Replication data for: Conditional Cash Transfers and Electoral Behavior: A Reassessment of a Randomized Experiment in Mexico

In this paper, we re-examine a recent finding that Progresa, Mexico’s conditional cash transfer (CCT) program, mobilized voters in the 2000 presidential elections in favor of the incumbent party. First, after running several sen- sitivity analyses and balance checks, we find that the significance of these results does not hold when two categories of outlier observations are dropped from the dataset. Second, using the R package Amelia II, we impute a handful of missing observations in the dataset that could possibly bias results. When missing data is corrected for, we do confirm the effect of early assignment to Progresa on incumbent vote share in the IV model, a result consistent with the original findings. Third, we create OLS and IV models that account for heterogeneous variation in socioeconomic covariates. The results from the new models disprove the finding that CCTs had the effect of mobilizing voters for the incumbent. Instead, we find that although Progresa’s favorable electoral effects for incumbents do hold, contrary to the existing finding, incumbent electoral gains are not a result of voter mobilization, but rather of vote choice switch- ing away from opposition parties. Moreover, this vote-switching effect is only observed in high poverty precincts and precincts with high voter registration.