首页 | 学科·学界 | 社会现象 | 文章·争鸣 | 读书 | 学者家园 | 文献服务 | 数据服务 | 中心网刊 | Blog | WIKI | 网上调查 | 社会论坛
  当前位置:首页 >> 读书 >> Re-Drawing Boundaries: Work, Households, and Gender in China >> 正文
7. Wage and Job Inequalities in the Working Lives of Men and Women in Tianjin

图书名称:Re-Drawing Boundaries: Work, Households, and Gender in China
图书作者:Barbara Entwisle and Gail E. Henderson    ISBN:
出版社:Berkeley: University of California Press    出版日期:2000年

Yanjie Bian, John R. Logan, and Xiaoling Shu

In this chapter we confront a paradox in the evolution of China's stratification system. On the one hand, income inequality has newly emerged as the most important dimension of inequality in people's life chances. As recently as the mid-1980s, bureaucratic allocation of commodities such as housing and health care seemed to have more importance than income (Henderson 1990; Walder 1986). By the 1990s, however, real incomes of urban residents doubled and income inequality rose by 50 percent (Bian and Logan 1996). In fact, a substantial literature has appeared, asking how these changes are linked to the transformation of China's economic institutions since 1978 under market reform (most recently, Bian and Logan 1996; Nee 1996; Parish and Michelson 1996; Szelényi and Kostello 1996; Walder 1996; Xie Yu and Hannum 1996).

On the other hand, gender disparities in wages seem to have continued untouched. Our previous multivariate analyses of wages in Tianjin have shown that the net effect of gender on wages, controlling for background factors and job characteristics, was essentially unchanged: a 15.2 percent male advantage in 1978 and an 18.3 percent male advantage in 1993 (Bian and Logan 1996, table 4). Further review of the Tianjin data reveals that the gross gender disparity, i.e., the effect of gender with no controls for background factors and job characteristics, was even more stable: a 26.1 percent male advantage in 1978, compared to 26.0 percent in 1993.[1]

How can one account for such stability in gender inequality during a period of rapid change in the economy and class structure as a whole? Researchers have almost universally expected market reform to change the standing of Chinese women workers, although they have not agreed on the direction of change. Let us review their arguments.

One view points to the equalizing effect of market forces as these are introduced into a formerly planned economy. In the planned economy, income inequality is derived from bureaucratic power (Szelényi 1978). However,

as power—control over resources—shifts progressively from political disposition to market institutions, there will be a change in the distribution of rewards favoring those who hold market rather than redistributive power. Opportunities are more broadly based and diverse when markets replace and augment the opportunity structure controlled by the state. (Nee 1996, 910–11)

This market-as-equalizer argument implies that, because women generally had less political power and therefore lower wages than men before reforms, market reform will tend to equalize labor market opportunities of men and women, and the gender wage gap will therefore decline (Whyte 1984; for Eastern Europe, see Einhorn 1993, 116–26; S⊘rensen and Trappe 1995). During the transition to a more market-oriented economy, wage differentials would come to depend more upon education and work experience—domains in which men have had smaller advantages over women. Hence Nee and Matthews specifically predict that “the more extensive the shift to markets and the higher the rate of economic growth,… [the greater tendency that] the gender wage gap [will] decline” (1996, 428–29).

Exactly the opposite position could be taken by those who credit the Chinese state with having equalized opportunities for women. This stateas-equalizer view is based on the Communist Party's ideological goals of liberating Chinese women after 1949 (Croll 1983; Stacey 1983). The Cultural Revolution of 1966–76 has been seen as an extreme case where the partystate forcefully pushed its equalization programs, including those for women's equalization with men, even at the expense of its economic interests (Parish 1981, 1984a; Whyte and Parish 1984). In the economic sphere, women's relative improvements in work participation, job security, and wages are seen as the coproducts of a strong, ideology-driven, and centralized party-state that achieved these goals through central planning and job assignments (K. Johnson 1976). From this perspective it might be expected that the delegation of authority to work unit managers under market reform would jeopardize women's position in the workplace.

Indeed there is some evidence that this has occurred. Social historians Honig and Hershatter (1988) have observed that work unit managers in the reform era, like employers in a capitalist economy, see women workers as less reliable (married women with children have the potential to interrupt work schedules in order to fulfill child care responsibilities), less efficient (married women are less amenable than men to overtime assignments), and more expensive (work units must provide maternity leave with full pay and child care services to married women with children) than men workers. These perceptions of women workers are not new, but in the reform era managers have greater incentives to reduce labor costs and more authority to adjust their hiring, job assignments, and wage assignments to their views of women (Ran 1988; Wu 1995) At one point even “reformminded” scholars regarded laying off women workers as a means to achieve economic efficiency and protect men's jobs (Stockman 1994), but this proposal was rejected by the still-influential All China Women's Federation. Nevertheless, in the 1990s more and more women have been asked by their work units to “stay home” (an alternative to a formal layoff) and receive a small percentage of their base salaries (Wu 1995).

Managers in public enterprises are not the only possible source of discrimination against women. Lee (1995) observes that profit-driven managers in the private sector and international joint ventures discriminate against their female employees. Discrimination also comes from the family (Davis 1995; Davis-Friedmann 1991). Forexample, in a multi province study, Entwisle et al. (1995) found that although rural women predominated in sideline family businesses before reforms, these women are surprisingly under-represented now that household businesses have become a main source of family income during the reforms.

