Uma aplicação de regressão quantílica para dados em painel do PIB e do Pronaf

Abstract

This study aims to assess the impact of Pronaf on the Brazilian GDP and on sectoral GDP (services, agricultural and industrial) from the application of quantile regression with fixed effects for panel data. Given the criticism regarding the distribution of resources, we also evaluate the different results within the regional context, i.e. the different impacts in the five geographical regions of Brazil (North, Northeast, Midwest, Southeast and South). We use the quantile regression model with fixed effects to panel database, since the variables used in the analysis provide strong inequality between municipalities. By observing the response of each quantile, not just from the average, as well as taking into account the control of fixed effects, the effects of Pronaf on GDP can be better captured. Although the program is national, the different results between regions suggest that (i) changes and improvements in Pronaf should be conducted properly to each region and (ii) it is clear the necessity to decentralize resources.