Bombs and ballots : estimating the effect of the Madrid bombings on the March 2004 general elections in Spain

Abstract

Whether or not the Madrid March 11th 2004 terrorist attacks affected the outcome of the Spanish general elections three days later has been the source of great controversy in the last years. This paper analyzes Spanish electoral data for the 2000 and 2004 Congressional elections, comparing the marginal effects of the proportion of voters who voted before the elections (and therefore, before the bombings in 2004) on the voting pattern in both years. A linear approach finds mild evidence that bombs undermined support for the incumbent conservative party and increased the share of the vote for the opposition socialists, similar to previous findings by Montalvo (2006) using a natural experiment design. A non-linear approach using binomial and multinomial logit models is not successful and yields no conclusive indications on how the attacks affected the outcome of the elections.