Effects of Refugee Inflow on the Voting Behavior of Natives: Application of Mobile Phone Data

Abstract

This paper investigates the effects of refugee inflow on the voting behavior of natives. I employ a difference in differences strategy to identify the voter reaction against the sudden inflow of Syrian refugees to Turkey and use unique data provided by Turk Telecom to proxy natives' refugee exposure. Data coming from mobile phone usage track the mobility of refugees throughout the year and enables the construction of refugee exposure measures across small geographical areas. I show a small drop in the government party vote share that announced an 'open door' policy for refugees. Furthermore, there is evidence for the reaction heterogeneity based on the socio-economic development levels of refugee-hosting areas. While there is a null response in ethnically Kurdish and economically lagging East, a negative reaction is concentrated in socioeconomically more developed areas. Then, exploring the effects in rural and urban areas reveals an interesting divide in the refugee exposure patterns of hosting areas. While in rural areas, natives react against the refugee presence, in urban areas, they react when the mobility (visibility) increases. Anecdotal evidence might help to put findings in context. Manipulating the concerns of voters on Syrians' receiving citizenship and demographic change might be regarded as two main strategies used by opposition parties. These strategies help convince the secular and conservative-nationalist voters in socioeconomically developed areas. (JEL codes: J01, J08, and J15)