A central question in public opinion is how voters respond to distributive policies that benefit them. The common expectation is that voters reward incumbents for personal benefit, following a pocketbook voting logic. Yet, the pocketbook explanation has received little direct empirical scrutiny. In this paper, I challenge the pocketbook account and argue that existing studies conflate personal benefits with perceived benefits to voters’ in-groups. I theorize a group-based mechanism in which voters respond to distributive policies based on how they believe those policies affect certain salient in-groups. I test the argument in two empirical studies. First, using survey data on COVID-era stimulus checks in Denmark and the United States, I show that check recipients became more likely to believe their racial or geographical in-groups also benefited. These findings reveal that distributive perceptions are endogenous to personal benefit, casting doubt on the common attribution of policy effects to pocketbook voting. Second, to isolate the causal role of group-based perceptions, I field three pre-registered experiments in the two countries, randomly varying features of hypothetical cash transfer policies. Across experiments, I find that voters’ political support depends at least as much on perceived in-group benefit as on personal gain. Importantly, these effects are highly group-dependent, emerging only for groups with strong political identities. Together, the findings show that group-based responses, not just pocketbook concerns, shape how voters react to policies that benefit them, helping explain the wide variation in policy effects across contexts.