Strategic campaigning with vote and turnout buying

We model a patron-client relationship where the incumbent may either persuade or mobilize voters; that is, he can pay citizens to vote for him -vote buying- or he can pay citizens to show up to vote-turnout buying- respectively. Furthermore, we are interested in the optimal budget allocation across groups of citizens, thus we focus on a single member, majoritarian election in a multidistrict environment. Our findings enrich the distributive politics literature that frames the problem into swing vs. core states, by showing that (1) the distributive/clientelist game favors districts where more voters prefer the opposition party, resembling Dixit and Londregan (1996); (2) in every district, citizens who support the incumbent party (weakly or strongly) are the first target of its distributive/clientelist eff orts; and (3) within those voters who receive transfers, mobilization occurs mainly across the weak supporters and weak opposers, while persuasion occurs across the strong opposers to the incumbent party, simultaneously. Using data from argentina we show that politicians behave consistently with the predictions, and we can estimate the cost of buying votes and turnout. Moreover, we show that gender and monitoring technologies are important in understanding not only vote and turnout buying behavior, but also targeting strategies. We estimate the number of votes bought during the 2003 presidential race, its average price, and their effect.