Abstract
We combine a model of symmetric information with selfish and office-motivatedpoliticians and a Regression Discontinuity Design analysis based on close municipalelections to study partisan bias in the allocation of drought aid relief in Brazil. We identifya novel pattern of distributive politics whereby partisan bias materialises only beforemunicipal elections, while it disappears before presidential elections. Furthermore, beforemayoral elections, it fades for extreme (high or low) aridity levels while persisting formoderate levels. Our empirical results show that in this case alignment increases theprobability of receiving aid relief by a factor of two (equivalent to 18.1 percentage points).