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
9. The Perils of Assessing Trends in Gender Inequality in China

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

Martin King Whyte

Those who study China have an understandable desire to reach conclusions about the overall extent of gender equality and inequality in that society, and whether things have gotten better or worse over time, particularly as a result of China's reforms. All of the chapters in this section, and to some extent most of the other chapters in the volume, wrestle with these questions. I want to stress how difficult it is to reach such summary judgments. Very little of what I have to say will be new, and in fact many of the same comments have been made about stratification in general and in any society, not just regarding gender inequality in China. Many of the points I make will seem quite obvious. However, given the frequency with which analysts ignore these complexities and problems while pursuing summary evaluations, I think these points are worth reiterating.


The first obvious point to stress is that if we want to assess the extent of gender inequality at one point in time, or judge the trend in such inequality over time, we need to decide what realms, dimensions, and indicators to consider. Obviously, there are multiple realms in which gender equality and inequality can be observed. One approach is to consider women's relative position in economics, politics, religion, cultural images, family life, and so forth. Even if we restrict our attention more narrowly, as in the present volume on gender and work experience, there are still multiple indicators and measures. At the outset we need to note that there are both objective and subjective dimensions of the relative position of women in the world of work. The subjective feelings women (and men) have about their jobs, their relations with colleagues and superiors, their chances for advancement, and the influence of such things on their social standing are the focus particularly of Hershatter's chapter 5, and also to some extent of Honig's chapter 6. While less often systematically studied than objective dimensions, these subjective dimensions are important socially and politically. Individuals and groups are moved to feelings of satisfaction or resentment not by trends in objective indicators of status, but in response to these subjective factors. But by their very nature these are quite difficult to systematically measure.

Consider the images of women conveyed in the Hershatter and Honig chapters. The young urban women described by Honig who were sent to the countryside in the wake of the Cultural Revolution saw themselves as achieving rough equality with their sent-down male peers, and perhaps even with village men. However, they viewed village women as locked into a subservient status. The boldness and pride of the rural female labor models of the 1950s portrayed in the Hershatter chapter is missing from this picture of subservient rural women. Obviously, the differences between the times and places dealt with in these two chapters and among the types of individuals who provided the information upon which they are based may help explain this contrast. But just as obviously, it is difficult to use such subjective accounts to construct a summary evaluation of the relative situation of women and whether it has improved or deteriorated.

Difficulties such as these lead most social scientists and informed observers to rely on objective statistics rather than subjective accounts. However, even if we focus on objective indicators of trends in women's relative standing in the world of work, we face a complex terrain. We can try to measure, for example, women's rates of labor force participation relative to men's, their relative income (or even relative share of various kinds of income—wages, bonuses, subsidies, and so on), relative representation in various occupations or industries, relative representation in different kinds of work units, relative access to various advantaged positions (e.g., work unit leadership positions, promotions in rank and pay, voluntary job changes, enrollment in specialized training, model worker status) and relative vulnerability to various disadvantages (e.g., being laid off, demoted, fired, involuntarily transferred, forced to work overtime). We can even spread our net more broadly and consider such things as relative autonomy in daily work, exposure to hazardous work conditions, difficulty of commuting to work, relative likelihood of being assigned housing, extensiveness of guanxi networks, and other realms. The list of possible indicators becomes longer if we consider, as do the chapters by Bian, Logan, and Shu (chap. 7) and by Michelson and Parish (chap. 8), that we also need to consider women's relative access to education and Party membership as well as divergent retirement ages, since these realms have clear influence on men and women's work status and opportunities.

Given such complexities, even if we are only concerned with assessing overall gender inequality at one point in time, we are likely to see a confusing and apparently contradictory picture. To start with, we rarely have access to the kinds of data that permit a confident assessment of the overall situation. For any single indicator, furthermore, we need some sort of comprehensive or representative statistics of the situation of women relative to men. However, we often have only fragmentary statistics from possibly atypical work organizations or localities or anecdotal information, and these refer only to the situation of women, rather than to that of women compared to men. Even when we can find systematic statistics of women's situation relative to men we may not be on safe ground, since we need to consider how accurate and meaningful those statistics are. In the case of statistics derived from Chinese publications, we may have little idea of how the data were collected and whether there are systematic biases built into them.

