The cost of living varies as much across regions as it does across time, but researchers have only begun to acknowledge the importance of controlling for regional differences in the cost of living when conducting cross-sectional analyses. We demonstrate the importance of considering regional cost-of-living differences by using empirical models of demand for state lotteries. Previous research on state lotteries has shown that the nominal-income elasticity of demand for lottery tickets is less than 1, suggesting expenditure on lottery tickets is regressive. We reestimate traditional models of lottery demand using a sample of metropolitan statistical areas, but we also control for cost-of-living differences across these areas. We find that, in accordance with our conceptual framework, estimated income elasticities are larger when we control for local cost-of-living variation – that is, the regressivity of state lotteries is overstated when local cost-of-living variation is omitted from empirical models of lottery demand. Our results are robust to several measures of the cost of living, including housing price indexes.