cross-sectional model, relating an indicator of current consumption to the current price, which does not take the effects of addiction into account.
Heterogeneity of consumers. Cocaine consumers include casual and heavy users. These two groups face similar prices (or distributions of prices) but have very different consumption patterns. The levels of the demand functions of casual and heavy users are obviously different. It is possible that the slopes (the changes in consumption due to a unit change in price) are also different. For example, casual users may be more responsive to changes in prices than are heavy users. If so, the high elasticities of participation found in recent studies of demand may mainly reflect responses of casual users to price fluctuations. In addition, the finding that frequency of use (conditional on participation) is less responsive to price than the number of users may be strongly influenced by the behavior of heavy (high frequency) users who are relatively insensitive to price changes. No demand model that the committee has seen allows for the possibility that casual and heavy users have different price sensitivities.
Cross-elasticities. Some illegal drugs may be substitutes or complements for others. If so, the demand for one drug may depend on the price of another. Most studies have ignored such cross-elasticity effects. Two exceptions are models of demand for heroin and cocaine by Caulkins (1995) and a model of marijuana participation by DeSimone (1998). Moreover, possible complementarities are the focus of research on what has been called the gateway effect (see Chapter 7 for further discussion).
The dynamics of drug use. No existing empirical model of demand for drugs describes the process by which individuals initiate and make transitions among different levels of drug use (for example, from nonuse to casual use, from casual use to nonuse or heavy use). Everingham and Rydell (1994), Rydell and Everingham (1994) and Everingham et al. (1995) propose a conceptual framework for modeling such transitions. However, the data that are required for empirical study of drug use dynamics and their dependence on prices and other costs of drug use are not available to researchers. Implementation of such a study would require a longitudinal dataset that describes drug use by individuals over time. The Monitoring the Future survey gathers longitudinal data on participation and frequency of use by youth, but it rarely makes the data available to researchers.
Heterogeneity of drugs. Cocaine is sold in several chemically distinct forms (mainly cocaine base and powder cocaine). These forms have different prices, and consumers consider them to be different products. It is therefore likely that different forms of cocaine have different demand functions and price elasticities. The committee is aware of no study that has estimated separate demand functions or price elasticities for different forms of cocaine.