The persistent gender wage gap in Tianjin supports neither the “market as equalizer” nor the “state as equalizer” argument. Chapters 5 and 6 in this volume have indicated that the Chinese state never equalized labor market opportunities between men and women. Further, Deng Xiaoping's market reforms were not designed to equalize gender differences. To be sure, Tianjin does not yet have a completely market economy, nor is it among the most reformed regions of China. As the third largest Chinese city with a population of 10 million, next to Shanghai (13 million) and Beijing (11 million), Tianjin is under the jurisdiction of and strongly influenced by the central government. The general wisdom is that reforms in Guangdong, Fujian, and Hainan provinces, all in the far south, were implemented earlier and more completely. Nevertheless, Tianjin belongs to the next category of coastal regions in east China, where reforms have been more aggressive than in inland regions of the northwest and southwest. What would account for the stability of the gender gap in such a place?


In this study we apply to this question a conclusion recent lyargued by Walder (1996) for stratification in general: that the market per se has no inherent implications for disparities between men and women in the labor force. Under certain assumptions (e.g., that productivity is the main determinant of wages), the market may be an equalizer. This is Nee and Matthews' (1996) assumption. Because women are no less productive than men with similar skills and experience, a free labor market should pay women and men equally. Under other assumptions, the market is not an equalizer. For example, if Chinese families favor employment of men even at the cost of forced retirement for women, the potentially large layoffs that market reform may eventually provoke in deficit-ridden work units will place women at a disadvantage. In the same way, under certain as sumptions (e.g., that state policy enforces formal equality of men and women), the state is an equalizer. Under others (e.g., if the ruling Communist Party favors membership by men, whose political authority is then translatable into higher wages), it is not. These issues cannot be resolved at the level of abstract principles. In fact, all the assumptions stated above may be at least partly correct, which means that the market—and the state too—could be both an equalizer and a source of inequality.

Therefore we proceed more cautiously to specify the sources of inequality and to see how these have affected successive cohorts of men and women as they entered the labor force and developed their work careers over the last half-century. To this end we make use of the same 1993 Tianjin survey previously analyzed by Bian and Logan (1996), dividing the sample into cohorts of men and women based on when they entered the labor force. Gender inequality potentially begins with differential preparation of young men and women (in terms of education, for example), placement in their first job, and initial wage offers for people in similar jobs. It evolves over people's careers as men and women experience different possibilities of job mobility (or pressures for retirement) and as their wages grow at different rates. All of these processes together yield the particular level of gender wage inequality that we measure at a single point in time. And they are all susceptible to change.

Respondents in the 1993 Tianjin survey entered the civilian urban labor force from 1923 to 1993. Based on this information, we have constructed five labor cohorts whose job placements and wages were influenced by the significant events and policies of the periods of their career entry and of subsequent periods: (1) pre-1949 cohort, whose job histories began prior to Communist rule; (2) 1949–57 cohort, who started their career in a socialist mixed economy of multiple property forms; (3) 1958–65 cohort, who entered the labor force when Mao's government gained total control of the urban economy through the Socialist Transformation (1956–58) and when the Great Leap Forward (1958–61) mobilized women and especially housewives to work outside home; (4) 1966–77 cohort, who acquired their first jobs during the Cultural Revolution, a political and economic crisis in which economic egalitarianism was believed to reach its zenith; and (5) 1978–93 cohort, who started their work career during the new era of market-oriented reforms under Deng Xiaoping.

People in these cohorts differ greatly in how they obtained their first jobs and in the conditions under which they have worked. The pre-1949 Tianjin economy was dominated by foreign investments, especially by Japan during its occupation of the city from 1937 to 1945. Many members of this cohort were migrants from nearby rural areas and their children. This rural-to-urban migration was restricted after January 1949, when Tianjin was liberated from the Nationalist government. However, searching for jobs was an individual action for most members of the 1949–57 cohort, because the new government did not have the capacity to implement full-scale job assignment programs until the Socialist Transformation (1956–58).

Members of the 1958–65 cohort were almost exclusively assigned jobs by the government in the state and collective sectors, but there was a small proportion (about 1–2%) who chose to work as independent laborers (e.g., shoe repairers, ice pop vendors). Many housewives were mobilized during the early years of this period to work for pay, but they were mainly assigned jobs in small collective enterprises managed by subdistrict governments. The next cohort, 1966–77, entered the labor force during the Cultural Revolution. From 1966 to 1969, graduates from all but technical schools were “sent down” to the countryside; only a small fraction could stay in the city for health or family reasons. However, the city was able to assign jobs to all youths who graduated from schools in 1970. Starting in 1971, being “sent down” resumed, but more and more youths began to be assigned jobs in the city. Finally, the 1978–93 cohort entered the economy in a period of market reform. The reform period has evolved considerably in these last 15 years: after 1986 most new workers signed temporary labor contracts (for 3–5 years) with the work units to which they were assigned, and after 1991, job assignments were abolished and replaced by emerging labor markets in Tianjin (Bian 1998). Most relevant to our concerns, work units achieved considerable autonomy in personnel and wage policies in the reform period.

If there have been changes in the extent and sources of gender inequality, these may emerge as differences among these cohorts of workers in any of the following characteristics:

  1. The backgrounds of young men and women as they enter their first job, and their achievements up to the time of their current (1993) job or retirement. Within each cohort, we will consider whether men and women entered the labor for ceat the same age, with similar education, or similar levels of Communist Party membership, and whether they subsequently had similar opportunities to improve their education or enter the Party, and similar likelihood of retirement at a given age. In a society where the consistent pressure of the state for greater equality was maintained, we might expect strong gender differences in earlier cohorts but a tendency for them to diminish among more recentcohorts.
  2. The kinds of jobs that men and women were able to enter, and their job mobility up to their present position. Several job characteristics have been found to affect wages independently of gender and other background factors. These include occupational category as well as the institutional sector and bureaucratic rank of one's work unit. Again, comparisons across cohorts provide a longitudinal perspective on the “gender-typing” of occupations and work units.
  3. The direct effects of gender on wages (after controlling for other characteristics), as well as indirect effects (through gender differences in other predictors and in the payoffs from those predictors). Cohort differences in the gender gap in initial wage can be tied directly to state policies and market conditions that prevailed at the time of job entry simply because they coincided in time. Cohort differences in the current (1993) gender gap are more complex. They depend on men's and women's starting points. (Therefore we will control for initial wage as a way of holding constant this factor.) But more important, the current wage of an older worker depends to some degree on what happened during his or her whole work career. Advantages or disadvantages incurred at any point could affect present standing. The current wage—hence, the current gender gap—for workers who entered in the most recent cohort reflects only the conditions of the reform period. Still it will be instructive to examine cohort differences in the direct and indirect effects of gender on current wage.

