Working Paper No. 04-05:
The Health Effects Of Alcohol: Do Controls For Demographics And Other Risky Habits Affect The Conclusions?
Date Posted: 2004
Abstract (below) | Full text (most recent) on SSRN
This paper uses data from the Health and Retirement Survey to measure the effects of alcohol on the incidence of morbidity and death. The study is able to reproduce the implied benefits of engaging in moderate levels of alcohol consumption, even after controlling for a large number of independent variables not usually available in health data sets. In fact, the controls work in the direction of supporting the benefits of engaging in even higher dose levels than conventionally recommended. It turns out, however, that smokers and quitters enjoy most of the benefits of unusually high alcohol consumption. Non smokers evince modest benefits that are completely captured at very low dose levels. In general, the results suggest that studies of alcohol intake on health need to pay more attention to the characteristics of users. It may be that alcohol is especially beneficial for populations that are deficient in their health for other reasons like smoking or poor eating, whereas populations who follow good diets and do not smoke benefit very little from alcohol use.