs to women. "The fundamental reason for women's disproportionate entry into new occupations," they stress, was "a shortage of male workers." In some occupations, men were unavailable because the service sector had expanded so rapidly that demand had exceeded the supply of men with the right qualifications. At the same time, often a combination of automation and reduced pay had rendered many jobs unattractive to men who looked elsewhere for work. Thus, the movement of women into new forms of work did not take jobs away from men. Women entered occupations that were expanding--where the number of openings was increasing--and/or ones that "were no longer worth preserving as male territory." (39)

In this situation, employers turned to women for a variety of reasons. Anti-discrimination legislation and law suits filed by women against employers coupled with government affirmative action regulation made discrimination costly. In some occupations, too, newly automated tasks seemed extensions of "women's work"--as in the introduction of keyboards that resembled typewriters or the emphasis on communications skills in dealing with clients. Women were sought, too, in occupations where the clients were, increasingly, other women. And, with the pressure of competition from de-regulation at home and globalized industry abroad, many employers looked for ways to cut costs. Because they would work for less pay, women suddenly became much more attractive. Women responded readily to increased employer demand. Anticipating opportunities, confident of the impact of affirmative action and public opposition to discrimination, many enrolled in M.B.A. programs and other varieties of professional training hitherto dominated by men. As women entered new forms of work, their presence created a virtuous circle that led to the employment of still more women. Employers less often preferred men, the sex-typing of many forms of work eroded, women used their informal networks to get jobs for other women, and some men "fled ... feminizing jobs ..." (40)

3. Rewards

How should we assess women's occupational and industrial changes during the twentieth-century? On the one hand, these patterns point to continued occupational segregation and to women's disproportionately low representation in the better jobs. On the other hand, they show immense change. In other words, they highlight the paradox of inequality. Women moved into a great array of occupations and industries. In large numbers, they did jobs almost never done by women before. They made special progress in the best categories of work--clerical early in the century, managerial at its close--and in attractive industries like real estate, advertising, and investment. One way to assess these changes is to follow the money. How, overall, did women's earnings compare to men's? Within occupations, did their pay approach parity with men's? Here, unfortunately, we can only provide data since 1940, the first year when the census asked for information on income. For the years since 1940, however, amid all the complex data, one finding stands out: the durability of women's inequality.

Between 1940 and 1980 the main story is the stability in women's earnings as a share of men's: (41) among full-time nonagricultural workers, 58 percent in 1940 and 55 percent in 1980. (42) (Figure 6) (Table 1) Then, in the 1980s, women's earnings moved upward. Women's median earnings as a fraction of men's rose to 67 percent. (43) However, black women's wages, which increased in the 1960s and 1970s, showed signs of reversal in the 1980s. Part of the reason was the cutback in publicly-funded jobs in which they had developed a relatively well-paying niche. (44) The gap between the earnings of men and women did not result from their occupational distribution. That is, it did not reflect women clustering in low paying jobs. In fact, in all major categories of non-agricultural work women consistently earned less than men, even though in most occupations women improved their economic position. In the same years, the fraction of women in poverty dropped sharply. But women composed a larger fraction of the poverty population, and the ratio of their poverty to men's widened. (45)

[FIGURE 6 OMITTED]

While women in managerial jobs did earn considerably more than women operatives, they still earned less than male managers--a point that highlights the limits of occupational titles for gauging economic progress. Job titles inflated to show progress toward ending sex discrimination often masked the real content of work, the relegation of women to less remunerative positions, and the glass ceilings against which they bumped. (46) Women often did not earn as much as men in less desirable or prestigious jobs. Male clerical workers often earned more than female managers and male operatives more than female crafts workers while predominantly female occupations paid less to begin with. (47)

Women's lower earnings did not result from less education. At every educational level, women earned less than men. In 1940, for instance, among individuals with 12 years of schooling, women earned 59 percent as much as men. In 1970, among workers with four or more years of college, women earned 58 percent of men's median income. In 2000, the ratio had improved, but, still, for women with four or more years of college was only 68 percent. However, for both men and women, returns to higher education increased sharply in the last half of the 20th century. Between 1940 and 2000, the ratio of high school graduates' to college graduates' median earnings decreased from 59 to 50 for women and 60 to 52 for men.