Even if we were able to assemble a range of representative statistics on the relative situation of women compared to men in various work domains, we could not expect these to reveal a clear or consistent picture. Instead we would find, for example, that women are fairly close to parity with men in labor force participation rates, at least between the ages of 20 and 50 (for rural China, see chap. 8); however they are much further away from equality in realms such as university attendance and Party membership (see chap. 7, table 7.1). The fact that we may get different pictures from alternative indicators leads to another cautionary note. Given such disparities, there is no single indicator that can be taken as the “best” or even a “good” summary measure of the relative status of women in China. This is not, I stress again, a situation distinctive to China, but rather a reflection of the complexities of gender stratification around the world (Whyte 1978). Therefore, in order to assess the relative situation of women we need to assemble as many meaningful indicators as possible.

The “one point in time” assessment is also made complex by the question “Compared to what?” and by its logical cousin, “Is the glass of water halffull or half-empty?” In every country in the world, gender inequalities are visible, and such inequalities were certainly sharp in China's past. By what standard can we judge whether the gender gap revealed by any particular Chinese indicator is large or small? In some cases the comparative verdict may be quite clear. For example, female labor force participation rates in China, at over 90 percent and close to parity with males, are among the highest in the world. In terms of wages the common finding in Chinese urban surveys—that women earn about 75–80 percent as much as men— also looks pretty good in international comparison (see chap. 7 and the comparative statistics cited there). However, in other cases what we should conclude is not so clear. Is the fact that women constituted only about 13 percent of all members of the Communist Party in the early 1980s bad or good, given the very considerable under-representation of women in political leadership in every society around the world? Even if we were able to assemble internationally comparable data on the indicators we are using for China, the picture is likely to be mixed, with Chinese women faring better than women in comparison societies according to some indicators and worse according to others. Which other societies are most suitable for such comparisons is also a matter of some debate. China is often compared to advanced industrial societies and particularly to the United States, perhaps primarily because those doing these comparisons come from such societies and are most familiar with the situations in their home countries. Arguably, comparisons with other East Asian societies, other developing societies, or Soviet bloc countries (pre-1989) would be more appropriate, but comparable statistics from such places may be hard to locate.

The problems of reaching summary judgments are compounded when we want to answer the question of whether gender inequality is increasing or decreasing over time. In addition to all of the complexities already noted, one needs comparable data from different time points for any single indicator in order to talk about trends. Unfortunately, analysts quite commonly find current data indicating something good or bad about women's relative situation and conclude that things have changed for the better or worse, without examining the situation at an earlier point in time.

Obviously, in order to make assessments about changes over time we need comparable baseline data for the earlier period in our comparison. Once again it may be difficult to locate such information for many of the indicators of interest. Few systematic statistics on the relative position of women in work and other realms were published in the pre-reform period. There was also little in the way of survey research using scientific methods carried out in China prior to the 1980s. In some cases, such as school enrollments at various levels, figures were reported and collected but not openly published during the Maoist period. In the more open atmosphere of the reform era such statistical series have been openly published, making baseline figures available for some indicators (e.g., Research Institute of All China Women's Federation 1991). Of course, with little information provided in most cases about how the data were reported and tabulated over the years, one may question how accurate these recently published statistical series are. In other cases estimates for earlier periods can be derived from retrospective data provided in current surveys or from cohort comparisons within such surveys, although such indirect estimation methods are far more problematic than would be data that had been collected at the earlier time.[1] However, for some indicators of interest it is very difficult or impossible to locate or construct baseline estimates for earlier periods. For example, it would be very difficult to get accurate retrospective figures from the Maoist era on indicators such as male versus female pay levels, overtime work stints, or personal guanxi networks. Carrying out the sort of regression analysis presented by Bian, Logan, and Shu (chap. 7) and by Michelson and Parish (chap. 8) to see the net effect of gender on work status, controlling for background factors, requires access to data on individuals, and access to any such data collected in China prior to the 1980s is rare. For a variety of reasons, then, we may hope that data collected in recent years will be analyzed and archived to provide baselines against which to compare the future situation of Chinese women. But the availability of only fragmentary baseline data on the past means that systematic examinations of the impact of the reforms on gender relations remain problematic at present.

