Educai os jovens para que não seja preciso punir os adultos: a relação entre o desempenho escolar e os crimes nas cidades brasileiras

Abstract

The present research has as its main objective to provide empirical evidence of the relationship of school performance to crime in Brazilian cities, in particular homicide crimes, classified as violent crimes against the person, and those of robbery and theft, considered crimes against property, using cross-sectional data from 2007, 2008, 2009 and 2010. We chose to address the crime problem from the point of view of prevention policies, using the lagged data from the high school National Institute of Educational Studies and Research Anísio Teixeira (INEP), in an attempt to understand the school performance in Brazilian cities, and its relation to the offer of crimes. The estimations were also in line with the Ecological Model, in which the historical process of individuals and the environment may influence on their decision of illicit practices. We used the method of Ordinary Least Squares (OLS), Bootstrap, Jackknife and Quantile Regressions (QR), using the variables of school performance and public security expenditures lagged, in addition to the other important variables in the understanding of crime. The criminal data was provided by the Department of Education, Research, Information Analysis and Personnel Development of the National Secretariat of Public Security, and the rest originated from sources such as the Applied Research Institute and the Brazilian Institute of Geography and Statistics. The results indicated that for the Brazilian average case there is evidence of school performance on crime rates, in particular, for robbery and theft. In the same way, it was observed that the most part of the variables showed signs and expected impacts as also estimated in the studied literature. The estimations using the Bootstrap and Jackknife methods corroborated the significance of the parameters since they provide more precise confidence intervals for the estimated parameters. Likewise, it was observed regarding the quantile analysis, that for the crime of theft those municipalities with higher crime rates had a greater effect of the education in the reduction of their criminal levels, having a decrease between 1.6 and 2.6 crimes per 100,000 inhabitants, facing a rise of 1% on high school approval rates. For the crimes of robbery and murder, the effect of the educational variable is small in the lower quantiles, i.e., in the cities with lower criminal rates, however in those cities it was verified higher criminal inertia. It was also pointed out that there is a directly proportional and statistically significant relationship between the lagged educational variable and crime in subsequent years.