Details of our 1993 Tianjin sample (N? 1042) can be found in Bian and Logan (1996). Briefly, the sample was based on a multistage probability sampling procedure, including all of the city's 124 subdistricts (the lowest level of government administration in the city residential management structure). Registered households were randomly selected from each sampled subdistrict and neighborhood. The respondent was either the head of household, if available, or an adult person of 18 years or older from the household. Because heads of households are mostly male, two-thirds of the resulting sample is male. The response rate was close to 100 percent, which is typical in Chinese surveys with official authorization (P. Blau and Ruan 1990; Lin and Bian 1991). Definitions of key variables are provided below, as we report the average values for men and women in various cohorts.


Information on background characteristics (age, education, and Party membership) is presented by gender and cohort in table 7.1. Men and women of the pre-1949 cohort entered the labor force at a younger age than did post-1949 cohorts. Little gender variation in age at career entry is observed for all but the 1958–65 cohort. For this cohort, women's average age at career entry is 2.6 years older than men, which suggests that some women not in the workforce in the early 1950s were mobilized to work after 1958.

More consistent gender differences are found in education, at job entry, and at the time of the survey. There is a clear trend of increasing proportions with a high school or vocational education across cohorts, from 15–20 percent in the pre-1949 cohort to over 50 percent in the most recent cohort. Gender differences are erratic: women appear somewhat favored in the two earliest cohorts, at a disadvantage in the next two, and equal in the last one. The male advantage appears, rather, in college education. At job entry there is a 2–8 percent gap between the percentage of men and women with a college education, and the gap widens after that. For some people, attainment of higher education occurs after they enter the labor force. This educational upgrading is more prevalent for men than for women in all but the 1949–57 cohort. We believe this is due to a work unit bias in favor of men. Employees must obtain permission from their work unit to temporarily or periodically leave their jobs in order to attend school. These employees continue to be paid their regular salaries, and they receive additional financial aid from their work units to cover their tuition and related costs in school. These data suggest a work unit bias in favor of investing in men rather than women.

The third background characteristic in table 7.1 is Communist Party membership. Party membership is more common among men than women in every cohort at the time of their first job, and the percentage point difference in the party membership of men and women increases from the first to the current job. Except for the most recent cohort, for whom we hesitate to give much weight because of the strong positive association between Party membership and age, this pattern is the same for every cohort—a surprising finding especially for the Cultural Revolution cohort of 1966–77, when the Party was thought to be at the peak of its promotion of gender equality.


Up to this point we have identified gender disparities in education and Party membership, disparities that increased during the working career of

<COL align=left> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right>
TABLE 7.1 Respondents' Background Characteristics by Gender and Cohort, Tianjin, 1993
  Labor Cohort (Time Entering the Labor Force) and Gender
  Pre–1949 1949–57 1958–65 1966–77 1978–93
  Men Women Men Women Men Women Men Women Men Women
SOURCE: 1993 Tianjin Household Survey.
Number of cases 69 18 161 39 111 75 206 106 154 98
Mean age                    
 First job 17.1 16.6 22.2 21.8 19.7 22.3 20.1 20.8 21.7 22.9
 Current job 66.5 62.9 63.3 62.9 52.7 56.0 41.8 42.6 34.4 35.0
High/tech school (%)                    
 First job 14.4 22.2 17.3 23.0 36.0 22.6 28.1 25.4 57.7 57.1
 Current job 17.3 27.7 21.1 20.5 32.4 18.6 29.1 29.2 50.6 56.1
College and higher (%)                    
 First job 7.2 5.5 10.5 3.5 15.3 8.0 8.2 2.8 7.7 3.1
 Current job 13.0 5.5 15.5 12.8 25.2 16.0 17.4 10.3. 18.8 9.1
Chinese Communist Party members (%)                    
 First job 18.8 5.5 10.5 2.7 7.2 4.0 6.3 2.8 5.1 1.0
 Current job 37.6 16.6 37.2 28.2 37.8 17.3 26.2 17.9 12.3 5.1
our respondents but did not change consistently across cohorts. We turn now to their actual jobs and wages. Table 7.2 provides average values on four dimensions of people's jobs: occupation, work unit sector, work unit rank, and wage.


Occupation is coded into five categories, based on an initial three-category classification developed by Bian and Logan (1996) to indicate the worker's participation in decision making or connections to market activity. This classification is intended to be sensitive to the postulate of market transition that rewards in a redistributive system go to those with decision-making power, while rewards in a market system go to those who participate more directly in the market. Two categories of workers participate in decision making: “administrative” (people with the formal authority to make decisions for the organization) and “clerical” (people whose main duty is to assist administrators but who are strategically placed to control information and routine processing of decisions). There are also two categories of market involvement: “market activity” (occupations directly involved in marketing, such as purchasing agents and sales persons) and “market potential” (occupations involving direct contact with customers, such as medical workers and drivers, or whose function is to develop or improve technologies that suit the changing market environment, such as engineers and technicians). We use the term “other” for occupations that fit none of these categories, mostly factory workers who are not in direct contact with customers. Bian and Logan (1996) find that jobs with redistributive power (the first two categories combined) had higher wages than “other” jobs from 1978 to 1993, net of job holders' education, experience, Party membership, and work unit placements. Jobs with market connection (the second two categories combined) had substantially higher wages than “other” jobs only in 1993, when comprehensive reforms had been introduced in Tianjin's economy.