A further obvious point is that to make meaningful statements about changes over time we need data not only on the situation of women but also on their situation relative to men in two or more time periods. Figures on changing numbers of female college graduates, administrative cadres, engineers, or other relevant categories may be affected by general trends in China, and without comparable data for men, we cannot tell from such figures how any increase or decrease compares with what is happening to men.[2]


The perils of drawing conclusions from current phenomena are perhaps especially great in China, where official history has been rewritten repeatedly. In addition, the reform era has seen a rise in “proto-feminism” within the official Women's Federation and elsewhere in Chinese society, with one manifestation of this trend being a sharp increase in exposés of the mistreatment and disadvantages suffered by women. We face a classic case of what social psychologists call a “labeling problem”—has mistreatment of women increased, or was it simply ignored and covered up in the past and is being highlighted now? Given this situation, presuming what the past situation was like is very risky. Obtaining accurate information about the past in order to make judgments about changes over time is quite difficult, as noted above. Given this cautionary note, however, I still applaud efforts such as those by Hershatter and Honig (chaps. 5 and 6) to use oral history interviews, contemporary documents, and memoirs in a cautious and critical manner to get us beyond prevailing stereotypes about the lot of Chinese women in the 1950s and 1960s.

Let me elaborate on the difficulties posed by the “labeling problem.” Conventional wisdom says that women have lost some ground relative to men during the reform period. Explanations for this deterioration are somewhat different as applied to urban versus rural China. For urban areas it is argued that market reforms and the profit motive lead decision-makers to prefer men and discriminate against women more than was the case during the Maoist era. In rural areas any harmful influence of market forces is compounded by weakened state control over peasants and the revival of traditional beliefs and customs, which encourage discrimination against women. For evidence of these alleged trends conventional wisdom cites figures on scant hiring and excessive layoffs of women by urban firms and of abduction and sale of women and female infanticide in rural areas.

There are at least two problems with using apparently “hard” statistics on such phenomena to conclude that women face increasing discrimination in the reform era: (1) we lack comparable data for earlier periods, and (2) we may have reason to doubt the objectivity of the sources of such figures. As already noted, a very healthy trend in the reform era is that muckraking journalism and critical social science have focused attention on the many disadvantages faced by women. At the same time, since there were no such critical voices raised during the Maoist period, it is hard to know whether the phenomena described have become worse, have existed all along but are only now being criticized, or perhaps have even improved as a result of the critical attention. We can try to examine some specific claims more closely and see whether we can find supportive or disconfirming evidence. In regard to claims about discrimination in hiring and excessive layoffs of women by urban work organizations, there are several reasons to be skeptical. First, we know from earlier research that at certain times during the Maoist era there were similar claims that women were disproportionately pushed out of jobs, leading to some doubt about whether the reform era really initiated a new trend (e.g., L. J. Huang 1963). More to the point, such systematic data as we have for urban China in recent years indicate either stability or improvement in the employment situation of women, rather than deterioration.[3] Finally, we need to keep in mind the observation in Michelson and Parish's chapter 8 that some of the difficulties affecting women in the early years of the reform may be transitory. In particular, any propensity to resist hiring women or to disproportionately select female employees for layoffs because of the supposed negative impact of women employees on productivity and profits is likely to disappear in the face of evidence of the dynamism and success of other firms employing substantial numbers of women workers.[4] In sum, a variety of evidence leads to considerable doubt about a consistent trend toward increased discrimination against women in urban employment during the reform era.

What about the claims that in rural areas the reform era has seen an increase in the abduction and sale of women? Accounts tell us that recent years have seen more and more cases in which thugs deceive and kidnap women and sell them as brides to men in other villages. These men are often portrayed as beating and raping the women in order to keep them from escaping. In some instances organized gangs are said to engage in this sordid trade (e.g., Sun 1992; Tefft 1994). Yet even in regard to such disturbing claims there are grounds for skepticism. Once again we are presented with anecdotal evidence and fragmentary statistics only for recent years,[5] but no comparable evidence on any earlier period. Might not these reports indicate a continuing phenomenon, rather than a genuine increase over time? We also know that Chinese authorities have made efforts to denounce trafficking in women and to arrest and imprison those engaged in the trade (e.g., Huang Wei 1991). One added consideration is the comment in Honig's chapter that urban observers during the Cultural Revolution (the sent-down “Iron Girls”) were startled at how traditional and down-trodden the village women were. Given these observations, how can we be sure whether the number of abductions and sales of rural women have increased, or whether critics and authorities (most of them of urban origin) have simply been drawing more attention to an existing phenomenon that has not increased?[6]

