Indicadores socioeconômicos como determinantes do nível de corrupção nos municípios brasileiros: uma análise a partir de regressão espacial

Abstract

Regional analysis of the determinants of corruption makes historical, cultural and political factors more homogeneous, making socioeconomic factors as the most important. This dissertation aimed to determine and analyze the socioeconomic factors of corruption at the municipal level in Brazil. For this, it was used data from a anti-corruption program based on the random auditing of municipal government’s expenditure, called in portuguese Programa de Fiscalização de Recursos Federais a partir de Sorteios Públicos, implemented by the Controladoria Geral da União (CGU) of municipalities audited between 2006 and 2010. For this analysis we use the evaluation by a spatial regression model. The results obtained from the study show a high spatial dependence in the data, indicating that in this case, the spatial regression model is more indicated. Moreover, the most significant variables as determinants of corruption is poverty (measured by the number of beneficiaries of the Program Bolsa Família), the illiteracy rate of people between 7 and 14 years and the value of GDP of agriculture, which correlated positively with the level of corruption. Still, two other socioeconomic variables showed statistical significance in at least one of the models tested, the investment expenditure, with negative correlation, and the share of women between 10 and 14 years who had children, with a positive sign.