The relative earnings of younger women, however, have improved. By 2000, the ratio of women's earnings to men's had risen from 62 for women born 1946-55 to 68 and 82 for the next two cohorts. If young women continue on this trajectory--if the gain toward income parity with men persists over time--the result will herald a revolution in one of the nation's most historically durable inequalities. (Table 2) That is one way to interpret these numbers. But there is another side to the story. Young men and women started with similar earnings, but men advanced further and faster. Take women born in 1956-65 as an example. In 1980, the ratio of their earning's to men's was 82; in 1990, it had declined to 75, and in 2000 to 68. For women born in 1966-75, the ratio went down from 92 in 1990 to 82 in 2000--still higher than for women of any other age, but lower than a decade earlier. Over time, men capitalized a relatively small initial advantage into a commanding lead.

4. Differentiation

By 2000, women were arrayed in a much wider variety of jobs and industries than earlier in the century, and their earnings ranged from the top to the bottom of the income scale. Differentiation had emerged as a key theme in their economic history. As they improved their occupations and incomes, women stratified, replicating among themselves the nation's durable structure of inequality. (48) Here, too, women's experience exemplified the paradox of inequality. For instance, in 1940, women with 4 or more years of higher education earned 33 percent more than those with only a high school education; by 2000, the difference had skyrocketed to 200 percent.

Women were much less concentrated in a small number of occupations at the close of the twentieth-century than at its start. The number of occupational titles needed to account for about three-quarters of the female labor force rose from only 11 in 1910 to 18 in 2000. (49) Together, the occupations that employed at least 1 percent of women workers accounted for 94 percent of all women workers in 1910, 87 percent in 1940 and 80 percent in 2000. The two largest female occupations in 1910--private household service and farm laborers--together accounted for nearly a third of women in the labor force. In 1940, the largest occupations, operative and private household worker together accounted for 28 percent; by 2000, clerical related white-collar work, the major category, contained 14 percent of women workers. Thus, as women moved into different kinds of work, they also entered a wider array of jobs.

[FIGURE 7 OMITTED]

By the end of the twentieth-century, working women were arrayed more evenly along the income scale than ever before. Consider women who worked fulltime divided into earnings quintiles by gender for 1940 and 2000. The enduring disparity between men and women is, again, the first pattern to stand out. In each year women were found more often at the bottom, less often at the top. But in 2000, the disparities, sharp as they were, were much less than those a half century earlier. (Figure 7) An index which measures the degree to which each of the sexes is under or over represented in an occupational quintile for a given year shows the same trend (Table 3). Women's over-representation in the lowest quintiles declined while women's presence in the highest two quintiles increased. Differentiation, it is clear, characterized the economic as well as the occupational profile of twentieth-century women. (50) Economic differentiation also emerges from statistics that measure income inequality. The coefficient of variation, one such measure, traces increasing inequality among both men and women after 1980. Among men, it rose 54 percent (from 72 to 111) and among women, 30 percent (from 76 to 101). (Figure 8) Gini coefficients, another standard measure of inequality, point to similar trends. (51)

5. Banking: a case in point

The contradictory trends in women's occupational and economic history--evidence of both progress and continued inequality--emerge clearly by looking at banking and credit, an industry that feminized during the twentieth-century. In 1910, 95 percent of its employees were men; in 2000, 69 percent were women. The share of women increased steadily after 1910, reaching a majority during the 1960s. Until 1990, most of the women in banking--80 percent to 90 percent--worked in clerical jobs. And, as banking feminized, most of the industry's clerical jobs went to women, jumping from 14 percent in 1910 to 43 percent in 1920--a reflection of the second wave of feminization in clerical work. By 1950, women were nearly two-thirds and, in 1980, 85 percent. As women moved into clerical work, taking over the position of teller, men moved into management and professional positions. Men were 93 percent of managerial workers in 1920, 80 percent in 1970, and 61 percent in 1980.

