This paper reexamines the effect of the business cycle on alcohol consumption using U.S. state-level analysis introduced by Rhum [Ruhm, C.J., 1995. Economic conditions and alcohol problems. Journal of Health Economics 14, 583-603]. Using an extended panel, this analysis finds that Ruhm's estimates are biased and inconsistent due to the problem of non-stationarity in alcohol consumption and state economic condition time series. Corrected estimation using logarithmic first differences confirms Ruhm's original finding of pro-cyclical alcohol consumption, but these results, unlike Ruhm's, are robust to sample period.