In the presence of competing interest groups, this paper examines how the form of votebuying contracts affects policy outcomes. We study contracts contingent upon individual votes, policy outcomes, and/or vote shares. Voters either care about their individual votes, or about the policy outcome. We find that vote buying is cheaper when what can be contracted upon coincides with what voters care about. Vote buying becomes extremely costly, or even impossible, when there is no such coincidence. Finally, vote buying is extremely cheap, or even free, when contracts can be contingent upon both individual votes and vote shares.