When it comes to the study of the effect of price on alcohol consumption, a wealth of information has been discovered. The impact of price has been widely discussed and fairly precisely estimated, paying attention to the differential impacts on light, moderate, and heavy drinkers. However, when it comes to studies on the impact of physical (geographic) access to alcohol, comparatively little is known. This application proposes to develop a research program that will move the study of access forward, probing the theoretical and practical details of what we mean by """"""""access"""""""". While many studies have investigated the aggregate effects of alcohol outlet densities on average drinking behavior, there are still many important issues to be resolved. First, how should geographic access be theoretically defined? Second, given the real-world limitations on data, how should access be empirically defined? Third, research needs to begin to address the ecological inference problems of previous studies on access to alcohol. In particular, we know that on average, consumers in a state will decrease alcohol consumption when access is limited. However, we do not know how individual consumers are affected. Consumers could simply make fewer purchase trips and """"""""stock up"""""""", could substitute more readily available forms of alcohol, or could consume less alcohol. Also of importance are the possible differing impacts among various types of drinkers. In particular, does the average consumption decrease because all consumers drink less? Or, do some consumers stop drinking altogether, and others fail to change their drinking behavior? This research will compare well-known measures of access with innovative new measures, develop theoretical economic descriptions of how different types of consumers might respond to changes in access, and analyze a unique data set containing liquor sales and store locations in North Carolina and Virginia. Specifically, census data on racial characteristics, income, commuting patterns, and population density will be used to explain variations in per capita alcohol purchases, death rates and hospitalizations rates due to chronic liver disease, DUI arrests, and arrests for drunkenness.