Table 7.2 shows that for the first three cohorts, women's initial job placements were more likely than men's to be in occupations with no market involvement or decision-making power, but this potential disadvantage was reversed for the two more recent cohorts. Have women therefore improved their occupational status? Taking into account changes during their work careers, we find considerable room for doubt. By the time of their current job, there is a general tendency among men and women to leave low-level occupations, but such mobility is greater for men.

The top decision-making occupations—administrators—have tended to be predominantly male for every cohort, both at job entry (by a smaller margin) and at current job (by a larger margin, especially for cohorts with

<COL align=left> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right>
TABLE 7.2 Job-Related Variables by Gender and Cohort, Tianjin, 1993
  Labor Cohort (Time Entering the Labor Force) and Gender
  Pre–1949 1949–57 1958–65 1966–77 1978–93
  Men Women Men Women Men Women Men Women Men Women
SOURCE: 1993 Tianjin Household Survey.
 Administrative occupation(%)                    
  First job 5.8 0.0 2.5 2.6 6.3 1.3 2.9 0.9 2.6 2.0
  Current job 17.4 0.0 13.1 7.7 17.1 2.6 10.2 4.7 4.6 3.1
 Clerical occupation (%)                    
  First job 8.7 11.1 12.4 15.4 8.1 5.3 8.3 17.0 8.4 13.3
  Current job 8.7 11.1 13.0 12.8 15.3 10.7 16.0 17.9 10.4 18.4
 Market activity occupation (%)                    
  First job 13.0 5.6 11.1 2.6 4.5 6.7 8.3 11.3 6.5 14.3
  Current job 10.1 5.6 13.7 5.1 5.4 6.7 10.7 18.9 11.0 14.3
 Market potential occupation (%)                    
  First job 20.3 22.2 29.2 20.5 31.5 28.0 30.1 24.5 35.1 27.6
  Current job 21.7 27.8 30.0 23.1 31.5 21.3 32.0 17.9 37.0 26.5
 Other occupation (%)                    
  First job 44.9 50.0 41.6 53.9 48.7 58.7 50.0 46.2 46.1 42.9
  Current job 40.6 55.6 32.3 48.7 30.6 58.7 31.1 40.6 36.4 37.8
Institutional Sector                    
 State sector (%)                    
  First job n.a. n.a. 86.3 82.0 90.0 82.6 85.9 71.7 81.9 64.2
  Current job 79.7 77.7 88.2 87.1 88.2 74.6 83.9 71.7 82.4 65.4
 Collective sector (%)                    
  First job n.a. n.a. 5.6 10.2 8.1 17.3 13.1 26.4 14.2 31.6
  Current job 13.0 16.6 5.6 7.6 9.0 13.3 11.6 25.4 11.6 29.5
 New sector (%).0                    
  First job n.a. n.a. 8.0 7.6 1.8 0.0 0.9 1.8 3.9 4.0
  Current job 7.2 5.5 6.2 5.1 2.7 12.0 4.3 2.8 5.8 5.1
Work Unit Rank                    
 Work Units with chu or higher rank (%)                    
  First job n.a. n.a. 44.1 33.3 50.4 28.0 40.0 26.4 40.2 38.7
  Current job 37.6 50.0 49.0 43.5 50.4 37.3 48.0 30.1 42.2 40.8
 Retirees (%) 62.3 77.7 44.1 76.9 9.9 50.6 0.0 8.4 1.3 2.0
 Average monthly wage (in yuan )                    
  First job 60.0 42.4 53.7 41.7 46.4 37.8 60.8 47.3 75.9 80.3
  Current job 288.9 229.0 278.7 227.9 290.6 226.6 287.7 222.9 259.9 218.8
more years of experience). The pattern across cohorts does not conform to any simple hypothesis about the effects of ideology or markets. One of the greatest discrepancies between men and women at first job and in mobility to current job is in the 1958–65 “socialist transition” cohort, and a clear gender gap is found even in the 1966–77 “Cultural Revolution” cohort. Thus it is hard to find any support for the view that state policy specifically favored women's chances of achieving administrative positions during those periods. A better case can be made for greater gender equality in the most recent cohort (as we also found in the distribution of “other” jobs). Women in this cohort remain almost as likely as men to be administrators in their current jobs, although here again we caution that much hinges on the process of future career mobility, which may already be favoring men in this cohort.

It is in the clerical occupations that women have a clearer and continuing (perhaps increasing) edge over men. It is important to note that these are not low-paid and low-prestige occupations. Our analysis of wages below shows that they are paid better than “other” occupations, and they are often perceived as desirable white-collar positions for men and women (note that they are positions that men often move into during their work careers). The most relevant shift across cohorts is what seems to be a greater predominance of women in this category in the two more recent cohorts, and a tendency in the 1978–93 cohort for women to move into this category (from their entry position) at a greater rate than men.

In the two market-connected categories, there seems to have been a shift over time toward female predominance in occupations with direct market activity (sales and purchasing). But occupations with “market potential” have remained predominantly male. As we shall see below, both of these market sectors have higher-than-average wages. Thus the channeling of men into one and women toward the other is neutral in its effects on gender wage disparities.