Banking had been a traditional man's domain for two reasons: sex stereotypes about women's interests and mental capacities and the physical demands of the job:

   Men handled financial matters because it was assumed that women were
   not interested in such activities and furthermore women's minds were
   incapable of and unaccustomed to what was referred to as, 'doing
   figuring' and making financial transactions. Since [the] early medium
   of exchanges included heavy gold and silver commodities as well as
   currency, women were presumably unable to handle such heavy items.
   Moreover, large posting and accounting books used in banking were
   presumed difficult for women to lift. (52)

Bank jobs opened to women in the twentieth-century when the increased volume of work required many more clerks to file paperwork and record transactions. Their numbers swelled briefly during World War I and then fell back again until World War II caused another labor shortage, which brought many more women into banks as tellers, bookkeepers, and minor officers--positions previously held by men. As bankers developed new services, expanded their customer base, and introduced "sophisticated accounting and financial policies," they needed to increase both the overall size of their staffs and the number with advanced training. For the latter, they turned increasingly to college-educated men who began their banking careers in managerial-training positions, rather than, as in earlier time, as clerks and tellers. Thus, "the teller's position diminished in prestige, responsibility, skill, and advancement opportunities. And in keeping with tradition, the teller's salary remained low." Few men applied for jobs as tellers, and, "by default rather than by design," women were hired in their place. "Teller" was now "redefined as a typical 'woman's job.'"

Nonetheless, after 1970, women made striking inroads into both management and professional and technical positions in banking and credit. Until the late 1960s, in fact, women generally were not allowed in banks' managerial training programs. In the 1970s, however, de-regulation put more competitive pressure on the industry, and banks consolidated and opened many more branches. Branch managers, often, were women. During the 1970s, bank employment grew 50 percent while the number of bank managers jumped 86 percent. At the same time, the earnings and prestige of bank managers deteriorated, and young men began to look elsewhere for white-collar careers. Banks also hired more women because of the pressure of federal regulations and anti-discrimination laws and because women launched successful and highly publicized suits against them for employment discrimination. Although many women were hired or promoted into managerial positions, they were shunted primarily into the less prestigious, powerful, and remunerative managerial specialties, which included retail banking; men still filled most of the managerial jobs in commercial banking, the most attractive branch of the industry. (53)

In 1970, women composed 21 percent of bank managers. In a decade, this fraction had almost doubled to 39 percent; in the next decade it rose again to 53 percent. In 2000, it was 55 percent. In the same three decades, the share of women among professional and technical employees also increased dramatically, from 34 percent to 52 percent. By 2000, 33 percent of women in banking, compared to 63 percent of men, held managerial or professional positions. Women remained under-represented in the best work, but they had made genuine inroads into the better jobs. In this, their experience in banking and credit encapsulated the wider economic history of women, the paradox of inequality: an increase in participation, a shift upward in occupational distribution, and a differentiation that brought their social structure closer to the overall structure of inequality.

But occupational improvement did not erase disparities in earning. Within the major occupational categories in banking and credit, women continually earned less than men. Women in managerial and professional positions earned a lot more than clerical workers: in 2000; the median annual earnings (in 1990 dollars) for women managers were $25,869 compared to $13,536 for women in clerical jobs. Movement up the occupational ladder, therefore, brought substantial increases in pay for women and boosted the number of women in banking and credit earning higher salaries. But in 2000, male managers' median annual earnings were $39,856--much higher than women's in the same category of work, and as late as 1970, the median earnings of male clerical workers were only slightly less than those of women managers. In fact, between 1940 and 2000, there was little change in the ratio of women's to men's earnings: for managerial employees it hovered between 50 and 65 percent, for professionals and technicians from 56 percent to 75 percent, and for clerical workers, from 39 percent to 87 percent (a sharp increase from 73 percent in 1980). Nonetheless, younger women in banking were closing the economic distance between themselves and men. In 2000, among managers, 50-59 year old women earned 57 percent as much as men; 40-49 year olds earned 53 percent; 30-39 year olds 63 percent; and 20-29 year olds, 76 percent. Among clerical workers, the ratio jumped from 56 for 50-59 year olds to 89 among 20-29 year olds. But, over time, as elsewhere in the economy, men outpaced women in economic gains. Consider, for example, the managers born between 1956 and 1965. In 1980, the ratio of women's income to men's was 88. By 1990, it had fallen to 71 and by 2000 to 54. For managers born between 1966 and 1975, the decline was similar. The pattern even held with clerical workers. There are two ways to interpret these ratios. One is as progress; young women had finally moved toward income equality with men. The other is less hopeful: young men and women started out at similar salaries, but, over time, men moved up the economic ladder faster, leaving women increasingly behind. Perhaps women's prospects had improved, but, then again, perhaps they hadn't.

