On Explaining Asia's “Missing Women”: Comment on Das Gupta

IN AN EARLIER issue of this journal, Monica Das Gupta (2005) comments on a recent paper of mine. Her comment was based on Oster (2005a), a work-ing paper version of Oster (2005b); in general, the two versions do not dif-fer, but I will distinguish between them when they do. In that paper, I at-tempt to connect the issue of the gender imbalance in Asia, the "missing women", with the prevalence of the hepatitis B virus. I present a wide va-riety of evidence suggesting that women who are carriers of the hepatitis B virus give birth to more male children than do noncarriers. I argue that perhaps as much as 45 percent of the gender imbalance observed in the Sen (1992) missing women populations in the period 1980-90 can be ac-counted for by hepatitis B. Further, I argue that the explanatory power var-ies significantly across space: 75 percent of the missing women in China are accounted for, versus around 20 percent in India. The connection between hepatitis B and sex ratios at birth relies on existing individual-level studies as well as on new analyses: a natural ex-periment based on recent vaccination campaigns and cross-country evidence (Oster 2005a, b). Hepatitis B (a viral disease of the liver) is common in China, with 10-15 percent of the population infected before vaccination. I argue that there is evidence that the sex ratio imbalance in China arises at birth, and, in addition to presenting evidence on the sex ratio-hepatitis connec-tion, I conclude that some of the sex ratio imbalance is naturally occurring. The analysis in Oster (2005a, b) is much more extensive than the short summary I present here, and it includes important caveats, robustness checks, and alternative counterfactual analyses. In her note, Das Gupta argues that the hepatitis B explanation is un-likely to be important, since sex ratios in China differ over time and among