To summarize this detailed review of occupational differences, the key male advantage seems to be in administrative occupations and occupations with market potential. Women have an advantage in clerical occupations and those with market activity. In the most “political” of these categories, administration, there is little evidence that egalitarian ideology became a more potent factor in the more intensely “socialist” years of 1958–77. There is some indication of female gains in the market reform period, shown as smaller proportions of women in “other” jobs, somewhat greater access to administrative jobs, and a stronger grip on clerical and market activity jobs. But in two of these comparisons (looking at “other” and administrative jobs) there is also evidence of a male edge in mobility from first to current job, which in previous cohorts eventually resulted in clear male advantages. This consideration makes us hesitate to emphasize what would otherwise seem to be a positive result of market reform for women.

Work Units

Work unit rank is a dichotomous variable, comparing ranks of chu or higher (i.e., division, bureau, ministry) with lower ranks (i.e., department, section, or no rank). Previous studies have shown that the higher the rank, the higher the wages for workers (Bian 1994b, 167–170; Bian and Logan 1996; Walder 1992b). Work unit sector is a categorical variable, having three broad categories: (1) state sector, (2) collective sector, and (3) new sector. The new sector consists of private and semi-private forms of economic organization, including individual laborers, household business, domestic private companies, foreign firms, international joint ventures, and organizations based on mixed property arrangements. Decomposing the new sector category into subcategories is not possible because of small numbers of respondents in each subcategory. These measures apply only to the post1949 Chinese economy; therefore we exclude the pre-1949 cohort from our analysis of sectoral placement at career entry.

Overall, at career entry and current job, women are more likely than men to be found in the collective sector and to be under-represented in the state sector. This disparity is stronger in younger cohorts than in older cohorts, and it is a great disadvantage for women—it is a greater disparity and has more impact on the gender wage gap than the differences in occupation discussed above.

Surprisingly, although women also are generally in work units of lower rank than men, there is no gender difference in rank in the 1978–93 cohort. This indicates that although women in this cohort tend to be assigned jobs in collective enterprises, it is the collective enterprises at higher bureaucratic levels (managed by industrial bureaus or district governments) that have recruited new workers after 1978.

Little gender difference is observed in percentages of men and women working in the new sector of private and semi-private economic organizations. Given the mixed economy of multiple ownership forms in the early 1950s, the 1949–57 cohort had 8.0 percent of men and 7.6 percent of women who started their career in the private or semi-private sectors, but the next two cohorts had only 1–2 percent of men or women having their first jobs in these sectors because of tightening state control over the urban economy during the 1960s and 1970s. The 1978–93 cohort entered the labor force during a new era of reforms, but in Tianjin only about 4 percent of men and women have started jobs in the private or semi-private sectors. As of 1993, each cohort had increased its rate of participation in private or semi-private sectors, with men and women moving equally into these activities.


Another indicator of labor force activity shown in table 7.2 is retirement. Here, as with sectoral differences above, we find sharp gender differences. The official age of compulsory retirement is 50 for women workers, 60 for men. Cadres and professionals at high levels are allowed to continue to work until age 55 for women and age 65 for men. These policies enhance opportunities for younger people in an urban economy that might otherwise have a high rate of unemployment for young adults. It is not surprising, therefore, that in the first three cohorts (all of whom include people in their 50s and older), women are more likely than men to be retired. This difference is particularly large for the 1958–65 cohort, in which half the women but only a tenth of the men are retired. But it also appears in the younger 1966–77 cohort, where 8.4 percent of women, but no men, are retired. These latter cohorts have been affected by a new practice of dingti, or children inheriting their parents' jobs. When urban youths sent to the countryside during the Cultural Revolution flooded back to the cities during 1978 and 1979, mothers rather than fathers usually took the option to retire early in order for children to work. Although this practice was interrupted during the 1980s (Bian 1994b; Davis 1990), we learned from our field interviews in Tianjin that work units in the state and collective sectors have begun using this practice again in order to relieve unemployment among their employees' children. Women may be pressed into early retirement at reduced pay as a byproduct of market reform.


We turn finally to wages. Our measure is the total monthly wages including base salary, bonuses, and other nonsalary payments from work. There are likely to be recall errors in first-job wages—more so for members of older cohorts. However, there is little reason to believe that there would be a gender bias in recall errors, so we can focus our attention on wage differentials between men and women within each cohort. Gender differences in monthly wage are observed for first job and current job in almost all cohorts. The exception is entry wages in the 1978–93 cohort, in which women earned more than men. At career entry, women's wage as a percentage of men's increased from 71 percent for the pre-1949 cohort to 81 percent for the 1958–65 cohort, and after a slight decline in the 1966–77 cohort (78%) it reached 106 percent in 1993. So, compared to men, women's wages have steadily improved at career entry. Nevertheless, gender inequality in current wage remains about the same from older cohorts to younger ones, with women earning around 80 percent of what men earn. Comparing first job to current job, it is surprising that women in the most recent cohort already have fallen behind men, reversing what otherwise might have appeared as an achievement of the market reform period.


Tables 7.1 and 7.2 have shown persistent disadvantages for women in college education (despite an equalization of high school education rates), Communist Party membership, access to administrative jobs and jobs in state sector enterprises, high-ranking work units (except for the most recent cohort), and retirement. These disadvantages in themselves could account for the lower wages of women (except for initial wage for the most recent cohort). We wish to look more closely at the wage determination process to answer the following questions: (1) Is there a direct effect of gender on wages net of background factors and job characteristics? (2) Does this gender effect vary among cohorts, either at first wage (where it is clearly linked to a particular historical period) or in current wage? (3) To what degree is the gender wage gap attributable to gender differences in average values on other predictors, and to what degree is it due to different returns to those other factors? Because there is a relatively small sample size in each cohort, our discussion will emphasize not specific coefficients, but sets of predictors taken together.

