Drought declarations and voting outcomes : evidence from Tunisia

Using a novel dataset in a difference-in-differences framework, I study the political implications of the change in policies protecting drought-impacted communities in Tunisia in 3088 districts. In 2016, the Tunisian government had drastically reduced support to farmers affected by droughts. I show that this resulted in lower vote share for the incumbent candidate in the second round of the 2019 presidential elections. The magnitudes are substantial: a one standard deviation increase in the district-level damage caused by droughts after the policy change leads to a decline in the incumbent's vote share by 10 percentage points. I show that the likely mechanism is the rise of public discontent: the drought-affected districts had a higher number of riots or violent protests relative to the policy change. I find no significant effects either on voter turnout, closeness to incumbent party at the parliamentary elections or on trust in political parties at the parliament.