The Kindness of Strangers: Adopted Girls in China

Abstract

It is well known that sex selection in favor of boys is taking place on a large scale in China. Less well studied is abandonment and adoptions of girls in China. The 1980s saw a surge in internal adoptions, reaching four percent of girls in the late 1980s, a development that was hailed as evidence that sex ratio imbalance in the 1980s overstated excess mortality of girls (pre or post natal), since abandoned girls are often omitted in self reported fertility histories. In this paper we seek to address two questions: first, why are girls adopted in a society marked by son preference; and second, did abandonments continue in the 1990s at the levels observed in the late eighties despite greater use of prenatal selection (120 boys per 100 girls were born between 2000 and 2005)? We investigate three hypotheses as to the motives for girl adoptions. One, parents may simply want a child, and girls make up 80 to 90 percent of abandoned children. Two, parents may want a girl, either because they have a preference for a girl, or they have a preference for children of both sexes and they already have a boy. Third, parents may prefer a girl for her services, notably domestic chores or securing a future daughter-in-law for a son, bride shortage looming large in China's future, with more than 25 million more males than females under the age of 20. It can be noted that the adoption of a little-daughter-in law was a common practise in the Pearl River delta and Taiwan well into the 20th century. To distinguish these hypotheses, we study adoption patterns using the 1992 National Sample Survey of the Situation of Children. We find that daughters are predominantly adopted by families with two older own-birth sons, evidence against the first motive of adoption of girls resulting from a gender neutral desire for a child. To separate the second and the third hypotheses, we study the educational outcomes of adopted girls and find that for children 8-13 years old -- an age when school enrolment and attendance reflect parental priorities more than child aptitude -- adopted girls were substantially less likely to attend school than own-birth children or adopted boys. Furthermore, we find adoption of girls to be particularly common in the two provinces where the adoption of little-daughters-in-law used to be a particularly prevalent form of marriage: Fujian and Jiangxi. These two findings we take as evidence for the third hypothesized motive: girls are adopted to serve, possibly including serving as a future-daughter-in law, or at least hedge against future bride shortages. Arguably, poor child outcomes could be attributed to adopted children being negatively selected. However, for primary education, the case can be made that this is a parent controlled outcomes. Moreover, the Chinese preference for sons suggests that adopted girls are likely no more negatively selected than adopted boys. Anecdotal evidence suggest that girl abandonments are triggered by the existence of an older sister, a circumstance largely exogenous to unobservable individual characteristics. Further to this argument, we show that adoptions (and presumably abandonments) of girls, but not boys, increased more in the 1980s in provinces that practised more strict birth-planning policies. To address the question of continued prevalence, we use the 1990 and 2000 censuses and note that sex ratios were unnaturally female in both the 1990 and 2000 censuses following sons. For instance, in the 2000 census, following two boys there were 0.66 boys/girls at third parity. Since there is no evidence that boys are aborted based on sex, these girl biased sex ratios may provide a lower-bound estimate of girl adoptions, and consequently, of girl abandonments. Employing this proxy, we estimate that adoptions continued through the 1990s at the level recorded for the late 1980s -- that is, three to four percent of girls were abandoned and raised by strangers. How these girls fared is an open question.