Response to NRC Assessment of RAND's Controlling Cocaine Study

Abstract

A 1994 RAND study, Controlling Cocaine: Supply vs. Demand Programs, drew widely cited conclusions regarding the relative cost-effectiveness of spending additional drug control moneys on treatment and various modes of enforcement. A National Research Council committee last year issued a critique of that report concluding that it was not a good basis for policymaking. Modeling is an inexact science, and there is plenty of room for experts to disagree on methods and conclusions. We feel, however, that the NRC's critique warrants a reply, for two reasons. First, it appears to rest on incomplete information regarding the RAND model. Our differences with the assessment are thus not simply a matter of varying judgment or opinion. Second, the critique was issued by a distinguished panel at the request of the Office of National Drug Control Policy. Barring objection, we fear that many in the drug policy community may tend to accept the conclusions of the panel as the "last word" on this topic. This document presents a detailed point-by-point response to the NRC panel's critique. It was prepared by Jonathan Caulkins with substantial input from James Chiesa and Susan Everingham. Reviewers were David Boyum and Peter Reuter, to whom we are all grateful for numerous comments that resulted in improvements to the text. We summarize the NRC panel's criticisms (in italics) and our responses here. • The RAND study assumes that the unit price of supplying cocaine decreases with the total quantity of cocaine supplied, whereas for conventional goods the usual assumption is of price increasing with quantity supplied. This could influence the study's conclusions about the effectiveness of enforcement policies. Though we stand behind the price-quantity relationship that we assume, we did reprogram the model to include the panel's alternative assumption and the results are substantially the same. • The RAND model uses cocaine seizures as the primary measure of supply control activity, when they alone may not accurately reflect that activity. Actually, the model considers asset seizure, arrests, imprisonment, and incarceration of cocaine-using sellers, although these are assumed to be proportional to product seizures. Rewriting the equations in terms of one of the other supply control effects would have no effect on the results. The approach taken here is consistent with our desire to test the effects of expanding average practice within each category of enforcement strategies, so that the approach would be parallel with our analysis of treatment. • The RAND study ignores the nonmonetary costs of dealer imprisonment (e.g., stigma). The model assumes that the price a dealer charges covers all costs and risks, monetary and nonmonetary (or the dealer would find some other line of work).

Rethinking America's Illegal Drug Policy

“…Rydell's & Everingham's initial study was criticized for underestimating the decrease in cocaine use stemming from increases in cocaine prices due to source-country control, interdiction, and domestic enforcement. Repeating their study of policy effectiveness in 2000 assuming a more elastic demand for cocaine, Caulkins, Chiesa, and Everingham (2000) determined that treatment has a four-to-one advantage over domestic enforcement in reducing the costs of crime and productivity losses.…”

Drug Policy

“…The right column of Table 1 lists explanatory constructs relied on most heavily in recent theoretical analyses of American drug policy (e.g. Behrens et al 2000;Caulkins et al 2000;…”

“…It is surely better to categorize users into "addicts" vs. non-addicted users, rather than mindlessly (and moralistically) lumping heavy users together with extreme casual light users (see Caulkins 1997). But we should be wary of reifying an extreme corner of a continuous, multidimensional space constituted by the dimensions of frequency of use, quantity consumed per use, and harmfulness of conduct while intoxicated.…”

Illicit drug markets and economic irregularities

“…Although the idea that enforcement swamping might generate industry-wide external economies of scale has been incorporated in various models [22,38], it remain controversial [13], even though it turns out not to be a major driver of the best known conclusions of those models [39]. Presumably at a minimum it could be a factor that tends to flatten any upward slope in the supply curve, even if it does not make the slope actually negative.…”