Time-series models of the demand for alcoholic beverages have been criticized for use of annual data; omitted variables; mis-measurement of advertising; simultaneous equations bias; and inadequate attention to nonstationarity and dynamics. This paper reappraises the relationship between alcohol advertising, price, and consumption in a manner which speaks to these issues. Using quarterly data from 1970:1 to 1990:4 on three beverages (beer, wine, and distilled spirits), we find evidence of cointegration between beverage consumption, prices, advertising, and real income. Elasticities obtained from the estimated cointegrating vectors indicate that long-run beverage demands are both price and income inelastic. Moreover, after correcting for each of the problems described above, advertising has virtually no influence on the steady-state level of alcoholic beverage consumption.