Spillovers of Conditional Cash Transfers across Localities: evidence from Progresa

The conditional cash transfer program Progresa (now Oportunidades),implemented in rural areas of Mexico from 1997 onward, is likely to have spillovers across villages on the education outcomes of many children who reside in non-beneficiary localities but attend secondary schools located in beneficiary localities. The effects of these spillovers are examined by breaking up the control group of the randomized evaluation, performed between 1997 and 1999, by location of the nearest secondary school into a group of localities that was exposed to spillovers and another that was not. The increase in the incidence of spillovers associated with the scaling-up of the program is exploited to identify the effects of these spillovers on the education outcomes of the children residing in control localities. Double and triple-difference estimators are implemented in order to control for the selection due to the non-random incidence of spillovers. Positive spillovers are found on both the attendance and attainments of boys at the secondary level of schooling. Some additional evidence on the competing mechanisms suggests that these spillovers are consistent with peers effect in secondary schools as well as with the effects of the local development spurred by the program.