Alcohol abuse is widely blamed for the very high rate of male mortality in Russia. I specify and estimate a simple structural model of the demand for alcohol that incorporates two key features of the Russian context. First, alcohol use – particularly incidents of heavy drinking – often involves friends and (male) family members. Second, there is strong habit persistence in alcohol use: depending on the degree of forward-looking behavior by consumers, responses to a tax policy will depend on beliefs about the future path of prices. I estimate the model using panel data from the Russian Longitudinal Monitoring Survey (RLMS), and two alternative sources of variation in alcohol prices: a 2011 change that shifted the trend in the rate of growth of the excise tax for alcohol, and regional variation in alcohol regulations over the 1995-2014 period. To obtain direct information on peer use of alcohol, I exploit the clustered design of the RLMS, which enables me to find close neighbors for nearly all sample members. The estimation results confirm that both peer influence and habit persistence are critical determinants of the longer-run response alcohol demand to price changes. One third of the predicted 30% reduction in the rate of heavy drinking caused by a 50% permanent increase in vodka prices, for example, is attributed to the social multiplier effect that emerges as groups of friends jointly reduce their consumption. Finally, I use the RLMS data to relate patterns of heavy drinking to mortality. The estimates imply that permanent increases in alcohol prices would yield significant
reductions in male mortality.