The history of gender relations traces what Charles Tilly calls a "durable inequality"--a persistent inequality among paired categories over time. These inequalities--black/white is another--retain their power through four mechanisms that Tilly terms "exploitation," "opportunity hoarding," "emulation," and "adaptation." By exploitation, he means that the more powerful member of a category uses its resources to appropriate a disproportionate share of what the less powerful member produces and perpetuates its power through any one of a number of means--legislation, work rules, or outright repression. With opportunity hoarding, the less powerful member of the category makes its peace, more or less, by finding ways to promote its own interests; it looks for ways to advance within a category rather than to break down the distinctions between them. The development of ethnic niches--where ethnic groups take hold of an occupation and pass it on to their members--is one historically very important example; another is labor union practices that favor relatives or friends for membership. The third and fourth mechanisms--emulation and adaptation--are general social processes; they can be seen at work outside of as well as within relations of inequality. Emulation is simply the tendency for individuals and groups to reproduce the organizational models with which they are familiar; adaptation refers to the ways in which members of each category build routines and social relationships that facilitate daily interaction and reinforce their interest in maintaining category boundaries. (54)

The historical record is clearer about how exploitation and opportunity hoarding shaped women's inequality in the twentieth-century than it is about the role of emulation and adaptation. Exploitation took various forms: rules that prohibited the employment of married women, actions by labor unions concerned with preserving male family wages, and hegemonic cultural ideas that assigned married women to domestic labor while devaluing the kinds of market work they performed, to name three that were particularly important. Opportunity hoarding refers to the creation of occupational niches by and for women, which is another way to describe the feminization of occupations. Excluded from most occupations, women seized what openings there were, making them their own, and filling vacancies with other women trained in new educational programs--from commercial courses in high schools to social work schools in universities--that channeled women into the expanding slots in a limited number of occupations. It took powerful exogenous forces to interrupt these processes. Actions by the state; demand induced by labor shortages, structural shifts, and technological change; supply created by higher education; economic pressures from inflation and rising standards of consumption; changes in patterns of marriage and divorce; and the militancy of women striving for equity: all these came together in the last decades of the twentieth century to mount a powerful assault on the enduring inequalities that separated the experience of women and men. And young women could count some major gains. But the structures they attacked were so powerful that victories were only partial. At the beginning of the twenty-first century, whether women would continue to modify this ancient form of categorical inequality, whether young women had finally begun to dismantle the durable walls that have limited women's economic achievements, or whether they had reached the limit of their power, remained to be seen.

In Tilly's account, the mechanisms that perpetuate durable inequalities are relentless rather than paradoxical. They do not account for the mobility of groups and individuals within the disadvantaged categories that compose social structures. They account for only one side of the paradox of inequality. They do not show why groups, such as women, remain unequal despite great individual gains in occupation, education, and income. The answer, a crucial supplement to the theory of durable inequality, lies in the process of internal differentiation--in American history, characteristic of women, African Americans, and most ethnic groups. The grooves etched deeply into hegemonic social structures provide the routes along which advancement takes place. Women assimilated into and reproduced existing economic hierarchies among themselves. For this reason, their mobility did not challenge the structure of inequality; instead it reinforced it.

