Making Policies Matter: Voter Responses to Campaign Promises

Can campaign promises change voter behavior, even where clientelism and vote buying are pervasive? We elicit multidimensional campaign promises from political candidates in consecutive mayoral elections in the Philippines. Voters who are randomly informed about these promises rationally update their beliefs about candidates, along both policy and valence dimensions. Those who receive information about current promises are more likely to vote for candidates with policy promises closest to their own preferences. Those informed about current and past campaign promises reward incumbents who fulfilled their past promises; they perceive them to be more honest and competent. However, voters with clientelist ties to candidates respond weakly to campaign promises. A structural model allows us to disentangle information effects on beliefs and preferences by comparing actual incumbent vote shares with shares in counterfactual elections: both effects are substantial. Even in a clientelist democracy, counterfactual incumbent vote shares deviate more from actual shares when policy and valence play no role in campaigning than when vote-buying plays no role. Finally, a cost benefit analysis reveals that vote-buying is nevertheless more effective than information campaigns, explaining why candidates do not use them.