Data on U.S. state-level beer shipments from 1970 to 2007 provided by The Beer Institute are used to estimate pooled time-series models of annual consumption regressed on economic and demographic variables, using the common correlated effects (CCE) estimators to control for unobserved common effects and to allow for heterogeneous responses across units. Beer is found to be a procyclical good, varying negatively with the state unemployment rate. Previous findings for the negative effect of excise taxes on consumption are supported, though the estimated elasticities are smaller than those reported in earlier research. Demographics have a significant and material effect on consumption, with larger shares of young adults in the population implying greater consumption of beer per capita.