There are powerful lessons in this history of the paradox of inequality. The most obvious is that access--political and civil citizenship--is not enough. Access promotes individual and group interests but does little to diminish the structures of inequality. What, then, erodes inequality itself? What would promote mobility without extreme differentiation? How can mobility be accomplished without replicating and reinforcing existing economic inequalities? The history of women--like the history of African Americans--is rich in examples of how not to meet these goals, and it poses a standard against which to analyze and evaluate "progress." But how to intervene--how to turn the history of inequality into a wholly positive story rather than a paradox--remains an excruciatingly difficult, but crucial, question for theory and policy.

Department of History

Philadelphia, PA 19104

School of Social Work

Philadelphia, PA 19104

Department of Sociology

Philadelphia, PA 19104

ENDNOTES

Research for this article was supported by grants from the Russell Sage Foundation and the Spencer Foundation.

1. Elsewhere, we have shown how poverty illustrates this theme: a very substantial decline in post-transfer poverty among women since 1940 has masked the persistence of pre-transfer poverty and left the distance between poverty rates for women and men wider than before. Michael B. Katz and Mark J. Stern, "1940s to the Present," in Gwendolyn Mink and Alice O'Connor, eds. Poverty in the United State: An Encyclopedia of History, Politics and Poverty (Santa Barbara: ABC-CLIO, 2004), pp. 33-47.

2. Steven Ruggles and Matthew Sobek et al., Integrated Public Use Microdata Series: Version 3.0 (Minneapolis, 2003). http://www.ipums.org

3. For a recent review, see Andrew Hacker, Mismatch: The Growing Gulf between Women and Men (New York, 2003).

4. For an historical overview, see Alice Kessler-Harris, Out to Work: A History of Wage-Earning Women in the United States (New York, 1982).

5. Excluding these activities has a large impact on women's rates of labor force participation, especially before 1940. For more on measurement of women's participation, see: Christine E. Bose, Women in 1900: Gateway to the Political Economy of the 20th Century (Philadelphia, 2001), 33-54; Claudia Goldin, Understanding the Gender Gap: An Economic History of American Women (New York, 1990); Matthew J. Sobek, "A Century of Work: Gender, Labor Force Participation, and Occupational Attainment in the United States, 1880-1990" (Ph.D. Dissertation, University of Minnesota, 1997). The other measurement issue relates to the source of the data--the census itself. According to Goldin, the census measures labor force participation in different ways before and after 1940. The differences do not produce dramatic variations in the results, but it should be remembered that the rates for the two periods do not have exactly the same base.

6. Figure 1 is adapted from the work of Sobek, 1997. It shows labor force participation rates for women based on census occupational data alone and with an adjustment for keeping boarders calculated by Goldin. It does not include unpaid farm labor. Some feminist historians, such as Bose, consider that all women living in farm households should be considered members of the labor force. The problem with this definition is that it loosens the boundaries of the idea of labor force to the point where they become almost nonexistent. As a consequence, it threatens to obscure important changes over time and to impede, rather than facilitate, analysis. In our subsequent writing, we will focus on domestic work, and the work that falls between home and market, as part of the analysis of family economies and strategies. (For an example of how we think about family strategy.)

7. Jacqueline Jones, Labor of Love, Labor of Sorrow: Black Women, Work and the Family from Slavery to the Present (New York, 1985) has pointed out that, for most of American history, work-on-account-of-necessity marked the experience of African-American women. Similarly, for women employed in factories early in the twentieth century, historian Leslie Tentler argues, wage work was a necessity, not the activity of choice [Tentler, Wage-Earning Women: Industrial Work and Family Life in the United States, 1900-1930 (New York, 1979)]. State mothers' pensions in the 1910's and 1920's and federal Aid to Families with Dependent Children in the 1930's, their promoters said, would permit women to remain at home with their children [see Joanne L. Goodwin, Gender and the Politics of Welfare Reform: Mothers' Pensions in Chicago, 1911-1929 (Chicago, 1997)].

8. See Samuel Cohn, The Process of Occupational Sex-Typing (Philadelphia, 1985) and Goldin, 1990.

9. James R. Wetzel, "Labor Force, Unemployment, and Earnings," State of the Union: America in the 1990s. Volume One; Economic Trends Reynolds Farley (Ed.) (New York, 1995), 63, points out that women "cut the average labor force gap with men by one third during the 1980s." (Italics in original).

