Economic analyses of multiple addictions for men and women
Goodman, A.C.
Journal of Mental Health Policy and Economics 12(3): 139-155
2009
ISSN/ISBN: 1091-4358 PMID: 19996476 Document Number: 630996
This study seeks to address analytical issues regarding the joint usage of alcohol, tobacco, and drugs, focusing on incomes, taxes, and gender-related differences. Many studies analyze a single addictive substance, with the maintained assumption (often due to data inadequacies) that the use of other addictive substances does not matter. Using a database that is uniquely suited to the task, this study examines economic determinants of addiction probabilities and decomposes the differences between men and women into risk factors and probabilities. The study uses the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) database. The NESARC, representing the entire non-institutionalized U.S. population age 18 and over, is the primary source for information and data on: (i) alcohol and drug use; (ii) alcohol and drug abuse and dependence; and (iii) associated psychiatric and other medical comorbidities. The study then proposes a multinomial logit modeling strategy that addresses endogeneity of smoking, drinking, and drug use. Parameter estimates then predict absolute and marginal probabilities and look at gender and age related differences. The study also develops and demonstrates a new decomposition for analyzing the differences between men's and women's uses of addictive substances. Women, Blacks, and Hispanics are less likely to engage in addictive behaviors. Increased cigarette and beer taxes negatively affect probabilities of smoking and drinking. Increasing both cigarette and beer taxes is related both to more abstinence (none of the three types of substances), and to more use of drugs (which are untaxed). The measured impacts of current income and current taxes on addictive goods are strong even though addictive decisions are almost certainly longer term decisions, reflecting both current and past prices. However, the impacts of current incomes and taxes in the multinomial logit formulations are highly significant and the results are plausible. To the extent that taxes can reduce harmful addictive behaviors, the utilization and cost of health care attributable to addiction may be reduced. Higher taxes have strong potential negative impacts on addictive behaviors. The effects differ, however, by gender, race, and age, and ethnicity. The analysis could be extended to two part models, in which quantities and/or expenditures on alcohol, tobacco, or drugs may be examined, conditional on the individuals' specific categories of addictive substance used. With panel data, decisions on starting and/or stopping drinking, smoking, or ingesting drugs may also be considered.