Using a Semiparametric Estimator to Forecast Education Outcomes in Nicaragua's Red de Proteccion Social

This paper uses baseline data from the randomized experiment of the conditional cash transfer program - Red de Proteccion Social, Nicaragua to conduct an ex ante evaluation and compares results to those of the experimental evaluation. Reduced form estimation of a behavioural model using a health production framework forms the basis of the evaluation. A Klein and Spady semi-parametric single index model is used to predict unobserved outcomes under the treatment. The sample consists of children aged 7-13 who have not completed grade 4 of primary school. The evaluation shows that the ex ante approach closely matches the experimental outcomes in most cases. ex ante evaluation conditional cash transfers single index models education