10. In 1900, single women were just over 8 times more likely than married women to work; by 1940 the gap had narrowed to about 3 1/2 times; by 1990, the difference was only about 8 percent, a distance that widened to 11 percent during the century's last decade.

11. See Suzanne M. Bianchi, "Changing Economic Roles of Women and Men," In State of the Union: America in the 1990s Volume One: Economic Trends. Reynolds Farley (Ed.) (New York, 1995), 111, 117.

12. See Cynthia Engel, "Health Services Industry: Still A Job Machine?" Monthly Labor Review 122 no.3 (1999): 3-14; Wetzel 1995, 74-5.

13. In 2000, 51 percent of both men and women in the workforce had completed at least one year of college. Among women, no trend was more striking than the increase in professional degrees received after 1970. In 1970, women received 5.4 percent of law degrees; this fraction rose to 30.2 percent in 1980 and 40.8 percent in 1990. In medicine, the increase went from 8.4 percent to 23.4 percent to 33.2 percent in the same years. See Bianchi, 1995; Daphne Spain and Suzanne M. Bianchi, Balancing Act: Motherhood, Marriage, and Employment among American Women (New York, 1996).

14. There are many sources documenting the rise in inequality and the decline in male wages. Two very good ones are: Annette Bernhardt et al., Divergent Paths: Economic Mobility in the New American Labor Market (New York, 2001), Sheldon Danziger and Peter Gottschalk, America Unequal (New York and Cambridge, 1995). See also Wetzel 1995, 68-9.

15. See Jones, 1985. See also Michael B. Katz, Mark J. Stern, and Jamie J. Fader, "The New African American Inequality" Journal of American History 92:1 (June 2005): 75-108.

16. Considering women of all ages together, the rate of labor force participation among black women-on-their-own declined from a high of 77 percent in 1910 to a low of 48 percent in 1970 before turning upward to 59 percent in 2000. The rate for white women followed a more erratic pattern.

17. See Lori Reid, "Occupational Segregation, Human Capital, and Motherhood: Black Women's Higher Exit Rates from Full-Time Employment," Gender & Society 16, no.5 (2002): 728-47.

18. We discuss the comparative educational experiences of blacks and whites in more detail in Katz, Stern and Fader, New African American Inequality.

19. See Reid, 2002.

20. See Karen Christopher, "Explaining the Recent Employment Gap Between Black and White Women," Sociological Focus 29, no.3(1996): 263-80.

21. See Elyce J. Rotella, From Home to Office: U.S. Women at Work, 1870-1930 (Ann Arbor, MI, 1981); Louise A. Tilly and Joan W. Scott, Women, Work, and Family (New York, 1978). See also, Jerome P. Bjelopera, City of Clerks: Office and Sales Workers in Philadelphia, 1870-1920 (Urbana and Chicago, 2005).

22. See Venus Green, Race on the Line: Gender, Labor, and Technology in the Bell system, 1880-1980 (Durham, N.C., 2001). On another industry, see Angel Kwolek-Folland, "Gender, Self, and Work in the Life Insurance Industry, 1880-1930," Work Engendered: Toward a New History of American Labor. Ava Baron (Ed.) (Ithaca, 1991, 168-90).

23. For much of the background on clerical work, we found Rotella, 1981 to be most useful.

24. See Margery W. Davies, Woman's Place Is at the Typewriter: Office Work and Office Workers 1870-1930 (Philadelphia, 1982).

25. Rotella 1981, 67-70; Davies 1982, 9-27.

26. See Davies, 1982.

27. See Cohn, 1985; Davies, 1982.

28. The share of women working as operatives remained around 20 percent from 1910 through 1950 when its decline started. By 2000, only 7 percent of women worked as operatives. (Figure 4) Women's industrial experience thus traced the arc of the nation's industrial era.

29. They were laundress for a private household (31 percent) followed by dressmaker and seamstress outside of factories (23 percent) and boarding and lodging house keeper (13 percent).

