Drought-Reliefs and Partisanship

Abstract

We analyse partisan biases in the allocation of central discretionary transfers in a federal country. We study drought aid-relief in Brazil, where presidential and municipal elections alternate every two years, to identify a novel pattern of distributive politics, determined by the sequence of central and local elections. In particular, we show that alignment advantage materialises only in the period before municipal elections, while it disappears in the period before presidential elections. Furthermore, we show that even before mayoral elections partisanship only counts for districts with intermediate levels of aridity, where being aligned causes an increase by a factor of almost two (equivalent to +18.1 p.p.) in the chances of receiving aid-relief. We rationalise this pattern in a model with office-motivated politicians and rational voters.