Ne parlons pas de l’affaire Dreyfus: Antisemitic riots and radicalization in Fin de Siècle France

Abstract

There is mixed evidence on whether riots lead to more support for or a backlash against the political side perpetrating the violence. While most recent literature has focused on contemporary left-wing riots, and the role of the media and parties in shaping public reactions, we look at right-wing riots in late 19\textsuperscript{th} century France, before mass media and mass parties. We use a combination of historical municipal-level election data and newly collected information on the location of antisemitic riots in early 1898, which followed the publication of Émile Zola’s \emph{J’accuse} in the context of the Dreyfus Affair. Using a difference-in-differences design -- comparing municipalities where riots occurred to those were riots did not -- we show that the riots caused an increase of about 5\% in the vote share in the 1898 elections of extreme right candidates at the expense of centre right candidates. These results are consistent to different specifications and robustness tests. The increase in votes for the extreme right are partially driven by an increase in turnout, but not by anti-incumbent voting, which suggests that the riots caused a shift in opinions amongst right-leaning voters. These results contribute valuable insights into understanding the consequences of political violence, suggesting that riots against an ethnic minority group can contributed to the mobilization of the extreme right, and that this effect already existed in the early years of democratic elections.