Pierre ANDRE

We estimate the extent to which international aid projects are subject to favoritism using differences in differences and RDD estimates based on presidential transition dates. We compare two different sources of aid, China and the World Bank, and two different types of favoritism based on the president’s region of birth or ethnic group. Consistent with the literature, we find that Chinese aid is massively targeted to the region of birth of new presidents. However, we also find evidence of a less visible and less intense form of favoritism in World Bank aid: it targets areas that are co-ethnic with a new president, without following major regional administrative boundaries. This is consistent with a better control of the World Bank aid. Moreover, the political economy of World Bank aid favoritism appears to be more complex: World Bank aid favoritism is more intense in dictatorships, when Chinese aid is more abundant, and in countries that have historically received more World Bank aid.

Link to the paper.

Link to the Working Paper.

Has secular education contributed to the decline of polygamy in Africa? To answer this question, we study a wave of public school construction in late-colonial Cameroon. Our difference-in-differences and event-study specifications show that school openings have simultaneously increased education and the chances to be in a polygamous union for men and, more surprisingly, for women. We estimate a structural model of marriage to explain why education made women more likely to be in a polygamous union. The main estimated channel is marriage to educated men who are more often polygamists than uneducated men, not direct preferences for polygamy.

Link to the paper.

Link to the last Working Paper.

We estimate whether opening of new schools increases educational attainment and affects civic engagement, political participation, and political selection in Malian villages. We compare the differences in educational attainment between individuals below and above the age of 9 as of school opening date by using a donut regression discontinuity design. Opening of schools drastically increases school enrollment. Using this exogenous variation in school enrollment as an instrument, we show that education increases participation in village associations, involvement in local political life, and the number of elected politicians from a village. Most of the effect of education is concentrated among individuals belonging to well-established families. This suggests a redistribution of roles to the dominant group of a village.

Link to the paper

Link to the last working paper

In developing countries, the opportunity costs of children’s time can significantly hinder universal education. This paper studies one of these opportunity costs: we estimate the agricultural productivity of children aged 10 to 15 years old using the LSMS-ISA panel survey in Tanzania. Since child labor can be endogenous, we exploit the panel structure of the data and instrument child labor with changes in the age composition of the household. One day of child work leads to an increase in production value by roughly US$0.89. Children enrolled in school work 26 fewer days than nonenrolled children. Compensating enrolled children for loss in income can be accomplished with monthly payments of $1.92. However, a complete simulation of a hypothetical conditional cash transfer shows that even $10/month transfers would fail to achieve universal school enrollment of children aged 10 to 15 years old.

Link to the paper

Link to the working paper

Link to the paper

Link to a more complete working paper

See here for details