This thesis has two main objectives: (i) to analyze empirically, using data from municipal elections in Guatemala for the period 1999-2011, the effect of incumbency on the probability of being reelected and (ii) to explore the possible cause of the effects obtained from the estimations. To do this, a regression discontinuity estimation strategy is implemented exploiting the inherent discontinuity in the system of simple majority voting. Paradoxically, the findings show that far from being an advantage, incumbency seems to be an electoral disadvantage. The result extends to political parties. In the second part, the results are mixed, but suggest that elections seem to play an accountability role, but also are an important mechanism to discover the type (or quality) of politicians. This work is an important contribution to the limited literature on the dynamics of local elections in Guatemala and opens the door to future research issues.