We consider the behavior of an incumbent that can deploy local public goods and private goods to buy votes, and is unable to verify vote choice but capable of monitoring voter turnout, a common scenario in secret-ballot polities. As advanced by recent literature, the ability to monitor turnout rather than vote choice implies that politicians should use targetable private goods to mobilize voters. However, politicians also deploy non-excludable local public goods, which have low mobilization potential because of free-rider incentives. We argue that vote-buying politicians reserve local public goods for loyal constituencies (where they enjoy support from most voters that bother to turn out) and provide private goods in other constituencies where they gain from motivating less committed supporters to turn out. We test aggregate-level implications of our theory on Venezuela’s social programs.