Further, many of the behaviors noted sound not too different from the features of Chinese arranged marriages over the centuries: a rural family, in exchange for a substantial payment (bride wealth), sends their young and often terrified daughter off into a strange and perhaps distant village to be married to a man she has never met. Her husband is expected to use force if necessary to initiate sexual relations and to prompt his unwilling bride to show proper respect for his family, and she cannot expect other villagers or even her own natal family to come to her assistance. Arranging such marriages becomes a lucrative business for some individuals (marriage gobetweens). We know that this kind of totally arranged marriage became substantially less common in rural China during the Maoist period but did not disappear entirely (Parish and Whyte 1978, 169–80). While not condoning such practices, we need to ask: when does an arranged marriage become a case of abduction and rape? If there is something new here, it may be that some of these unwilling brides are educated and able to utilize the provisions of China's marriage laws (of 1950 and 1980) and sympathetic urban allies to escape from such arrangements and bring down sanctions against those responsible for them. If this interpretation rather than the conventional “rise of abductions of women” is correct, perhaps we should regard these cases as a sign of progress, rather than as evidence of deterioration in the treatment of women.

The third claim I want to examine is the purported rise in female infanticide in the reform era. Here one has some of the same reasons for doubt as in the two cases just discussed. Evidence about distorted sex ratios and anecdotal information about female infanticide provided a crucial opening for Women's Federation cadres and others to make better treatment of women part of China's political agenda. Precisely for this reason one might wonder how typical or new the described phenomena are. In this case we have a basis for checking these claims against systematic evidence regarding the past. This is possible because demography leaves indelible traces in the characteristics of the living. Even though data on female infanticide and sex ratios at birth were not published during the Maoist era, with proper estimation procedures one can use information on the gender, birth order, and other characteristics of Chinese cohorts born over the years to determine what the trends in excess female infant mortality have been. The results reveal a very significant decline in excessive deaths of infant females beginning before 1949 and continuing into the 1950s, a significantly lower but still nonzero level maintained during the 1960s, and the beginnings of a new rise in excess female infant mortality in the 1970s, which escalated in the early 1980s (Coale and Banister 1994; K. Johnson 1996). Based on this evidence, we can have confidence that the rise in excess deaths of infant females in China in recent years is not simply the result of selective data and “labeling” processes, but represents a genuine deterioration in the treatment of Chinese women.

That said, it should also be noted that this trend is a product not so much of economic reforms, but of China's draconian birth planning policy. The latter corresponds roughly in time to the economic reforms, but its spirit and implications are quite different. Conventional wisdom says that in the reform era decentralization has weakened the power of the state to carry out “affirmative action” policies that benefit women, allowing traditional patriarchal (and new capitalist) practices to emerge that harm women. In the case of the resurgence of female infanticide, however, this is not what has happened. Instead we see here a still surprisingly strong state successfully enforcing policies on the population that directly endanger women.[7]

I have discussed quite briefly three specific claims about trends in the treatment of women, and my verdict is mixed. I am quite dubious about the claims of increased job discrimination against urban women during the reform era, troubled but still not persuaded by claims about a rise in abduction and sale of rural women, but quite convinced by the evidence of a new rise in female infanticide.


To attempt an overall assessment of whether women have gained or lost ground relative to men during the reform era, a systematic analysis is necessary of a much larger number of indicators of gender status carried out in a less cursory way than I have done here for these three indicators. I presume that such a comprehensive analysis would again lead to a mixed picture. In some realms, such as the relative educational attainment of young women in urban (Bauer et al. 1992) and rural areas (chap. 8), female Party membership,[8] and particularly the rise of what I have called “protofeminism,” we see clear signs of improvement for Chinese women; in others, such as urban relative income (as discussed in chapter 7), perhaps our verdict would be one of little change; while in still other realms, such as the survival chances of infants, we would conclude that women have been harmed by recent changes. In regard to still other indicators, as Michelson and Parish suggest (chap. 8), we would not find any clear or linear trend. This is the case, for example, in regard to female access to university-level education. During most of the pre–Cultural Revolution period (before 1966), female students constituted only 23–26 percent of all university students; when universities reopened after the Cultural Revolution, this figure rose to 32–33 percent; during the initial phases of the post-Mao reforms of university enrollment, the percentage fell back to 24–27 percent; but by the late 1980s it had risen again to 33–34 percent.[9] Clearly, in such a case no simple conclusion about the impact of the reforms captures the actual trend over time.

As a further cautionary note, the part of China's population to which any statements about trends apply and how broadly any conclusions can be generalized are also important considerations. As the chapters in this volume and my earlier comments imply, there are major differences between the social worlds of rural and urban China that affect the lot of women. Quite possibly, women's lot in urban areas as measured by a particular indicator will have improved over time, while in rural areas their lot will have deteriorated. Important regional and local variations, age cohort differences, class and status differences, subethnic variations, and so forth are also likely in regard to gender relations. Because most of the data we use come from particular locales or types of work organizations, it is important to carefully consider the extent to which any trends observed could be atypical.