30. We infer that most of self-employed women with the occupational title "managers, officials, and proprietors" (36 percent) owned small businesses as did saleswomen (10 percent).

31. Ownership of business services and temporary work are inferred from the occupation "stenographers, typists, and secretaries" among the self-employed; in 2000, 96 percent of self-employed women listed as "attendants, professional and personal service" were in the "educational services" industry--hence, the inference that they were independent teachers. Many may have been music teachers reclassified from the music industry in earlier censuses.

32. Jerry A. Jacobs, "Women's Entry into Management: Trends in Earnings, Authority, and Values among Salaried Managers," Gender Inequality at Work. Jerry A. Jacobs (Ed.) (Thousand Oaks, CA, 1995), 152.

33. See Kessler-Harris, 1982; Barbara F. Reskin and Patricia A. Roos, Job Queues, Gender Queues: Explaining Women's Inroads into Male Occupations (Philadelphia, 1990).

34. See Yvonne D. Newsome and F. Nii-Amoo Dodoo, "Reversal of Fortune: Explaining the Decline in Black Women's Earnings" Gender & Society 16, no. 4 (2002): 442-464.

35. In 1970, 10 percent of men and 3 percent of women were managers; in 2000, the fractions were 13 percent and 9 percent.

36. Reskin and Roos, 1990, 145.

37. Donald, Tomaskovic-Devey, "Sex Composition and Gendered Earnings Inequality" Gender Inequality at Work, Jerry A. Jacobs (Ed.) (Thousand Oaks, CA, 1995).

38. Bianchi, 1995. Baunach also finds that two-thirds of women would have to change occupations to equalize distribution [Dawn M. Baunach "Trends in Occupational Sex Segregation and Inequality, 1950 to 1990," Social Science Research 31, no. 1 (2002): 77-98]. The consensus among scholars studying women's occupational change is that most resulted from occupational desegregation rather than from structural shifts. See, for instance, David A. Cotter, Joanne M. DeFiore, Joan M. Hermsen, Brenda Marsteller Kowalewski and Reeve Vanneman, "Occupational Gender Desegregation in the 1980s," Work and Occupations 22, no. 1 (1995): 3-21, Thomas Wells, "Changes in Occupational Sex Segregation During the 1980s and 1990s," Social Science Quarterly 80 (1999):370-81.

39. Reskin and Roos, 1990, 302-3.

40. ibid, 305.

41. The numbers here compare full-time (worked at least 48 weeks) non-agricultural workers for 1940 through 1990 with full-time and part-time combined for 2000. Hours worked was not available on the 2000 sample.

42. In the run of annual figures, 1950, when women's earnings temporarily increased, appears an anomaly, the result of a postwar labor shortage.

43. "Something rather dramatic happened in the 1980s," observes Bianchi. "After at least two decades in which the ratio of women's to men's earnings fluctuated but remained at about the same level, there was a sizable increase in the ratio during the 1980s" (1995), 127.

44. See also Reskin and Roos, 1990. A number of researchers have highlighted the reduction in the earnings gender gap during the 1980s. The exact percentage change varies with the data set and method, but there is consensus on the trend. See, for example, Javed Ashraf, "Is Gender Pay Discrimination on the Wane? Evidence from Panel Data, 1968-1989," Industrial and Labor Relations Review 49, no. 3 (1996); Annette Bernhardt, Martina Morris, and Mark S. Handcock, "Women's Gains or Men's Losses? A Closer Look at the Shrinking Gender Gap in Earnings," American Journal of Sociology 101, no. 2 (1995): 302-328; June O'Neill and Solomon Polachek, "Why the Gender Gap in Wages Narrowed in the 1980s," Journal of Labor Economics 11, no. 1 (1993): 205-28. On the interpretation of trends in women's earnings, see Reynolds Farley, The New American Reality: Who We Are, How We Got Here, Where We Are Going (New York, 1996).

