Forty Million Missing Girls:



In the study of China’s rural development, economists and political scientists have frequently examined land policy, while demographers, sociologists and anthropologists look at family planning. Yet in real life the two domains are closely related as households attempt to match and manage their land and labor resources. This article brings together questions about land, gender and family planning in relation to both policy and practice. It draws on fieldwork in rural north China and comparative data to examine and assess local and regional variations in the critical gender imbalance in contemporary rural China.


“The dearth of girls is now more extreme in the PRC than anywhere else in the world.”


One of the toughest challenges is to modify China’s rigid custom of patrilocal and patrilineal marriage, the restriction of land rights to the males of the patrilineal clan.” Banister (2004)


Missing girls


Recent studies suggest that over 40 million girls and women are ‘missing’ in China (Klasen and Wink 2003). The problem was driven home in various ways. One article stressed the “frustrations of as many as 40 million single men by 2020.”[1] A major book has warned of the impending danger of social and political instability stemming from a surfeit of young men with no prospects of marriage (Hudson and den Boer 2004). Reports from specific places illustrate the concern. The city of Guiyang in southwest China had 129 male babies born for every 100 female babies. Guiyang banned abortions after 14 weeks of pregnancy in January 2005. A large street banner bluntly proclaimed, “Firmly crack down on the criminal activity of drowning and other ways of brutally killing female babies.” But a clinic down the street advertised ultrasound tests, allowing people to choose methods other than drowning daughters.[2]


Discrimination against girls has undergone what Amartya Sen termed a radical change as it shifted from female disadvantage in mortality to female disadvantage in natality (2003). Over the past two decades, sex selective abortions in rural China and a number of other Asian countries have rapidly been replacing the brutal methods of gender (sex) selection mentioned above.

Three basic explanations for Chinese gender discrimination occur again and again in scholarly and popular discourse. These are that sons are necessary for heavy farm labor, to support their parents in old age, and to carry on the family line. These explanations offer convenient stereotypes but little deep analysis. A brief look at them will raise some questions.


1) Heavy labor. Rural women have contributed a large proportion of the farm labor in China for nearly fifty years. The fact that many men have left farming to women and migrated to towns and cities in search of better incomes since the 1970s undercuts the “heavy labor” argument.[3]


2) Old age support. The patrilineal family system requires that a son stay home to support parents and that daughters who marry out support their husband’s parents. Parents expect to be supported by sons and daughters-in-law. Sons are given the means, in the form of heritable local property rights, to support parents. Yet it is not uncommon in my fieldwork to run across parents who receive as much or more economic and personal-needs support from their married-out daughters, even those living in another village or district, as from their sons or daughters-in-law. Why is it inconceivable or unacceptable for daughters and sons-in-law to support parents in old age? Usually, the answer to this question is to move on to the “family line.”


3) The family line. Continuing the family line from father to son is a mandate and perhaps even a mantra, repeated so often in some areas that it is never questioned. Without doubt, the concept of the family line is a powerful force in Chinese society, one closely linked with historical lineage concepts that were suppressed in the Mao years. Why have lineages re-emerged and become so important in some areas in the post-collective era since the 1980s? Why do some regions and groups place much more importance on lineages than others (M. Han 2001, Cohen 1990, 2005)? Why does the search for expanded networks and social relationships often continue to exclude women from its formal mapping?


In addition to the three explanations above, social scientists have probed more deeply. Demographer Judith Banister (2004) examined the reasons for China’s growing sex imbalance including poverty, the political or economic system, socio-economic development, educational level, Chinese culture, the one-child policy, low fertility, and ultrasound technology. She found that China’s shortage of girls cannot be explained by poverty, political or economic system, by the level of socio-economic development or by educational level as these variables do not correspond to the demographic evidence. However, daughter shortage within China is closely associated with the distribution of Han Chinese culture within China proper, while the peripheries and most minority areas have more balanced sex ratios. Banister argues that the introduction of the state family planning policy, and its more rigorous enforcement in Han areas, is associated with a rising proportion of sons. This trend became even more marked once ultrasound technology for sex testing became available (though illegal for this purpose), followed by abortion for sex choice. The recent demographic data support the view that female disadvantage in mortality has been transformed into female disadvantage in natality. Banister notes that “the combination of continuing son preference, low fertility and technology is causing the shortage of girls in China.” (2004:13). Yet she also mentions a number of more specific factors.


