This study analyzes how people make two sets of choices that affect drunk driving: the decision to drink alcoholic beverages, and the decision to drive after having consumed alcoholic beverages. The study consists of three parts, all interconnected. The first part provides a broad study of the demand for alcoholic beverages, including beer, wine and distilled spirits. This study relies on information contained in the National Health Interview Survey (NHIS), coupled with alcoholic beverage price data contained in a survey from the Association of Chambers of Commerce Research Affiliates (ACCRA), and improves on previous similar studies by (a) using more complete price data, (b) extending the analysis to the entire population of drinkers, rather than just to young drinkers, and (c) employing more general and power statistical techniques. This work builds on a model of demand that allows both for addiction (as previous studies have done) and incomplete information about the risks of alcohol (not previously incorporated into such demand models). The second part of the study employs information from roadside surveys to analyze the probability distribution of blood alcohol levels of drivers as a function of socio-economic characteristics of the driver and prevailing laws pertaining to drunk driving. This study relies on two nationwide roadside breath testing surveys conducted in 1973 and 1986. The estimated probabilities are converted, using Bayes Theorem and separate estimates of the unconditional probabilities of driving and drinking, into probabilities of driving given specified levels of drinking. These probabilities allow the analysis of the decision to drive after drinking. The final part of the study combines the two previous components into a simulation analysis, allowing for interactions of regulations and laws affecting both the demand for alcoholic beverages and the decision to drive after drinking.
