Presented at the 2012 meeting of the Midwest Political Science Association.
Standard explanations of voters' compliance in vote buying assume either that parties circumvent the secrecy of the ballot or that voters' political inclinations are observable. Neither of these assumptions hold for many elections in developing democracies. I present a model in which the buyer sustains compliance without knowing voters' preferences or vote choices. Compliance is sustained by conditioning future bribes on whether the party's votes reach an optimally-set threshold. This mechanism generates a collective action problem among bribed voters that is consistent with previous findings: bribed voters comply if they believe others are doing the same, and compliance is harder to sustain in large populations. The model also shows that it is easier to induce compliance of voters whose utility depends on the welfare of other voters. This accounts for the observed tendency of parties to target groups with strong social ties among their members. pdf.
Presented at the 2013 meeting of the Southern Political Science Association.
Corrupt politicians use a variety of manipulation strategies to win elections. How do they choose between them? While the current literature on electoral manipulation has concentrated on voters' characteristics that affect the incidence of one form of manipulation, this paper studies the effects of election-level factors on the relative incidence of distinct strategies. Using a new data set of citizens' and election monitors' reports of electoral crimes in Colombia, I find that larger numbers of voters per polling station reduce reports of vote buying, but do not affect the number of reports of restrictions on turnout. The finding lends support to the claim that vote buying in secret ballot elections can not be sustained when results are given at high levels of aggregation. A separate finding is that the size of the electorate has a larger negative effect on vote buying than on fraud or restrictions on turnout. This is consistent with vote buying having larger marginal costs of implementation than that of alternative methods.
I present a model of civilian cooperation with an armed group in an irregular war. Unlike previous models of civilians' interactions with combatants, in this model civilians consider the effect of their cooperation on territorial control in an incomplete information setting where other civilians' motivations and cooperation choices are unobserved. There are three main results derived from the model: first, superior military force is not sufficient to achieve full civilian cooperation, which is attained only if military power comes with expectations of punishment for past defections. Second, selective post-control reprisals bring higher cooperation than indiscriminate ones regardless of how effective is the armed group at identifying enemy informants and third, communities that have a highly centralized process of decision making cooperate only with one armed group reducing their exposition to violence.
Presented at the 2012 meeting of the Midwest Political Science Association.
We propose a MAP-EM estimator that recovers (stochastic) national electoral thresholds and disproportionality from observed seats/votes allocation data. We apply the procedure to 101 electoral systems used in 415 elections to the lower house across 30 European countries since WWII. We find that over half of these systems exhibit a statistically significant positive electoral national threshold of representation with the median threshold among these systems being 2.75%. In 46 out of the 101 systems we cannot reject the hypothesis that the allocation of seats for parties above the threshold is proportional to vote shares. The estimates produce a directly interpretable summary of the electoral system and, if appropriately purged from estimation contamination, can serve as independent variables to evaluate the consequences of electoral institutions.
Presented at the 2011 meeting of the American Political Science Association and the 2010 meeting of the Society for Political Methodology.
The popularity of propensity score matching has given rise to a robust, sometimes informal, debate concerning the number of pre-treatment variables that should be included in the propensity score. The standard practice when estimating a treatment effect is to include all available pre-treatment variables, and we demonstrate that this approach is not always optimal when the goal is bias reduction. We characterize the conditions under which including an additional relevant variable in the propensity score increases the bias on the effect of interest across a variety of different implementations of the propensity score methodology. Moreover, we find that balance tests and sensitivity analysis provide limited protection against overadjustment. pdf.