Do larger immigrant populations increase homicide rates? : a global study, 1995-2014

Abstract

Relating high immigration to violent crime has been a sensitive topic in the media for quite some time, often concerning highly populistic politics. Still substantial research that validates or disproves this assumed relationship is missing, and that is the motivation behind this thesis. More specifically I examine the statistical relation between immigration stock and the reported homicide rates in countries worldwide. The results show a small positive association for a world sample, but other variables seem to be more important. The effect of immigration on homicides seem to be conditioned on the access to political power among the lower socioeconomic groups in society. Looking explicitly at the OECD countries the effect goes away. I use data from 1994-2014 for over 150 countries globally and 33 countries in the OECD. Country- and time fixed effects are included in OLS regressions both with Newey-West robust standard errors and spatial correlation (Driscroll-Kraay) in STATA for both samples. The results taken together do not support the simple view that a larger share of immigrants increase violent crime.