Um modelo espaço-temporal para a criminalidade nos bairros de Fortaleza: a influência de fatores comuns e o efeito vizinhança

Abstract

One of the most common analyses in crime studies is the identification of spatial patterns (Almeida, Haddad and Hewings, 2005). However, the spatial dependence found can be influenced by cyclical changes in the state of the economy and by public security policy. The presence of these common factors, may inhibit a good capture of the effect of spillovers that permeate between neighbors (PESARAN et. al., 2015). In this perspective, this work seeks to infer patterns of spatial association of crime in Fortaleza, through a model that accommodates common factors effects and neighborhood effect, estimated by the Principal Component Analysis (PCA) method, proposed by Pesaran et.al. (2015), using the rates of robbery and theft, in the period from 2009 to 2019. The model estimates suggest the existence of a certain degree of influence of common factors on the spatial behavior of crime, concentrated in the Midwest region of the city. It is also verified that spatial clusters of robbery crimes are intensified in the peripheral regions of the city, while for robbery crimes, this association extends to neighborhoods in the prime region, indicating the presence of this type of crime throughout all regions, regardless of socioeconomic status.