45. Katz and Stern, "Since the 1940s."

46. See Chloe E. Bird, "High Finance, Small Change: Women's Increased Representation in Bank Management" Job Queues, Gender Queues: Explaining Women's Inroads into Males Occupations. Barbara F. Reskin and Patricia A. Roos (Eds.) (Philadelphia, 1990); Reskin and Roos, 1990.

47. Income varied, as well, by ethnicity and labor markets. Black women and Latinas earned less than white women while gender inequality among African-Americans and between blacks and whites varied with the concentration of blacks in metropolitan labor markets. See Cohen, 1998; David A. Cotter, Joan M. Hermsen, and Reeve Vanneman. "The Effects of Occupational Gender Segregation Across Race," Sociological Quarterly 44, no. 1(2003): 17-36; Paula England, Karen Christopher, and Lori Reid. "Gender, Race, Ethnicity, and Wages," Latinas and African American Women at Work: Race, Gender, and Economic Inequality. Irene Browne (Ed.) (New York, 1999), 139-180.

48. See Frank Levy, "Incomes and Income Inequality," State of the Union: America in the 1990s. Volume One: Economic Trends. Reynolds Farley (Ed.) (New York, 1995), 15; Spain and Bianchi, 1996.

49. This account uses occupations recoded according to the 1950 census occupational classification.

50. See Bernhardt, Morris, and Handcock, 1995; Mary Bowler, "Women's Earnings: An Overview," Monthly Labor Review 122, no. 12 (1999): 13-21; Newsome and Dodoo, 2002.

51. The increase was: among white males 36 percent, white females 29 percent, black and Latino males 14 percent, Latino females 19 percent, and most sharply of all, among Asians--males 47 percent and females 33 percent. (The Asian increase probably reflects the immigration of poor refugees. The 4 percent decline in the gini for black women likely reflects the great reduction in very poorly paid agricultural and domestic employment.) Inequality is also a regional story. Among regions, inequality at the end of the 1930s was highest in the South. Everywhere, it decreased in the first post-World War II years. Over the 60 years from 1939 to 1999, incomes became more equal in the South and less equal elsewhere. As a result, distinctions among regions largely disintegrated. In 1939, for instance, the gini coefficient was .37 in New England and .41 in the Middle Atlantic compared to .48 in the East South Central and .51 in the West South Central census divisions. In 1999, in the same divisions, it was .44, .47, .45, and .47--virtually the same.

52. Jane E. Prather, "When the Girls Move In: A Sociological Analysis of the Feminization of the Bank Teller's Job," Journal of Marriage and the Family 33, no. 4 (1971): 777-82.

53. See Bird, 1990; Kessler-Harris, 1982.

54. Charles Tilly, Durable Inequality. (Berkeley and Los Angeles, 1998).

By Michael B. Katz, Mark J. Stern, Jamie J. Fader

University of Pennsylvania 

Table 1 Female to Male Ratio of Median Real Personal Earnings (1990
Dollars) for Nonfarm Workers, 1940-2000

                       1940  1950  1960  1970  1980  1990  2000

Full Time & Part Time  58.0  51.0  45.9  46.1  51.2  60.0  65.9
Full Time Only         57.7  60.8  52.7  52.9  55.0  66.7  75.0

Table 2 Female to Male Ratio of Median Real Personal Earnings (1990
Dollars) for Year-Round Non-farm Workers by Cohort and Year, 1940-2000

         1940  1950  1960  1970  1980  1990  2000

1906-15  64.6  61.2  48.9  56.5
1926-35        90.2  62.5  50.8  48.5
1936-45              66.1  62.7  49.0  52.0
1946-55                    85.3  63.5  60.8  62.2
1956-65                          82.1  75.0  67.5
1966-75                                91.7  81.7

See figure 2 for source documentation

Table 3 Index of Representativeness, Earnings Quintiles by Sex, 1940 and
2000

              1-20   21-40  41-60  61-80  81-100

1940  Male     67.3   82.1  104.8  118.8  128.5
      Female  192.8  150.8   86.4   46.7   18.9
2000  Male     71.1   83.0   95.9  114.2  136.9
      Female  136.4  121.4  105.1   82.1   53.5