Economic approach to intergenerational mobility : measures, methods, and challenges in developing countries

This paper provides a critical survey and synthesis of the recent economic literature on intergenerational mobility in developing countries, with a focus on data and methodological challenges. The attenuation due to measurement error is compounded by sample truncation resulting from co-residency and causes substantial downward bias in intergenerational regression coefficient. In contrast, intergenerational correlation and intergenerational rank correlation are much more robust to such data limitations. To understand heterogeneity across groups, cohorts, and spatial units, reliable estimates of both the intercept and the slope are necessary. The ordinary least squares estimate of the intercept is biased upward, but less so in the rank-rank regression. Sibling correlation is a broader measure of mobility, particularly useful with limited data. Estimating causal effects is challenging because one needs credible exogenous variation in the parental generation and then to track down and collect information on all the children when they grow up. A less demanding but policy-relevant approach is to focus on the causal effects of policies on intergenerational regression coefficient, intergenerational correlation, and intergenerational rank correlation, without trying to disentangle the role of genetic inheritance.