Infrastructure for Votes? Experimental & Quasi-Experimental Evidence from Mexico

• In Event: The Political Economy of Voting

Abstract

A core assumption in the political accountability literature is that voters reward incumbents for policy. For the case of public works infrastructure, causal evidence is scarce. We examine the electoral returns to a randomly allocated federal program in Mexico which spent USD $68 million on neighborhood infrastructure improvements between 2009--2011. We find that while the program increased turnout in the 2012 elections, it had a precise zero effect on precinct-level vote share for municipal or federal incumbents. We then examine the non-experimental scale-up of the program. Using a difference-in-differences approach, we find that the program increased the municipality-level vote share of incumbents from the party in federal office. One possible explanation for this divergence between precinct- and municipal-level effects is that the program was a source of irregular campaign funds for municipal candidates from the party in federal office. We find suggestive evidence in favor of this explanation using a close-election regression discontinuity design. Partisan alignment between the municipal and federal government increased allocation of the program and also increased the amount of misused federal funds in the municipality, as measured by an independent agency.

Authors

• Francisco Garfias, University of California, San Diego

• Bruno Lopez-Videla, UC San Diego

• Wayne Aaron Sandholtz, Nova School of Business and Economics