Empirical Issues in the Homicide-Income Inequality Argument

Abstract

The argument that income inequality increases homicide rates has provoked scholarly debate, with some studies not supporting this position and providing evidence to the contrary. We identify several empirical issues with the current body of evidence, as well as their underlying problems. We challenge these issues by using more robust techniques than those typical of this literature. Based on the case of Mexican municipalities, we provide evidence that in fact, homicide rates correlated negatively with income inequality between 1990 and 2015, and that this relationship was moderated by levels of socioeconomic marginality. Likewise, we show evidence of spatial dependence in the relationship, thus challenging the assumption that geospatial units of analysis are probabilistically independent of each other.

¿Por qué tanta violencia homicida en América Latina? Caracterizando el fenómeno y expandiendo su marco de interpretación

“…• Por otro lado, Vilalta et al (2023) argumentan que las tasas de homicidio han aumentado en todos los países latinoamericanos independientemente de su nivel de ingresos medios durante la última década.…”

A quantile panel examination of the moderation effects of guardianship on residential burglary

“…When testing the moderation effect of informal guardianship on the relationship between formal guardianship with residential burglary, we must take into account the methodological implications of the data, the selection of variables, and the functional forms of the relationships being tested. For instance, cross-sectional specifications impose restrictive assumptions and limitations, among them, endogeneity bias from different sources such as measurement error, omitted variable bias, and/or simultaneity bias (Vilalta et al, 2022 ). Furthermore, we do not know whether main and interaction effects are constant throughout the residential burglary distribution –we have neither theoretical nor empirical grounds for this assumption.…”