What Do Campaign Contributions Buy? Deciphering Causal Effects of Money and Votes

Abstract

In the 1987-1988 election cycle, political action committees (PACs) spent $176.6 million on candidates who competed for a seat in the House of Representatives. What do PACs buy when contributing this money? If campaign are rational allocations of a contributor's income, campaign expenditures could increase both the probability that the contributor's preferred candidate is elected, and the likelihood that, if elected, the candidate will vote in the interest of the contributor. Most studies do not find that incumbents' expenditures have a significant positive effect on the number of votes incumbents receive [11; 14; 15; 17; 29]. The evidence that influence congressional votes is mixed. Most studies find weak or no support for this hypothesis. For example, Kau, Rubin and Keenan [17] find that from businesses mattered in only two of the eight votes they analyzed. Others argue that campaign expenditures play no important role in the voting decision of congressmen [5; 9; 12; 31]. The lack of finding of significant effects of money on votes has been explained with the need for better models to explain campaign contributions [5]. However, the theoretical developments of, for example, Ben-Zion and Eyton [4], Bental and Ben-Zion [3], and Baron [2], have not been incorporated in the empirical analysis of the effect of campaign on the voting behavior of legislators. Empirical models that test this relationship must allow for the endogeneity of campaign contributions, for the dichotomous nature of the dependent vote variable and the nonnegativity constraint of campaign contributions. Most empirical work in this area utilize models that do not incorporate these restrictions. In this study, I plan to incorporate these aspects of the data generating process into a more detailed model of the effects of campaign on congressional votes. Of all the reported empirical studies on this issue, only Chappell [5] uses a model which can properly assess the impact of money on votes. Chappell finds that campaign do not affect votes in Congress when his more detailed econometric technique is used. The best test of the theory results if votes have a clear economic payoff to contributors, benefits of the votes are concentrated and costs are