First Wage

We begin with wage at job entry. We first regress initial wage on gender only. This is the “gross” effect—no other variables are controlled. Then, we add age, education, and Party membership, and the occupation, employment sector, and unit rank of the first job to the equation. The effect of gender controlling for these other variables is the “net” effect.[2] Results are presented in the first section of table 7.3.[3] Gender is coded women? 1, men? 0. Negative coefficients in the table show the monthly deficit for women, in yuan; positive coefficients show a wage advantage.

At career entry, women's wages are lower than men's in all but the most recent cohort, but for the most part, these differences are small and not statistically significant.[4] Taking into account gender differences in men's and women's background characteristics and in jobs and work units (hereafter “gender differences in levels”) makes little difference in gender wage gaps.

Current Wage

A different pattern is revealed for gender differences in current wages, in the second section of table 7.3. We first regress current wages on gender

<COL align=left> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right> <COL align=right>
TABLE 7.3 Gender Effects on Monthly Wages (in yuan ) for First and Current Jobs
  Total Pre–1949 1949–57 1958–65 1966–77 1978–93
SOURCE: 1993 Tianjin Household Survey.
First job            
 Gender effect with no controls −6 −18 −12 −9 −14[∗] 5
 Gender effect net of controls[a] −5 −8 −9 −4 −15[∗] 5
Current job            
 Gender effect with no controls −57[∗] −60 −51[∗] −64[∗] −65[∗] −41[∗]
 Gender effect with first wage controlled −54[∗] −62 −49[∗] −50[∗] −58[∗] −43[∗]
 Gender effect with all controls but retirement −37[∗] −17 −37[∗] −39[∗] −36[∗] −32[∗]
 Gender effect net of all controls[b] −28[∗] −21 −20 −22 −31[∗] −32[∗]
with no controls. Second, we add wage of the first job as a control variable. This means that we are estimating the gender effect on change from first to current wage. Third, we add background variables (education, Party membership, seniority—years worked), and characteristics of the current job (occupation, work unit, rank). Fourth, we add retirement. We considered as predictors marital status, whether respondents have children, and whether respondents had been sent to the countryside or military prior to their first civilian urban job. None of these variables has an independent effect on current wages and therefore we removed them from the final equations.

Women's current wages are significantly less than men's for all cohorts except the earliest. This is not due to women's lower wage at career entry, but due mainly to slower wage growth. After first wage is taken into account, the gender gap in current wage remains statistically significant and is almost as large as that found when first wage is not controlled for. But, the wage gap substantially narrows when gender differences in background factors and job characteristics are taken into account. For the 1978–93 cohort, despite the fact that at career entry women on average makes lightly more than men, in current wage, women make 32 yuan (or 12%) less than comparable men per month. Given the relatively short period from first to current job for most members of this cohort, women's relative wage reduction to men is rather a significant change during the reform era. Women's disadvantage in wage growth in this and other cohorts is mainly because of their lower levels of college education and Party membership, and their under-representation in jobs and sectors with wage advantages. An additional disadvantage comes from the significantly lower returns for women than men working in the new sector of private and semi-private organizations. Finally, after retirement is taken into account, gender wage gaps in the three older cohorts, to which retirement is most applicable, become statistically insignificant. Gender differences in retirement are thus a most important source of the gender gap in current wage.

Predicted Gender Wage Gaps

Table 7.4 presents predicted average monthly wages at current jobs for men and women by cohort, taking into account average levels of the predictor variables. Men's wages are predicted by applying the coefficient estimates for men to the mean levels of the predicted variables for men (tables 7.1 and 7.2). Women's wages are predicted analogously. The gender wage gaps approximate those seen in the actual means in table 7.2, as they should, given the procedure we have followed. Overall, men average 280 yuan per month, compared to an average monthly wage of 223 yuan for women (80% of men's).

How much of the wage gap is due to different background characteristics of men and women, and to differences in the jobs they currently hold? To answer this question, we perform a kind of thought experiment, where we artificially give women the characteristics and jobs that men enjoy. If women retired at ages equivalent to men, had men's levels of human and political capital, were placed in the same work unit ranks and work unit sectors as men, and hadmen's occupations, women's wage levels would improve to 252 yuan (still only 90% of men's wage). How much of the wage gap is due to differential returns (e.g., the premium attached to a college education, or Party membership)? Again we perform a thought experiment, this time retaining women's actual background characteristics and jobs, but giving them the returns that men enjoy. With men's rates of wage returns for seniority, human and political capital, and job placements, women's wage would still be 10 percent less than men's. Clearly, differential earnings payoff and gender differences in the predictors both contribute to the gender gap in pay.

<COL align=left> <COL align=middle> <COL align=middle> <COL align=middle> <COL align=middle> <COL align=middle> <COL align=middle>
TABLE 7.4 Predicted Monthly Wage (in yuan ) for Current Jobs
  Total Pre–1949 1949–57 1958–65 1966–77 1978–93
SOURCE: 1993 Tianjin Household Survey.
Predicted Wage            
 Men 280 289 279 291 288 260
 Women 223 230 228 227 223 219
 Women, assuming men's characteristics and jobs 252 348 255 265 241 231
 Women, assuming men's rates of return 253 251 245 255 260 253


One mechanism whereby men and women can improve their wage levels is job mobility. In discussing table 7.2 we have already described some of the differences in career advancement between men and women. We return to them more explicitly here (see table 7.5). Career advancement was not a common reason given for wanting to change jobs before the reforms (Davis 1990), but it has become a driving force for job mobility during the post1978 period. In this section, we analyze gender differences in occupational change and work unit change from first job to current job.

Of the 1037 respondents, 41 percent have never changed either occupation or work unit after entry into the labor force. But women were 10 percentage points more likely than men to continue to work in their initial occupations and work units. Women, therefore, have less mobility. Of those who ever changed jobs (58%), most (three-quarters) did so after 1978, rather than before (this is most true of the younger cohorts, of course). This should be interpreted as a consequence of the loosening up of labor control by the government and by work units. However, this policy shift did not have an equal effect on men and women: 46 percent of men changed their jobs after 1978 compared to 37 percent of women—a significant difference.