“One of the toughest challenges is to modify China’s rigid customs of patrilocal and patrilineal marriage, the restriction of land rights to the males of the patrilineal clan, the traditional weakening of daughters’ ties to their natal families after marriage, the dependence on sons but not own-daughters for old age support, and other customs that make daughters worth little in the eyes of their natal families (Das Gupta et al. 2004). So that China’s daughters may survive and be valued as much as sons, they need rights and responsibilities to have lifelong close ties to their natal families. The government has promoted some of these changes and has passed egalitarian laws. The need now is to more vigorously enforce the laws giving daughters equal rights and responsibilities.” (Banister 2004:14).


Here we find one of the few clear references to the property system as a factor in China’s son preference. I will argue that more than labor or the family “line” is at stake. The system of family property and the political institutions for holding and transmitting land and property rights need to be more carefully examined.


Conventional explanations pointing to labor, old age support and family line do reflect the way many people talk about son preference. They point to relatively benign and uncontested cultural assumptions without calling attention to the more controversial and conflictual aspects of the property system that discriminates so persistently and pervasively against girls and women. This article considers the links between property rights and population control to see how these two policies work together to produce the enormous deficit of daughters. Who exercises de facto control over land and property? Despite government efforts to legislate gender equality, patrilineal institutions retain considerable power in many rural areas. Their workings within village power structures need to be made more visible. The close ties to women’s natal families that Banister advocates would be stronger if daughters were in line for a share of the family house and land.


Pradeep Panda and Bina Agarwal, writing about India, claim that few people study women’s property status in relation to violence. Examining marital violence, they argue that for women economic independence can deter violence. “In all existing research, however, a significant unexplored factor is the impact of women’s property status on the likelihood of violence. In fact, we came across no study either for India or elsewhere where this had been studied empirically” (Panda and Agarwal 2005). Panda and Agarwal are concerned with property ownership as a way of protecting women from marital violence, but I will consider the idea that property ownership for women may also increase chances of pre and postnatal survival for girls.

Facing massive sex imbalance, the Chinese government has belatedly moved from silence and cover-up to taking measures to encourage better treatment of girls. Yet while some policies reflect thought about the causes of discrimination, they ignore the core institutions that shape rural society to the detriment of women and girls. Some of the measures that have been reported in the press are:


1. A ban on ultrasound for sex determination,

2. A ban on late abortion,

3. Propaganda that girls are good, and

4. Provision of pensions for parents with one child or two girls.


The first three measures deal with symptoms rather than causes. Forbidding the use of ultrasound for sex determination, and forbidding the use of abortions to try again for the desired sex do not address the reasons why parents are making such a choice. Public denunciation of female infanticide provides a warning but does not address the reasons why people take extreme measures against newborn baby girls. Proclaiming that daughters are as good as sons is simply blowing in the wind. The fourth measure directly responds to the claim that sons are needed for old age support. In some places the government has promised to provide pensions for parents who have no sons, but no national resources have been allocated for such pensions, which can be found only in a few communities. Moreover, it remains to be seen whether promises for pensions at age 60 will significantly affect the reproductive choices of parents aged 20 to 30. [4] In sum, the kinds of vigorous measures required to end the killing of girls are nowhere evident.

