A two-stage control function approach with settlement-specific residual variances to identify effects of family size on school achievement in Ouagadougou

Résumé

Abstract
In Ouagadougou (the capital of Burkina Faso), at least 33% of the 2.5 million inhabitants live in informal settlements, while the remainder lives in the formal settlements. The sub-population in the informal settlements lacks social infrastructure and has extremely low education attainments. Motivated by this specific context, we develop a method that can be used to investigate differences in family behaviour towards schooling in the two settlement types and, in particular, the trade- off between child quantity and quality. Our method uses a two-stage control function approach that allows for settlement-specific residual variances as well as for the endogeneity of both the number of children and the settlement choice. After controlling for a set of observable covariates, we find evidence of a significant non-linear quantity-quality trade-off in the informal settlements. We also find evidence that significant differences in quantityquality behaviour exist between the two settlements, but that the differences depend on both the education outcome considered as well as the gender of the child.

JEL subject classification : I25, I24, J13, J16, J18, C36.