Little Emperors Pose Behavioral Challenges

Abstract

In response to our Report “Little emperors: Behavioral impacts of China's one-child policy” (22 February, p. [953][1]), X. Zhao et al. argued that only children resulting from China's one-child policy (OCP) are successful relative to their peers (“China's little emperors show signs of success,” Letters, 22 February, p. [905][2]). However, our methodologies differ from those of the studies that Zhao et al. cite. The papers cited by Zhao et al. rely on a simple comparison of behavioral outcomes between only children and children who have siblings. It is unclear whether the observed behavioral differences (or lack of behavioral differences) reflect parental behavioral differences or the effect of growing up without siblings. The effect of family background would be even more pronounced in families that had more than one child under OCP; these families would likely live in more rural environments, be minorities, or be immune from OCP through wealth or political connections. We addressed this issue by comparing people born just before and just after OCP. Our instrumental variables method allows us to compare individuals who grew up as only children because of the OCP and who would have otherwise grown up with siblings, with individuals who were born just before the introduction of the OCP and hence grew up with siblings. We analyzed and controlled for the possibility of sample bias in education, wages, and age. (More details can be found in the supplementary online materials of our Report.) To clarify the power of our approach, we compared our results with those one would find if one did not account for the endogeneity problem explained above. In the section on “Evidence of Endogeneity” and in table S5 of our Report, we show that without the instrumental variables approach, we would have found differences between the two groups only in the trust game. However, with our approach, we detected that growing up without siblings also has an impact on risk and competition preferences. Our Report also differs from previous studies because it uses experimental techniques in addition to survey questions to measure behavioral differences. We conducted games from the economics literature that are designed to elicit altruism, the ability to trust, trustworthiness, risk preferences, and competitiveness, allowing researchers to observe specific, well-defined types of behavior. Monetary incentives within the games were used, as experimental economists argue that they provide greater motivation for participants to reveal their true preferences ([ 1 ][3]). Our data may thus serve as more reliable indicators of behavior than self-reported surveys. 1. [↵][4]1. D. Davis, 2. C. Holt , Experimental Economics (Princeton Univ. Press, Princeton, NJ, 1993). [1]: pending:yes [2]: /lookup/doi/10.1126/science.339.6122.905-b [3]: #ref-1 [4]: #xref-ref-1-1 View reference 1 in text