While it is common to point to son preference in Chinese culture, sex ratios evened out during the collective years (Banister 2004, Greenhalgh and Li 1995, Johnson 2004), with the surplus of males only rising sharply during the reform period. For two decades, up to 1980, collective resources supported children so that the economic incentive for couples to keep reproduction in tune with household resources (or else become poorer) was lifted. In China as a whole, the combination of growing population and high state extraction of grain at low prices kept rural standards of living low (Huang 1990; Friedman, Pickowicz and Selden, 2005).


The 1980s reforms gave the rural population more incentive to produce, and greater control over household economic affairs. The distribution of land to farmers encouraged households to control births or face reduction in per capita land holdings. The state did not, however, leave this to chance; from the late 1970s it enforced rigorous birth quotas, initially one child per household.


Rationing land and rationing children was the state’s solution to the land problem. Rural families would receive equal amounts of land to use (not to own), and would have equal numbers of children. This would provide a form of social insurance for all at no cost to the state, and prevent the reduction in resources and income per capita caused by population growth. Surprisingly ignored, or dismissed, is the fact that children are not equal, and in Chinese society, men and women are not equal.


In historical understandings of gender, marriage was followed by patrilocal residence and patrilineal heirs. Parents supported sons by giving them houses and land, and sons supported parents in old age (Croll 2000). The link between them was property. Parents did not form a multi-generational contract involving land with daughters. These local understandings were the products of long-standing patrilineal tradition in most of China, supported up to the mid-twentieth century by organized lineages and clans. Maoist revolutionaries, while denouncing feudal thought, never challenged the tradition of daughters marrying out while sons inherited their parents’ living space and their father’s kinship ties. Sons were crucial to the contract between generations and to defining village membership.


The policy of limiting parents to one or two children directly affected the ability to obtain sons. Its impact was differentially felt, according to what Attwood (1995) has called “demographic roulette” or to what peasants traditionally called “fate” — that element of life that they could not control.


The family arithmetic is not too hard. At the birth of the first child, roughly 50 percent of the families will get a son and be satisfied, having produced an heir and a pillar of support in old age. The remaining 50 percent will eagerly await the birth of the second child. At the birth of the second child, half of these will get a son, bringing the total number with sons to 75 percent. A 75 percent majority will then have met their basic demand for a son, but 25 percent of families will then have two daughters. These are the families that, in a society in which regulations have made adoption extremely difficult (Johnson 2004), will take desperate measures which may include attempts to break the quotas, hide or give away girls, or if necessary abandon, kill, or neglect their infant daughters in the hope that they will then be able to have a son. Consistent with this, demographers find a “two-decade trend of deteriorating infant mortality rates of females relative to males” (Cai and Lavely 2003:17).


To these daughter-only families, ultrasound machines and abortion of female offspring are relatively low-cost alternatives in their quest for a son, and undoubtedly preferable to bearing and hiding, giving away, or abandoning unwanted second or third daughters.


Under the circumstances, local officials who enforce policies are caught between pressures from above to enforce birth quotas and angry villagers demanding the right to have sons. What of the anger and desperation of households that failed to produce a son — twenty five percent of households in areas where families honor the quota and do not prevent female births or kill newborn girls?


The outcome, now well demonstrated in massive statistical evidence of millions of missing daughters, is a product of the contradictions between national regulations and local culture in a system in which the state provides little social welfare and few safeguards for property rights that might, in the long run, mitigate the problem. By ignoring gender factors in framing the birth control policy, China now faces the problem of millions of missing girls and women, and millions of bachelor men. The rules and structures of village life combined with those of the national government produced the demographic and social consequences that challenge China today.


