Price and enforcement effects on cocaine and marijuana demand

Economic Inquiry, Jan, 2003 by Jeff Desimone, Matthew C. Farrelly

This article estimates equations for past year cocaine and marijuana use among adult and juvenile respondents of the 1990-97 National Household Surveys on Drug Abuse. Unlike most previous studies, we control for the monetary price of marijuana, probabilities of arrest for marijuana and cocaine possession, and state fixed effects. Results indicate that cocaine prices are inversely related to adult cocaine and marijuana demand but are unrelated to juvenile drug demand, marijuana price effects are always statistically insignificant, estimated price effects are inflated when state effects are omitted, and increases in each arrest probability diminish both types of drug use. (JEL K42, 118, D12)

I. INTRODUCTION

The responsiveness of cocaine and marijuana demand to changes in their prices is a key determinant of the effectiveness of illegal drug enforcement policy. By harassing sellers and seizing drugs, enforcement attempts to reduce the consumption of illegal drugs by restricting their supply and thereby raising their prices. Even if enforcement is able to increase drug prices, its success in reducing illegal drug use depends on the elasticity of drug demand with respect to drug prices. Reciprocally, unless this elasticity is close to zero, legalization of cocaine and marijuana would likely increase their consumption substantially by drastically reducing their prices.

A complementary goal of enforcing cocaine and marijuana possession violations is to reduce their demand at prevailing prices. This occurs through both incarceration of drug users who will no longer be able to purchase drugs and deterrence of drug consumption by potential users. Price and enforcement effects may be dissimilar if consumers respond differently to changes in their budget constraints than to changes in expected punishment. In particular, the relative magnitudes of the responses in drug demand to changes in possession arrest probabilities and prices is an important determinant of how enforcement resources can most efficiently be allocated between buyers and sellers. But in spite of this policy relevance, there is little direct evidence on the relationship between arrest probabilities for cocaine and marijuana possession and demand for these drugs.

Meanwhile, the relationship between the consumption of cocaine and marijuana is both theoretically and empirically uncertain. In theory, cocaine and marijuana act as substitutes in the production of intoxication but also can provide complementary intoxicating effects. Empirically, this relationship determines whether policies designed to reduce demand for one drug have effects on the other that reinforce or counteract the impacts of policies designed specifically for that other drug. For instance, marijuana possession arrests more than doubled nationally between 1990 and 1997, both in number and as a fraction of total arrests. This might have reinforced any effect of cocaine possession enforcement on cocaine use if the two drugs are complements but had an unintended counteractive effect if they are substitutes.

This study provides evidence on the impacts of cocaine and marijuana prices and possession violation enforcement on the demand for these drugs. We analyze data on past year cocaine and marijuana use among 12- to 39-year-old respondents to the annual 1990-97 National Household Surveys on Drug Abuse (NHSDA). Along with various individual characteristics, the set of explanatory variables includes regional prices of cocaine and marijuana, state-level measures of the probability of arrest for cocaine and marijuana possession, and fixed effects for states and years. Our goals are to estimate the size of the response in the demand for cocaine and marijuana to changes in their prices, to do the same with respect to changes in possession violation enforcement intensity, and to examine whether unmeasured state characteristics can potentially bias estimated price and enforcement effects.

The analysis is novel in several ways. Most important, it is the first study of cocaine demand to control for state fixed effects in nationally representative data. Previous studies impute cocaine prices at the state level and use both cross-state and temporal variation in prices to identify price effects. But it is possible that a substantial component of cross-state price variation is explained by unobservable, time-invariant state-level factors that also explain cross-state variation in illegal drug use. For instance, states with more permissive attitudes regarding drug use are likely to have both lower cocaine prices and higher drug prevalence rates than antidrug states. Cocaine price elasticities estimated from analyses that ignore these fixed state effects might overstate the impact of an exogenous price change on the change in drug use within an average state, which is the true elasticity of interest.

The inclusion of state fixed effects necessitates constructing our cocaine prices differently than previous studies. Motivated by evidence from. Rhodes et al. (1994) and Caulkins (1995) that region and population size are the crucial geographic determinants of cocaine prices, we calculate cocaine prices that vary by census division and metropolitan area size. This allows for the temporal price variation necessary to simultaneously identify cocaine price effects and fixed state effects. In addition, our analysis is the first to examine the effect of marijuana prices on cocaine demand and to explicitly estimate the effect of cocaine possession arrest rates on cocaine and marijuana demand and of marijuana possession arrest rates on cocaine demand.