Do actions speak louder than words? Evidence on voter behavior from Madrid Rio Park

Abstract

In this paper we estimate the electoral gains and the timing of the gains, from the promise to build and the subsequent construction of a large infrastructure project, Madrid Rio Park, which was promised in the 2003 Madrid Mayoral election. We use as a case study the construction of the Madrid Rio Park because its unique circumstances provide us with a quasi experiment that allows us to causally separate the prospective and retrospective behavior of voters regarding electoral pledges. We find that voters behave mostly retrospectively. The retrospective effects are sizable in magnitude: we find that after the Madrid Rio Park was completed voters rewarded the incumbent with an additional three percentage points. Unexpectedly, but interestingly, we find only weak evidence of prospective behavior: voters do not strongly react to the promise to build the infrastructure.