The Yellow Earth: Huang Tu Village, Henan


I began fieldwork in Huang Tu (a pseudonym) in spring, 1989 with brief visits in the 1990s and again in 2004. I selected this farming village for two reasons. First, as early as 1981 the census data for Henan Province showed a demographic pattern favoring sons, with a birth sex ratio of 110 males per 100 females (State Statistical Bureau 1984:420-427). Sex ratios that rise above the expected norm of 106 male births per 100 female births are a warning sign that female children are endangered. [5] In 1989, Henan’s birth sex ratio rose to 116, the third highest in China (Zeng Yi et al (1993:294). Second, the village was located in a wheat-growing region, like most of the north China plain, with cotton as a subsidiary crop. Historically, these crops provided the main food and clothing for the farmers and for the market, until home spinning and weaving was largely replaced by factory-made cloth, reducing women’s work and their economic value to the household. When I arrived in 1989, over 80 percent of Huang Tu Village labor force was employed in farming, with construction as the next largest occupation. Women played an active role in the agricultural labor force during the collective era and in the reform period women did even more of the farm work as large numbers of men sought work outside the village. Women were among the family workers sent to dig ditches for a village irrigation canal, and they worked in the fields to plant and harvest crops. Huang Tu Village seemed an appropriate place to examine the relationship between gender and development and a possible setting in which to examine the institutions that devalue women. To what extent did development contribute to or undermine what has been called “fierce” patriarchy (Drèze and Gazdar 1996) in which daughters’ lives are systematically devalued?


Henan Province and one county: population and missing girls


By 2000 Henan Province, with a population of more than 91 million, was China’s most populous province (Henan 2005). [6] The sex ratio for Henan’s total population in 2000 was 106 males per 100 females, having risen from 104 in 1981. [7] That for children aged 1-4 had shot up to 136 males per 100 females, with nearly half a million girls missing for the previous four years, 1996-1999. [8] Henan’s child sex ratio is among the highest in China which reported a national ratio of 121 males per 100 females aged 1-4 in 2000 (Banister 2004).


The county in which Huang Tu Village is located had about 529,000 people in 1988, and reached nearly 643,000 in 2000 (China census 2000). Sex ratios began to rise in the mid-1980s (Figure 1), shortly after the one-child policy was established, and rose rapidly at the end of that decade.


Beginning in 1985, a similar rise in the proportion of boys was recorded at the township (administrative village) level, a unit that encompasses more than ten large villages. Abnormally high sex ratios thus appeared in this rural area fairly soon after the family planning policy was introduced, and before ultrasound machines were commonly available in rural areas and county hospitals (see Figure 2). In the early 1990s, I observed that county level family planning offices in the region were still checking contraceptive use by using primitive x-ray machines to detect the presence or absence of an IUD, or a pregnancy.


These county and township data show that the broad pattern of missing girls was not acute until the 1980s. In 2000, the pattern of elevated child sex ratios persists. Compared to Henan province child sex ratios (aged 1-4) with a ratio of 136 males per 100 females, the county shows an even higher sex ratio of 146 males per 100 females for children aged 1 to 4. Based on the number of recorded male children, 23,244 females were expected, but only 16,859 were counted. An estimated 6,385 female children were missing. In 1988, Huang Tu Village documents recorded a balanced population with 1206 males and 1210 females. This figure is misleading, however, as men made up only 47 percent of the official village labor force. Village women, at 53 percent, were the dominant source of village labor. The shortage of men was caused by their transfer to nonfarm (urban) household registration status when they obtained government jobs in towns or cities. [9] Thus, village records excluded about 50 adult men whose wives and children retained village registration. This suggests that sex ratios among children were probably higher than normal, with about 50 extra boys. Of registered villagers counted in the labor force, all but 8 (of 495 women and 435 men in the workforce) were in “household-managed” occupations, a reference to farming.


From my first visit to the village, the state birth control policy has been a highly sensitive topic. In 1989, I photographed village officials confiscating household furniture and sealing up the empty house of a family that had fled the village in order to have another child. The offending family’s first child was a girl, and the daughter-in-law was pregnant again before the required four years had elapsed when a second try would be permitted. Higher officials confiscated my pictures.


Two sets of local data, from 1989 and from 2004, supplement official data on sex ratios from higher levels of government. In 1989, I conducted a survey of 50 households selected from each of the eight teams. My sample of 83 children aged 0 to 15 shows a sex ratio of 152 males per 100 females. The data suggest that the sex bias may have been present even before the family planning policy began to be strictly enforced in the early 1980s.


