Analysis of Child Gender Discrimination Based on Adults ’ Consumption

The applications of the Rothbarth model of inferring child gender discrimination from the variations in parental living standard have consistently failed to uncover evidence for bias from surveys in countries with some of the world's worst welfare outcomes for girls. This paper demonstrates the importance of the remedies required for an effective implementation of that model with applications to two surveys from urban China and Vietnam. The paper obtains econometric evidence for the presence of child gender bias for the Chinese survey by non-­‐parametric and semi-­‐parametric methods in addition to the standard parametric estimates. The results reported for three categories of adult goods all suggest bias against girls, in contrast to those reported in earlier applications of the model to China. The additional probit estimates of the probability of having a second child conditional on the gender of the first child provide further support for our findings. The two-­‐stage least squares child gender test results for Vietnam, the only study of its type as far as we are aware, also suggest the presence of child gender effects in adults' consumption patterns. The combined test results from the two surveys cast doubts on the previous findings that claim the Rothbarth model of gender discrimination to be ineffective in identifying evidence of bias from consumption patterns.