Does Drinking Really Decrease in Bad Times?
This paper investigates the relationship between macroeconomic conditions and drinking using individual-level data from 1987 to 1999 interview years of the -behavioral risk factor surveillance system‖ (BRFSS). We confirm the procyclical variation in overall drinking identified in previous research using aggregate sales data and show that this largely results from changes in consumption by existing drinkers, rather than movements into or out of drinking. Moreover, the decrease occurring during
... occurring during bad economic times is concentrated among heavy consumers, with light drinking actually rising. We also find no evidence that the decline in overall alcohol use masks a rise for persons becoming unemployed during contractions. These results suggest that any stress-induced increases in drinking during bad economic times are more than offset by declines resulting from changes in economic factors such as lower incomes. (2000) provides a useful recent literature review on price elasticities and alcohol taxation. The effects of availability constraints and minimum drinking ages are discussed in Gruenewald (1993) and Wagenaar (1993) . Cook and Moore (2000) provide an excellent overview of economic research related to alcohol. Article: US Department of Health and Human Services 2. For example, a 1 standard deviation (2.12 percentage point) increase in the state unemployment rate is predicted to reduce drinking by 1.3 % and traffic fatalities by almost 7%. The consumption of distilled spirits is more sensitive to macroeconomic conditions than purchases of beer or wine. These results are consistent with earlier research findings by O'Neill (1984) and Evans and Graham (1988) indicating that vehicle fatalities and single vehicle night-time crashes (which frequently involve drunk-driving) are procyclical and those of Wagenaar and Streff (1989) suggesting that alcohol consumption increases in good times. 3. Freeman finds that the failure to correct for non-stationarity yields parameter estimates that are sensitive to the choice of time periods; however, this may be an artifact of the data he uses. Whereas Ruhm transforms beer, wine, and distilled spirits consumption into ethanol-equivalents using constant conversion rates, Freeman use information from the -alcohol epidemiologic data system‖ (AEDS) where the conversion factors vary over time-wine is assumed contain 16. 0% alcohol prior to 1972,14.5% from 1972 to 1976, and 12.9% after 1976, and distilled spirits 45.0%, 43.0%, and 41.1% alcohol in the three periods. These sharp discontinuities could render Freeman's data non-stationary. 14. The year variables eliminate 60.3% of the variation in unemployment rates remaining after including month dummy variables, state fixed-effects, and state-specific time trends. By contrast, the addition of state time-trends absorbed just 27% of the variation remaining after controlling for month and state fixed-effects. 15. This problem is likely to be less severe (but not completely eliminated) when considering state unemployment rates, to the extent that state economies fluctuate independently. 19. Similar evidence of larger short-run than long-run changes in alcohol consumption has been obtained by previous researchers examining the effects of treatment programs (Humphreys et al., 1997) , divorce (Hartford et al., 1994) , DUI-legislation (Ross, 1984), and the privatization of alcohol sales (Mulford et al., 1992) .