In 2004 I was given access to the household registration records for one of the eight teams. These records suggest that Huang Tu Village sex ratios are roughly comparable to the abnormally high sex ratios reported in census data for the township, county, and province up to the year 2000.


Table 1. Huang Tu Village child sex ratios by age for one team, 2004

Years born


Number of births

Sex ratio

(M/F x 100)





















Source: village officials and household registration records, 2004


* Note: the small number of children recorded from 1995 to 1999, and 2000 to 2003 make it difficult to identify significant trends at the local level. From 2000 to 2003, only 21 children were recorded: 7 boys 14 girls. Only 26 children were recorded for the five preceding years: 17 boys and 9 girls. In such small samples, percentages can fluctuate wildly. In a sample of only 21, the chance of getting 14 or more of one sex (using a probability of .5 for either sex) is close to 20% when the true sex ratio is equal. The same applies to the high sex ratio of 189 for 1995-1999 for only 26 children. The sex ratio of 113 for the total of 130 children is not statistically significant but the direction of the result is consistent with sex selection


Table 1 shows that from 1990 to 1999 quite a few more boys were born. After the 2000 national census revealed the shocking scarcity of girls across China, government promotion of girl children through propaganda may have made local officials and villagers aware of the need to have daughters (or simply to conceal the birth of sons because illegal abortions of females could be inferred from improbably high sex ratios). Given the small numbers of births allowed in any village under the family planning program, it is difficult reliably to identify trends at the local level without referring to larger samples of births or age-sex distributions from more villages or longer time periods. All we can really conclude is that the data for Huang Tu Village are consistent with the larger trends.


The local demand for sons is also reflected in family composition. Using team data from 2004, I calculated the number of households with children who remain sonless and daughterless. Out of 88 households, 66 had children under age 15. In these 66 households, 36 had a child of each sex, leaving 30 couples with children of only one sex. Eighteen couples had no daughter (13 of them had two sons), but only 12 couples had no son. Seven of the sonless 12 had only one daughter and would be able to try again. Only five of the daughterless 18 had just one son and would be allowed to try again. Thus at a minimum 13 families end up with no daughter, and only 5 end up with no son, unless they either have a child outside the quota or adopt one. This gives a rough measure of the number of families that have already broken the rules, as well as those that might want to do so. Nine of the daughterless couples have not borne a child in 10 years, suggesting their reproductive period has ended. Of the five couples who have no sons and already have two daughters, only two have not borne a child in ten years. This suggests that the other three may well try to break the quota.


This exercise in numbers resembles the kind of thinking that family planning cadres across China must conduct if they are to keep village reproduction within the quotas that their superiors require. This team has 17 households with couples who have gone over the quota of two children in the past 17 years. Nine of them had several daughters followed by a son. Only one had two sons first with a daughter as the final child. Of the 12 couples that have not yet had a son, only three are likely to have passed fertility. The others will most likely keep trying despite heavy penalties such as fines and job loss that can accompany out-of-quota births. Assuming half will still have a son, then only eight or nine out of 66 (less than 15 %) couples with children under age 15 will lack a son. Although severe penalties (fines, job loss, confiscation of personal property) may be imposed on those who exceed the quotas, these are often at the discretion of the local leaders who face threats from fellow villagers yet may be punished themselves by higher officials for the inconsistent way they enforce the policy.


The women’s director


In 2004, a central building in the village carried a large banner proclaiming, “Girls are as good as boys.” The women’s director told me, “The population must not increase. We must keep the proportion of births to 1 percent of the population per year. The whole village has about 3,000 people so we can have about 30 children per year.”


