53904 China's "Missing Girls" -- Son Preference or Hepatitis B Infections? The data on gender imbalance support the cultural rather than the biological explanation C hina has very large numbers of "missing girls," sex-selection technology (when households resorted to resulting in significant gender imbalance in the postnatal discrimination), show excess mortality con- population. Government's policies have sought centrated among girls born into families that already to change the cultural preference for sons which is be- have a girl. lieved to influence sex-selective abortion, but a recent After sex-selective technology became widely avail- study suggested that up to 75 percent of the "missing able, the South Asian and Korean data indicate that the girls" in China is due mostly to a high prevalence of use of this technology is strongly correlated with the Hepatitis B infection.1 If true, then the first plank of sex composition of existing children--just as in China. policy should be immunization programs. World Bank Across East and South Asia all the indicators--sex ra- research confirms, however, that son preference is the tios at birth and sex ratios of child mortality--show main factor behind the "missing girls" and policies do that whether or not girls "go missing" is determined by need to focus on changing the cultural roots of son the sex composition of children in the family. Again, preference. this suggests that son preference is the predominant underlying factor. The Hepatitis B hypothesis is difficult to reconcile For Hepatitis B prevalence to be the culprit, women with demographic and historical data who have borne a daughter would have to be especially Data from a large dataset show that the only women with prone to contracting the disease. Or the disease would elevated probabilities of bearing a son are those who have have to lead somehow to women first bearing daughters already borne daughters (fig.1).2 Those who have borne followed by an excess of sons. Either of these scenarios only sons show a mildly elevated probability of the next requires a complex set of biological factors. Is it possible child being a girl--indicative of a mild preference for that Hepatitis B works in these very complex ways? having a daughter if the sons are already safely in place. Historical data for China show that the fluctuations A similar pattern of son preference is found across in sex ratios during the 20th century correspond to pe- Asia. Studies in India, Bangladesh, and the Republic riods of severe resource constraints (fig. 2). Sex ratios of Korea, based on data from before the availability of rose sharply with wars and famines in the first half of the century, and overall fertility declined in the latter decades of the century. It is implausible that the preva- lence of Hepatitis B infection fluctuated in tandem with Figure 1. Probability of bearing a son, by sex composition of these political shifts. woman's existing children, China 1989­90 Medical evidence suggests the impact of Woman's first child: Hepatitis B on sex ratios at birth is marginal (53% of births) normal probability of A national longitudinal dataset collected in Taiwan having a boy (China) from 1988 to 1999 permits a robust test of the impact of Hepatitis B infection on sex ratios at birth. Woman has only boy(s): Woman has only girl(s): However, the data show that the infection raised wom- (18% of births) (21% of births) en's probability of having a son by only 0.25 percent, 8% reduced probability of 36% excess probability of next birth being a boy next birth being a boy suggesting that a 15-percent disease prevalence would raise the overall sex ratio at birth from a baseline of 105 Woman has both boy(s) and girl(s): to only 105.165.3 Since the sex ratio at birth for all births (8% of births) in the dataset was 109, Hepatitis B infection accounts for 12% excess probability of next birth being a boy just a tiny part of the female deficit. The sex ratio at birth rises with birth order in Tai- Source: Das Gupta (2008).put at the regional level, measured as number of patients per 1,000 inhabitants. wan (fig. 3). The panel data indicate that the impact of September 2008 econ.worldbank.org/research Hepatitis B is fairly constant across birth orders, again Figure 2. Excess of males to females, by five year cohorts, China 1920­95 rejecting the hypothesis that Hepatitis B infection ex- plains the elevated overall sex ratio at birth. Percent excess of males to females Another hypothesis has been put forward that 20 Hepatitis B skews the sex ratio of births not through the mother's infection status, but through the father's.4 16 Internal wars Communist However, neither premise explains why only women government 12 established Fertility with daughters have elevated probabilities of bearing decline Great Leap a son. Nor is either premise supported by the fact that Japanese 8 invasion Famine Sub-Saharan African countries with a high prevalence of Hepatitis B have normal sex ratios at birth. Finally, 4 a very recent review of data from China now confirms that Hepatitis B has little impact on the sex ratio at 0 birth.5 It appears the Chinese government has been cor- 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 rect to assume that son preference is the main factor First year of 5-year period of birth behind the "missing girls" and to focus their policies on Source: Monica Das Gupta and Li Shuzhuo. 1999. "Gender Bias in China, South Korea and India changing the cultural roots of son preference. 1920­1990: The Effects of War, Famine, and Fertility Decline." 30(3):619­52. Figure based on data from the 1953, 1964, 1982 and 1990 Population Censuses of China and the 1995 National One Percent Sample Survey. Monica Das Gupta, Senior Social Scientist mdasgupta@worldbank.org Notes Figure 3. Sex ratio at birth, Taiwan 1988­99 1 Emily Oster. 2005. "Hepatitis B and the Case of the Missing Sex ratio at birth (male to female) Women." Journal of Political Economy 113(6):1163­1216. 116 2 Monica Das Gupta. 2008. "Can Biological Factors like Hepatitis B Explain the Bulk of Gender Imbalance in 112 China? A Review of the Evidence." World Bank Research Observer 23(2): 201­17. 3 Ming-Jen Lin and Ming-Ching Luoh. Forthcoming. "Can 108 Hepatitis-B infection Hepatitis B Mothers Account for the Number of Missing Women? Evidence from Three Million Newborns in Tai- No Hepatitis-B infection 104 wan." American Economic Review. 4 Baruch Blumberg and Emily Oster. 2007. "Hepatitis B and Sex Ratios at Birth: Fathers or Mothers?" Processed. 100 1st 2nd 3rd http://home.uchicago.edu/~eoster/HBVfathers.pdf. Birth order of child 5 Emily Oster and others. 2008. "Hepatitis B Does Not Ex- Source: Derived from table 2 in Lin and Luoh (Forthcoming). plain Male-Biased Sex Ratios in China" Processed. http:// home.uchicago.edu/~eoster/hbvnotecon.pdf Human Development and Public Services Research econ.worldbank.org/programs/hd_and_public_services