Magdalena Cerda (New York Academy of Medicine)
03/16/2009, at noon in room 6050 ISR-Thompson.
According to the theory of neighborhood “collective efficacy,” a key factor in lowering neighborhood crime rates is the coupling of (a) social cohesion and mutual trust among neighbors with (b) shared expectations that neighbors will informally take action to maintain public order and prevent crime. To date, this theory has been tested in a handful of cities in the United States, in Stockholm, and (to some extent) in cities in Brazil and China. Still, there are theoretical and empirical reasons to question how generalizable the results of these studies are to cities with very different macro-level conditions. In this paper we analyze the relationship between collective efficacy and (a) neighborhood homicide rates and (b) individuals’ perceptions of violent crime in their neighborhoods in Medellín, the second largest city in Colombia. Our data include measures of neighborhood structural characteristics from census and municipal data, geo-coded police homicide records, and household survey data from 2494 respondents in 166 neighborhoods.
Medellín presents a very interesting deviant case for collective efficacy theory because the poorest neighborhoods in the city also have the most collective efficacy, and collective efficacy is positively associated with homicide (before adjusting for other factors) and with perceived violence. We suggest that one reason for this counterintuitive finding is that paramilitary gangs located in economically disadvantaged neighborhoods deliver important economic and social resources and (perhaps) promote informal social control, but at the same time they also bring violent crime to their neighborhoods. Moreover, we find that after adjusting for the level of economic disadvantage and paramilitary presence in the neighborhood, collective efficacy actually has a protective effect on homicide. Thus, our findings support collective efficacy theory with respect to homicide rates after controlling for the two important confounders: disadvantage and paramilitary presence. At the same time, even after adjusting for level of disadvantage and the concentration of paramilitaries, collective efficacy is associated with higher perceptions of neighborhood violence, perhaps because perceptions of crime spread through social networks/interactions and are heightened in more collectively efficacious neighborhoods. This paper sheds new light on how collective efficacy might operate in a non-Western context, and it also is one of the first studies to incorporate measures of paramilitary gang activity - a big issue in some Latin American cities - in studying neighborhood effects on crime.