Uma contribuição para prevenção da criminalidade

Abstract

This thesis comprises three articles about criminality. Article one, Urban violence: a comparative analysis of victimization between São Paulo, Rio de Janeiro, Recife and Vitória, reviews the routine activities approach trying to incorporate some elements of rational choice models. The empirical analysis estimates the probability of victimization. The database is a victimization survey conducted by FIA/USP and ILLANUD, in 2002, which was carried out in the Brazilian capital cities of Recife, Rio de Janeiro, São Paulo and Vitória. I consider five categories of victimization: aggression, robbery, theft, burglary and auto-theft. The more important factors to explain victimization are both the proximity between the victim and the offender and the victim attractiveness. The comparison between these cities shows that Recife is the one with greater probability of victimization, followed by São Paulo, Rio de Janeiro and Vitória. Article two, Evaluation of Fica Vivo Program in the Brazilian City of Belo Horizonte, carries out an evaluation of a State of Minas Gerais program aimed at preventing homicide in this Brazilian federation unit. In the case of the capital city, Belo Horizonte, it focus on the slums areas and currently it covers the dwellings of Morro das Pedras, Pedreira Prado Lopes, Alto Vera Cruz, Taquaril, Cabana de Pai Tomás, Ribeiro de Abreu e Conjunto Felicidade. As database I use both the geographical information system of the Minas Gerais police and the Brazilian Demographic Census of the National Statistics Bureau IBGE. The econometric procedures are based on double-difference and double-difference matching methods. The impact variable to be explained is the homicide rate, from 2000 to 2006, annually broken down into two six-month periods. We divided the program into four uneven periods according to each area expansion date. In general the results show differentiated impacts on the homicide rate for each slum. Whereas the most relevant impact was in Morro das Pedras, which is the pilot program, the impacts in the remaining slums are less relevant although statistically significant. Finally, article three, Economic Evaluation of the Homicide Prevention Program in a Slum Area of the City of Belo Horizonte, undertakes an economic evaluation of Fica Vivo Program in the slum of Morro das Pedras in Belo Horizonte. The cost-effectiveness and cost-benefit analyses were run for the 2004-2006 period. The costs were accounted for both the aggregated slums and Morro das Pedras, individually, based on information by the Social Defense Secretary (SEDS) and the State Police (PMMG). The main results show the existence of scale economies in program implementation and social protection actions. If one takes into account conservative estimates of the program costs, the average cost of Fica Vivo by beneficiary is about one third of the value spent by the Federal government for each beneficiary of the National Social Program of Income Transfer (Bolsa Família). Effectiveness is measured according to the methodology carried over in article two based on double difference modeling, taking the homicide ratio per one hundred thousand inhabitants as the impact variable. Using a conservative estimate, the cost of one homicide prevented by Fica Vivo is about one hundred forty four thousand dollars and its return rate is 99%. Compared with international evaluations of similar control and prevention programs the Fica Vivo Program is within the ones of greater return to society.