A final cautionary point concerns the difficulty of attributing causation for any trends we do observe. In regard to the “impact of reforms” question in particular, the post-1978 period has seen not only the launching of a market-based transformation of the economy, but also rising incomes, urban expansion, growing contacts with the outside world, the collapse of the Soviet bloc, increasingly harsh enforcement of birth control, and a number of other trends. In this situation it is risky to attribute any particular change in gender inequality to market reforms unless one can rule out some of the other social forces operating during the same period as possible causes.

My conclusions should now be fairly clear. Appropriate evidence is not currently available to make any sort of overall assessment of the relative situation of Chinese women and how it has changed over time. If more systematic evidence can be assembled, I think it is very unlikely that such evidence would lead to a simple answer about whether women's situation relative to men has improved or deteriorated in the reform era and as a result of those reforms. By presenting my laundry list of cautions and complexities here, my intention is not to persuade analysts to throw up their hands and stop their research on gender inequality in China. All really important subjects for social investigation are fraught with similar complexity and difficulty. My purpose is simply to point out that the search for an overall conclusion to the question of the impact of reforms on Chinese women can easily lead to biases and oversimplifications. The quest for sexual equality, in China or in any other society, is a battle in which progress or retrogression occurs in many domains and subpopulations, and only by documenting and explaining the outcomes of these many separate skirmishes will we gain a sense of how the overall battle is going.


1. Retrospective reports may be biased; the use of older cohorts to estimate earlier experiences ignores period and aging effects; differential mortality and migration make the current population sampled different from those who would have been found in the same place if an earlier survey had been carried out; and so forth.

2. Obviously there are some indicators for which no meaningful comparative indicators for men exist, such as the frequency of abortions or cases of women being sold into prostitution.

3. Chapter 7 by Bian, Logan, and Shu and also an earlier article by Bian and Logan (1996) support claims of stability in the treatment of women in work in Tianjin over the course of the reform era. In a related analysis by Tang and Parish (2000), urban Chinese survey data reveal that younger women are generally faring better in terms of employment and income than did older women when they started work, suggesting an improvement in working conditions for women, rather than a deterioration. For similar results based on the urban portion of the 1988 national income survey, see Rublee (n.d.).

4. The claims made during the 1980s that Chinese employers were loath to hire women primarily involved the contention that managers believed the family burdens disproportionately shouldered by women and the paid maternity and sick leave provided by the work unit would make a woman more costly and less “efficient” than a man. However, other stereotypes—that women are more diligent, less likely to engage in behavior that interferes with work (alcohol abuse, violence), and willing to work for lower wages—point in the other direction and might lead managers to prefer to hire women. Research in the United States provides no support for the view that family burdens mean that employing a woman entails a cost for the firm compared to hiring a man. See Bielby and Bielby (1988).

5. The Tefft article quotes the Chinese newspaper Legal Daily as claiming that in 1991 and 1992 there were 50,000 cases of abduction of women in China.

6. Not all accounts of abductions of rural females claim that the phenomenon has increased during the reform period. Some of the proto-feminist voices argue simply that this is a social evil that has been ignored rather than attacked (with the question of prevalence during the Maoist era left ambiguous). It may be impossible to assemble accurate information on how common such abductions were in earlier periods, thus preventing any confident assessment of trends over time.

7. To be sure, rural Chinese families would not selectively abort female fetuses and abandon infant daughters if they didn't hold traditional attitudes favoring sons. However, the analysis of Coale and Banister (1994) indicates that those traditional attitudes did not produce much excess infant female mortality when rural families had three or more children; they only had such repercussions once the state tried to enforce a limit of two children or fewer.

8. In 1983 it was reported that women constituted only 13 percent of all members of the Chinese Communist Party (see People's Daily, 11 December 1983, p. 5). By 1997, the figure had risen to 20.1 percent (from Xinhua News Agency report, Beijing, 7 July 1997, translated in Summary of World Broadcasts, Asia-Pacific, July 9, 1997, FE 2966, p. G/3.)

9. See the figures presented in Research Institute of All China Women's Federation (1991, 168). In 1992 the figure was 33.7 percent. See Beijing Review Staff (1995, 9).

【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