Abstract :
We examine the relationships among alcohol prices, consumption, and traffic fatalities using data across U.S. states from 1982 to 2000. Some previces studies have found large, negative associations between alcohol taxes and fatalities. However, commonly used price data suggest little or no connection between alcohol prices and fatalities. These apparently conflicting findings may result from measurement error and/or endogeneity in the price data, which biases ordinary least squares estimators toward a finding of no price effects. Using alcohol taxes as instrumental variables, fatalities are found to be negatively related to prices. In addition, alcohol consumption is strongly positively related to fatalities. However, biases may still remain, because taxes are not entirely suitable as instruments.
