The effect of 3.6 million refugees on crime

Murat G. Kırdar, Ivan López Cruz, and Betül Türküm

Journal of Economic Behavior & Organization, Volume 194 (2022), Pages 568-582


This paper estimates the causal effect of inflow Syrian refugees on crime rates in Turkey. By the end of 2020, the number of Syrian refugees in Turkey had reached 3.6 million, almost all of whom live outside of refugee camps in host communities.

The authors examine variations in refugee numbers and crime rates per 100,000 inhabitants (including natives and refugees) across Turkish provinces and over time for the period 2008-2019. The analysis is disaggregated by category of crime, including: assault, crimes involving firearms and knives, homicide, robbery, smuggling, theft, sexual crimes, kidnapping, defamation, use and purchase of drugs, and production and commerce of drugs. The analysis is based on provincial crime data from the Turkish Statistical Institute (TURKSTAT) together with several additional province-level datasets.

Main findings:

  • Inflows of refugees led to a reduction in aggregate crime rates in host provinces. A ten-point increase in the percentage of refugees in the provinces’ population led to an 8 percent decrease in crime rates.
  • The reduction in crime rates with the arrival of refugees does not result from an increased presence of armed forces (civilian and military personnel) in refugee-hosting provinces. Rather, there is suggestive evidence of a decrease in the per capita number of armed forces in refugee-hosting provinces.
  • The decrease in crime rates is observed across all categories of crime, except for smuggling. The authors suggest that the increase in the rate of smuggling crimes could reflect the prevalence of use of human smugglers to facilitate the entry of refugees from Syria into Turkey, or from Turkey to the EU.

The authors conclude that there is a negative immigration-crime relationship in this context. The authors note that this finding is surprising given that inflows of refugees could be expected to increase crime rates due to several criminogenic factors such as: (a) Syrian refugees are on average less educated and younger than Turkish natives; (b) refugees in Turkey face impediments to accessing the formal labor market and are subject to movement restrictions, leading to a potential skills mismatch; and (c) many Syrian refugees work in the informal sector, displacing low-skilled Turkish workers who may consequently resort to illegal activities. The authors offer several possible explanations for this counter-intuitive result: (i) the threat of detention or refoulement may deter refugees from crime; (ii) employment of refugees in the informal sector and cash assistance programs may provide adequate income, eliminating the necessity to resort to illegal activities; and (iii) many Turkish natives displaced from the formal labor market found employment and increased wages in the expanding formal labor market.