Overall, slightly more than 60 percent of men and slightly more than 50 percent of women have ever changed jobs. Gender differences are concentrated among those who changed both occupations and work units: 34 percent of men, compared to 25 percent of women, are in this group. Women are not only generally less mobile than men, but they are less able to change occupations and work units simultaneously. These findings are replicated in a more elaborate analysis (not presented in table 7.5). Controlling for education, Party membership, job type, work unit rank, and work unit sector, compared to men whose mobility rate is set to be 100, women have an overall job mobility rate of 64, but their rate of moving between

<COL align=left> <COL align=middle> <COL align=middle> <COL align=middle>
TABLE 7.5 Job Mobility, Tianjin, 1993
  Total Men Women
SOURCE: 1993 Tianjin Household Survey.
Number of cases 1037 701 336
Never changed jobs (%) 41.5 38.1 48.5[∗]
Ever changed jobs (%) 58.5 61.9 51.5[∗]
 When changed      
  Before 1978 15.4 15.8 14.6[∗]
  After 1978 43.0 46.1 36.9[∗]
 What Changed      
  Occupation only 22.9 23.8 21.1[∗]
  Work unit only 4.5 4.1 5.4
  Both occupation and work unit 31.0 34.0 25.0[∗]
Average number of occupations ever worked in 2.2 2.3 1.9[∗]
Average number of work units ever worked in 1.6 1.7 1.4[∗]
work units with or without occupational change is only 55. The average number of occupations that people ever worked in is 2.3 for men and 1.9 for women, and the average number of work units that people ever worked in is 1.7 for men and 1.4 for women. Both gender differences are statistically significant.

What are the outcomes of occupational and work unit change? Table 7.6 shows that when men change occupation, work unit, or both, they have the clear advantage of ending up in administrative, market activity, and market potential jobs rather than other jobs. In contrast, while women in any job change category have an advantage of moving into clerical jobs, they have no opportunity to move into administrative jobs, and their chances of moving into jobs with market involvement are less than their chances of moving into jobs with no market involvement or decision making. As shown in the last row of results in table 7.6, men in any job-change category enjoy a 14–18 percent wage gain; for women, however, wage gains only follow changes involving both occupation and work unit, a move they are less likely to make than men.[5]


Perhaps we should not be surprised that gender inequality has changed solittle in China over the last 15 years. In Russia, another “post-socialist” country,

<COL align=left> <COL align=left> <COL align=left> <COL align=left> <COL align=left> <COL align=left>
TABLE 7.6 Effects of Job Mobility on Current Occupation and Wage among Those Changing Jobs since Job Entry, by Gender, Tianjin, 1993
Dependent Variables Changed Occupation Only Changed Work Unit Only Changed both Occu-pation and Work Unit Model Chi Square N
SOURCE: 1993 Tianjin Household Survey.
p <.05 (two-tailed test).
NOTE: The top part of the table reports odds ratios of the effect of occupational change only, work unit change only, and a combined work unit and occupational change on occupation at the time of the survey.
For the dependent variables, the contrast category consists of occupation with no market involvement or decision-making power. Control variables include: seniority, seniority squared, high school, college, Party membership, the 4 job dummies, and the 2 sector dummies. The job and sector dummies are measured at the time prior to job changes.
The bottom part of the table shows the effects of job mobility on the natural logarithm of current monthly wage. The only control variable is natural logarithm of first wage.
Current Occupation: Clerical vs. other job          
 Men 5.33∗ 10.21∗ 5.62∗ 183.06∗ 319
 Women 1.17 16.12∗ 5.08∗ 116.90∗ 200
Administrative vs. other job          
 Men 2.19∗ 5.94∗ 4.19∗ 209.36∗ 305
 Women 0 0 0 85.55∗ 162
Market Activity vs. other job          
 Men 5.27∗ 9.89∗ 8.38∗ 189.26∗ 298
 Women 0.32 1.72 2.45 123.40∗ 191
Market Potential vs. other job          
 Men 1.39 4.64∗ 3.72∗ 325.69∗ 441
 Women 0.11 0.27 0.28 210.38∗ 220
Current Wage:          
 Men 1.14∗ 1.18∗ 1.14∗ R2 =.05 690
 Women 0.95 1.03 1.22∗ R2 =.03 333

Gerber and Hout (1998) show that women's average wage was 63 percent of men's in 1991, fell to 53 percent in 1994, but recovered to 65 percent in 1995. In fact, this last figure is the same as that reported for 1970 (Whyte 1984, table 2). Trends in gender inequality in wages hinge on changes in a wide range of institutions and policies, from those that give women less education and political standing, to those that pressure them to leave the workforce, to those that place them in different kinds of jobs, and finally to those that grant them different wages for the same job. Furthermore, there is a great deal of inertia built into the occupational structure because overall wage differences between men and women at any time are strongly affected by the background, job assignments, and career mobility of cohorts of workers who entered the labor force decades before.

