A recent article in the Journal of Wine Economics by Duquesne University Economics Department associate professor Pavel A. Yakovlev and graduate student Walter P. Guessford offers research findings so obviously pleasing to conservatives that you might wonder whether they were perpetrating a hoax. What their findings show is a positive relationship between measures of the ideological liberalism of a state and measures of how much alcohol was consumed in a state for the years between 1967 and 2010. Note that the authors are careful not to claim directly that their findings show that liberals drink more than conservatives. However they do reference a theory in their introduction that points to just such a conclusion. Moral hazards theory is deployed to suggest that dependency on government provision of health care makes people irresponsible about behaviors that affect their health. Mind you, if that is what explains these findings, then the real news here is that Americans in more liberal states began drinking to celebrate Obamacare four decades before it arrived!
Logic be damned, conservative reporters, editors and bloggers loved the idea that liberals are boozers. Jennifer Harper at the Washington Post led with, “New academic research correlating a link between unhealthy behaviors and political ideology does not bode well for liberals.” John Aziz at The Week wrote that the, “findings are consistent with other recent studies in other parts of the world showing that people with socialist views tend to drink more.” Liberals and socialists are basically indistinguishable don’t you know?
Beyond the time order problem with the moral hazards argument, this research suffers from a methodological flaw that anyone with a bachelor’s degree in the social sciences would have recognized: the ecological fallacy. The rule is that evidence of behavior at one level of analysis should not be used to reach conclusions about behavior at another level of analysis. So the statistics analyzed by Yakovlev and Guessford could only tell them about aggregate behavior at the state level and not about the behavior of state residents at the individual level. What that means is that the larger amounts of alcohol consumed in more liberal states were actually consumed by minorities of unhappy conservatives. Picture in your mind all those deeply depressed Republicans on the West Coast and in Colorado, Iowa and New England trying to drown their partisan sorrows in rivers of Coors Lite, MD 20/20 and Old Crow. (OK, I’m just guessing.) Alternatively, perhaps both liberals and conservatives in more liberal states may drink more because they are celebrating the superior quality of life in states where the problems of poverty, obesity, smoking, teen pregnancy etc, are less severe. We don’t know…and neither do Yakovlev and Guessford. Sláinte!