The local electoral impacts of conditional cash transfers Evidence from a field experiment

I develop and test two competing models assessing the impacts of targeted government transferson a local incumbent's electoral performance. I take advantage of the randomized roll-out of alarge-scale Conditional Cash Transfer program in the Philippines, which offers an ideal setting totest the models. Although the program was usually implemented in all villages in a municipality,a subset of beneficiary municipalities were randomly selected to receive the program in a randomlyselected subset of villages. I find that, in a competitive political environment, incumbentvote share is 26 percentage points higher in municipalities in which the program was implementedin all villages than in municipalities in which the program was implemented in half of the villages.The program had no impact in municipalities with low levels of political competition. Further,within municipalities, there is evidence consistent with the argument that incumbents compensatedhouseholds in control villages by redistributing their own budget there. Results suggest thatanti-poverty programs might have nefarious long-term consequences by preventing replacementsof local incumbents.