Much as I’d like to rationalize alcohol consumption, it seems the studies showing moderate drinking leads to better health have the usual problem according to the New York Times:
No study….. has ever proved a causal relationship between moderate drinking and lower risk of death — only that the two often go together. It may be that moderate drinking is just something healthy people tend to do, not something that makes people healthy.
Health economics might provide a treasure trove for economists well-versed in the techniques of “instrumental variables” typically used to determine causation.
Cheap Talk 
2 comments
Comments feed for this article
June 16, 2009 at 4:56 pm
michaelwebster
I doubt that they have even established proper covariation in most of these cases, just thinking about the various 2 x 2 tables.
June 27, 2009 at 9:21 pm
Thomas
You might try reading some of the original studies before dismissing them…
Economists have tried on a number of occasions to evangelize for instrumental variable methods in epidemiology, and have largely discovered that good instruments are extremely hard to find. There has been some use in studying randomized trials with poor compliance, and under the name of ‘Mendelian Randomization’ the method has been used with mixed success in genetic epidemiology. The (more) cynical members of the biostatistical community sometimes wonder whether instrumental variables actually work better in economics or whether it’s just harder to tell that they have failed.