"Regression results from a 30-year panel of the state-level data indicate that changes in alcohol-excise taxes cause a reduction in drinking and lower all-cause mortality in the short run. But those results do not fully capture the long-term mortality effects of a permanent change in drinking levels. In particular, since moderate drinking has a protective effect against heart disease in middle age, it is possible that a reduction in per capita drinking will result in some people drinking "too little" and dying sooner than they otherwise would. To explore that possibility, we simulate the effect of a one percent reduction in drinking on all-cause mortality for the age group 35-69, using several alternative assumptions about how the reduction is distributed across this population. We find that the long-term mortality effect of a one percent reduction in drinking is essentially nil"--National Bureau of Economic Research web site.