ONLINE APPENDIX: Financial Incentives and the Fertility-Sex Ratio Trade-off

This section presents a stylized discrete-choice model to derive predictions about the effects of a financial incentive program that rewards parents that have fewer children and more daughters. For simplicity, assume that (i) the program provides benefits only to parents that have one child, conditional on no second births, and (ii) the benefit received by parents of one girl ($G) is larger than the benefit received by parents of a boy ($B).1 Suppose this program is announced in year t. Couples that have no children or one child at the time of announcement are, thus, “treated” by the program, whereas the rest are not. However, differences in the future outcomes of couples with dissimilar child compositions in year t cannot be solely attributed to variation in the program’s incentives. For example, one-boy families may be less likely to sex-select at higher parities than daughtersonly families even in the absence of incentives. Moreover, characteristics such as son preference that determine the child composition in year t are also likely to influence the effectiveness of incentives in altering childbearing behavior. Thus, the child composition at the start of the program, although pre-determined, is certainly not exogenous.