Estimating an educational production function for five countries of Latin America on the basis of the PISA data

This paper takes a new look at the determinants of cognitive ability. Using the results of the 2006 PISA survey for five Latin American countries (Brazil, Chile, Colombia, Mexico and Uruguay) it adopts an efficiency analysis perspective. Rather than selecting as inputs a few variables, it integrates the maximum amount of available information via the use of correspondence analysis (CA). Efficiency is estimated at the individual level via corrected ordinary least squares and is then regressed on explanatory variables obtained again via correspondence analysis. Unobserved school effects are taken into account by estimating a random effect model. Finally, using the so-called Shapley decomposition, we determine the exact impact on individual efficiency of each of the factors that are considered as determinants of this efficiency.