This paper investigates whether the relocation of approximately 6,400 migrants from the Calais “Jungle” to temporary migrant-centers (CAOs) in France affected votes for the Front National in the 2017 presidential election. The assignment of CAOs was not random, and was likely to be endogenous to political outcomes; citizens of volunteering cities are arguably more tolerant toward migrants and less likely to be affected by the presence of a CAO. To isolate exogenous variation in the exposure of French municipalities to relocated migrants, the authors instrument the presence of a CAO with the presence and size of holiday villages in the same municipality. (One of the criteria for locating CAOs was potential additional space for migrants in holiday villages. The stock of holiday villages was determined before the current migrant surge that led to the creation of the CAOs). Key findings:
- The presence of a CAO had a negative effect on the vote share of the far-right party Front National. The presence of a CAO reduced the vote share increase of Front National by about 15.7 percent (about two percentage points) in those municipalities. At the same time there was an increase in the share of votes received by the far-left party Front de Gauche, which has a more open stance towards migrants, but similar political platform on other issues
- Spillover effects of the presence of the CAOs on neighboring municipalities. Municipalities within a five km radius decrease their support for the Front National by around 1.8 percent.
- A stronger decrease of vote shares of the Front National in municipalities with a larger share of younger people and existing migrants. The authors suggest this may be the case because younger people have “less fortified” opinions towards migrants and so may be more willing to connect with new people joining their municipality; and pre-existing communities from the same country of origin of the migrants could facilitate initial contact.
- Effects are dampened in municipalities that were exposed to larger numbers of migrants and where the mayors volunteered to welcome them. In municipalities that had over 39 migrants per 1,000 residents the impact on the Front National vote outcome is positive, a result consistent with findings on large inflows of migrants. The authors posit that citizens living in volunteering cities are less likely to be prejudiced against migrants, so actual contact with them is less likely to affect their political choices.
- The arrival of migrants does not appear to have affected local economic activity. While migrants in CAOs do not legally have the right to work and do not receive any monetary allocation, their arrival might have an effect on local economic activity through increased demand in the catering or building sectors; these potential variations in local economic activity might affect electoral outcomes. The authors do not find any significant difference of net job creation per inhabitant over the period 2012-2014 between municipalities that received a CAO and those that did not.
The authors conclude that the results are consistent with the contact theory (Allport, 1954). Citizens in exposed municipalities developed a greater degree of acceptance towards migrants and were therefore less likely to vote for the Front National. However, the electoral response to actual migration seems to depend crucially on the size of the inflow.