Exit polls and voter turnout in the 2017 French elections

Résumé 0

Belgian and Swiss media regularly interfere during French elections by releasing exit polls before polling stations close. These foreign media profit from a law forbidding the same behavior by their French counterparts to receive large inflows of web visits from France. We exploit the unusual timing and degree of confidence with which exit polls were released in the second round of the 2017 presidential elections to investigate their effect on voter turnout. Our analysis is based on comparing turnout rates at different times on the election day, in the first and second round, and with respect to previous elections. We find a significant decrease in turnout of around 3 to 4 percentage points after the exit polls' publication which is suggestive of a causal effect, although similar trends were observed in previous elections. The effect is stronger in departments close to the Belgian border shortly after the release of the exit polls. We do not find clear evidence that either candidate benefited from the decrease in turnout, yet we cannot exclude the presence of a small underdog effect which reduced the winning margin by around 1 percentage point.