Tuesday, July 07, 2009
Better proxy for human capital variable
Got problem in choosing good proxy for human capital in your estimating regression? Try school intensity (Duflo 2001, Comola and de Mello 2009 -- thanks for the paper, guys). It is the share of population born after 1963 multiplied by the number of schools built per children under SD Inpres program during 1973-78. The rationale is simple: the cohort of individuals born in districts that benefited from the program was more likely to stay longer at school and to earn more once joining the labor force.
Posted by Aco at 9:06:00 AM