Estimating the Price Elasticities of Demand for Illicit Drugs: Real-World vs. Laboratory Data

This study estimates the price elasticities of demand for heroin, cocaine, marijuana, and alcohol, using both real-world and laboratory data.  Data were collected in two waves from the same group of 120 heroin polydrug users, with each wave lasting two weeks and waves separated by 6 months.  During each wave, subjects recorded their daily drug purchase information via an interactive voice response (IVR) system.  At the end of each wave, the same subjects participated in a laboratory experiment that was designed specifically to mimic the scenarios reported on the IVR calls, thereby permitting a direct comparison of the real-world and laboratory methods of estimating elasticities for illict drugs.  Preliminary analyses, based on approximately 1,600 IVR calls and matched laboratory scenarios, show a strong correspondence between purchases made in the real-world and choices made in the laboratory, especially for the primary drug of abuse (heroin).  For example, the correlation coefficient for the quantities of drugs purchased in the real-world vs. the quantities chosen in the laboratory are .85, .65, .40, and .39 for heroin, cocaine, marijuana, and beer, respectively.  The two methods also yield similar elasticity estimates for the primary drug of abuse (heroin), but not for the secondary drugs of abuse (cocaine, marijuana, beer).