The director also affirmed that for everyone, the third birth is the last. “If a woman is pregnant with the third, she must get an ultrasound test (B-chao) and see if it’s a girl. If it’s a girl, she must have an abortion (yin chan). If it’s a boy, she can have it and pay a fine.” This statement suggests that some family planning officials, under heavy pressure to reduce births, are colluding with the private use of ultrasound tests in order to meet their population quotas and allow villagers to meet their demand for a son. These practices obviously contribute to the sex imbalance.


Conversations on family planning and land


One former village official and his neighbor told me, “All the figures for family planning are jia de (false), or biaomian (superficial).” “None of the statistics are to be believed,” they said, “because they don’t report girls.” They also noted government attempts to crack down on the use of ultrasound to check the sex of the fetus, including fining private clinics and destroying their machines. They were also aware of the national campaign to promote girl children with slogans like, “Girl children are the nation’s future.” But these measures ignore the root causes of the pressure on rural people to produce a son. “For city people, it does not matter,” they said. “But rural people still believe they must have a son to guarantee their basic living or survival in old age”[10]. They went on to explain that rural people no longer give up their land when they move to the city. “Even if the whole family moves to the city, having land and grain is a form of basic welfare. As long as rural people have agricultural registration, they keep their village land and rent it out to others, often relatives.”


Land, landholding group, and lineage


“The god of land brings a lawsuit against a farmer who has too many children.” Li Hongkui, fine-arts teacher, Beiguodong Village, Wuzhi county, Henan (People’s Daily Jan. 1, 2000)


Each of the eight teams in Huang Tu is associated with a section of a large rectangle formed by the grid of streets that compose the village. Each team has its own household registration records, conducts its own land allocation, and is responsible for its own family planning. The village as a whole has a village council, a set of leaders responsible for managing local government, with ultimate local authority vested in the Party Secretary.


Land was contracted to individual households in 1980. In 1985 land was re-divided and administered by the eight teams, each of which could reallocate land among the constituent households. Team subgroups (xiaozu) subdivide land according to population changes with readjustments every five years. These are the groups with 20 or 30 households each that readjust the amount of land, drawing lots each five years.


In 1988 the village had 4155 mu of land with about 1.7 mu of land per person. [11] In 1980, when the collective land was divided, everyone — male or female, young or old — had exactly the same amount distributed to her or his family. By 1985, due to family variation in births, deaths, marriages and migrations, differences in land per capita had emerged. Some teams lost members due to official outmigration for urban jobs, such as a man who obtained a government research position in Zhengzhou, and another who studied physics at a university in the United States and later moved to Holland. Others transferred out of the village when they passed the examination to become a state sector teacher, and work in the administrative village school. These individuals may keep their houses or house plots in Huang Tu Village, but no longer receive farm land. The village redistributed land within each team in 1985. Thus, some teams had larger or lesser amounts per capita according to whether they had gained or lost population. In 1989, Team 1 had 2.0 mu per capita, and Team 5 had only 1.3. The village had conducted a second land redistribution in 1990 and at that time the variation ranged from 1.8 mu to 1.3 mu per person, with the average amount of land per capita being 1.55 mu.


In 2004, population growth had reduced the amount of land per capita to 1.25 mu. The team with the least land per capita had dropped to .9 mu per person, while the team with the most had 1.5 mu per capita. The population was then 3207 and farmland had declined to 3953 mu because some land was converted to residential house plots. At the same time, increasing numbers of villagers, certainly a majority of young adults, turned to seasonal and long term nonagricultural work for a large part of their income.


Land policy and population policy in practice


I recently interviewed a farmer who makes a good living from farming about his land. He explained that his household has land for seven people because he has five children. He had three daughters by his first wife who got sick and died (bing si le). The three daughters are 24 or younger, that is, all were born under the birth control policy. All three of these daughters work outside the province in a mobile phone factory. The oldest is to marry a soldier from China’s far west, but soldiers are not allowed to marry until age 25. If Huang Tu Village redistributes land next year, the father will get land for his unmarried daughters for five more years despite the fact that none is in the village or even in the province. He has two sons by his second wife, ages ten and seven.


There are s

Leave a comment