In this context, we believe that the competing views about the impact of economic reforms on Chinese women's economic status relative to men's, reviewed at the beginning of this chapter, are too general and abstract. The Chinese state was never able to deliver equal opportunities to men and women with respect to higher education, Party membership, administrative and market jobs, and wages, even before 1978. Chinese women in the post1978 reform era also fall short in terms of human capital and their working careers. If we try to squeeze our results into these theoretical boxes, we draw the following conclusions:

The market-as-equalizer argument suggests that the gender wage gap is narrowing because in a more market-driven economy women may enjoy marke to pportunities not available to the min the previous redistributive system. However, the Tianjin surveys support neither the predicted outcome nor the hypothesized causal process that would lead to a declining gender inequality in pay. First, gender wage gaps persist from 1978 to 1993. They also persist across the five labor cohorts who entered the urban labor force from 1923 to 1993. Second, job assignments after 1956 clearly disfavored women, but post-1978 reforms also did not give advantages to women in job mobility. On the contrary, men are more able than women to move into jobs with power, jobs with market connections, and jobs in the state or new sectors, consequently obtaining higher wages. Although younger persons in the 1978–93 cohort were no longer assigned jobs by the government in the early 1990s, the improved job placements and higher first wages of women in this cohort—comforting from a market transition perspective—must be reconciled with the negative results for women's job mobility occurring mostly after 1978.

Stability in wage inequality challenges the state-as-equalizer model just as it challenges the market-as-equalizer model. The state-as-equalizer model implies that gender inequality in wages should be widening over time, but this does not occur. Even at the entry level, gender gaps are quite persistent from the 1949–57 cohort to the 1966–77 cohort: on average, women's first wages are 21–25 percent less than men's. This is primarily because women are assigned to jobs with no market involvement or decisionmaking power—the residual category in our occupational classification— and placed in the collective sector or work units with lower rank. As Honig and Hershatter (1988) observed, work units continued to discriminate against women in job assignment in the 1980s, as they had in previous decades. There is evidence that differential job mobility worsens women's placement in occupational and work unit structures, but we found no evidence that before 1978, when job mobility was highly controlled by government bureaucrats and work unit managers, there was any different pattern. There seems to have been persistent gender inequality in job mobility in Tianjin before and after reforms.

What accounts for the unexpected persistence in gender inequalities? First, China's policy continues to require working women to retire ten years earlier than working men. Seniority has a consistent linear effect on base salary over time, and women are largely excluded from salary increases beyond age 50. Second, women continue to have less access to higher education and Party membership—important determinants in the distribution of base salaries and other wages before and after reforms. Third, women are under-represented in administrative jobs and market potential jobs. Fourth, women are more likely than men to be placed in collective enterprises rather than state work units. Finally, women change occupations and work units less than men, and when they do change, it helps them less. Their wages increase at a lower rate than men's, given these changes.

It appears that neither the state nor the market alone can explain the gender wage gap. Rather, the state and market are structural contexts for the power relations between women and men. These power relations are not necessarily, or automatically, evolving with the economic reforms.


Revision of this chapter has benefited greatly from the comments of the participants at the conference “Gender, Households, and the Boundaries of Work in Contemporary China” in October 1996, especially Barbara Entwisle, Gail Henderson, and Martin Whyte. We are also grateful to Peggy Marini and Bill Parish for their helpful comments on an earlier draft and to the Tianjin Academy of Social Sciences for its collaboration on the 1993 Tianjin survey. Funding for various stages of this research has come from a National Science Foundation grant (SES-9209214) and grants-in-aid from the University of Minnesota Graduate School.

1. This is surprisingly close to the report by Hare (1995) that men's average wage is 25 percent higher than women's in Chinese village factories. It is beyond the scope of this chapter to analyze why China has had a persistently smaller gender wage gap than the United States or other Western countries. We do note, however, that it seems to be closer to that found in other East Asian economies (Lau et al. 1994). For a comparative analysis of gender inequalities in market capitalist and state socialist systems, consult Whyte (1984).

2. We did not include any variables of parental status as predictors because too many respondents (50%) provided no data. In a preliminary analysis (not presented) of the respondents who reported their parents' education, occupation, work unit sector, and cadre status, we found no gender difference in the effects of these variables on respondents' first wages. In a 1985 Tianjin sample, Lin and Bian (1991) show that the father's work unit sector has a significant effect on son's first work unit sector, but not on daughter's. This differential parental influence is partially and indirectly taken into account in our regression analysis of first wages as we include respondents' first work unit sector as well as other first-job indicators in our wage equations. Full results from the regression analyses are available from the authors. 

3. In this table, “gross gender effect” is the gender wage gap on the margin; the coefficient presented is obtained from a regression equation in which gender is the only predictor considered. In contrast, “net gender effect” is the gender wage gap between men and women with comparable characteristics measured by the predictors considered in the equations; so the coefficient for gender is obtained from regressing first wage on gender as well as all other control variables. Although first wages vary from cohort to cohort because of inflation effects, which are greatest in the youngest cohort, we still decided to use untransformed wage as the dependent variable because regressions using log form wages underestimate gender wage gaps (see Portes and Zhou 1996 for discussion on estimated effects of self-employment on wages from regressions using original and log form earnings).

4. There is a significant gender effect for the 1966–77 cohort, gross and net. This is surprising, given a widespread belief that economic egalitarianism reached its momentum during the Cultural Revolution (Parish 1984a). Parish's theme originally points to destratification in the post-1966 period by class, education, and occupation, and indeed receives support from the 1993 Tianjin survey: education and occupation dummies have weakening effects on initial wages from the 1958–65 to the 1966–77 cohort (results available from authors). Although the net gender effect on first wage is puzzling, a 1988 study of Tianjin reports the same result (Bian 1994b). Further analysis (not shown) indicates that the net gender wage difference can mostly be traced to persons in the category of “other” jobs (50% of men and 46% of women). Among occupations in this category, men were more likely than women to be assigned jobs in heavy industry in which starting salaries were higher than in other industries.

5. However, there is little gender difference in mobility into state and new sectors, as compared to moving into the collective sector, a result not presented in table 7.6.

【字体: 】【打印】【关闭

附 件:

| 中心简介 | 网站介绍 | 版权声明 | 服务条款 | 站点导航 | 请您留言 | 网站地图 | RSS聚合资讯