Executive Summary

Since 1991, the Office of National Drug Control Policy has published a biennial report on expenditures by Americans on illegal drugs and on legal drugs used illegally. This version of that biennial report provides estimates of cocaine, heroin and marijuana consumption from 1988 through 1998 and projects estimates for 1999 through 2000. For the first time, it provides comparable estimates for methamphetamine. This version improves and updates estimates of the supply of cocaine to the United States, and for the first time, provides estimates of the supply of heroin to American consumers. Finally, this version reports improved and updated estimates of trends in the domestic price of cocaine, heroin, methamphetamine and marijuana.

Using a consumption-based approach, we investigated the dollar expenditures by Americans on illicit drugs. We estimated that:

• In 1998, Americans spent $66 billion on these drugs (Table 1):¹

  • $39 billion on cocaine
  • $12 billion on heroin
  • $2.2 billion on methamphetamine
  • $11 billion on marijuana
  • $2.3 billion on other illegal drugs

• Between 1988 and 1998, expenditures on cocaine appear to have fallen. This trend results partly from a decrease in the number of users, but mostly from a decrease in cocaine's street price.

• Heroin expenditures fell from 1988 to the middle of the 1990s. Heroin expenditures appear to have increased since then.

• Trends in methamphetamine purchases are imprecise because of significant measurement problems. While expenditures may have fallen due to changes in the consumer price index, consumption levels have remained about the same over the last decade.

• Between 1989 and 1998, expenditure on marijuana increased slightly (as marijuana prices increased) then decreased slightly (as marijuana prices fell).

• Between 1989 and 1998, expenditures on other illicit drugs, and on legal drugs used illicitly, remained fairly constant.

Figures developed in estimating the retail sales value of illicit drugs consumed in the United States were compared to estimates of the amounts supplied to the domestic market. To investigate the reasonableness of our approximations of cocaine consumption in the U.S., we compared our consumption estimates with two estimates of cocaine supplied to the domestic market. The first comparison was an extrapolation of coca cultivation estimates (calculated by the Sequential Transition and Reduction (STAR) Model). The second comparison was an extrapolation of cocaine departing South America developed by the U.S. intelligence community, based on quantifying the density and loading of cocaine traffickers departing South America. See Table 4 for the results.

• The cultivation-based consumption estimates are high relative to our consumption estimates. Also, they decrease from 411–559 metric tons (1996) to 176-324 metric tons (1999), and that trend is not reflected in other measures of cocaine use.

• After 1996, the event-based consumption estimates are smaller than the Abt consumption estimates: 154 metric tons in 1997, 212 in 1998, and 191 metric tons in 1999. Moreover, their variability is not reflected in other data about cocaine use.

• Roughly 12 to 13 metric tons of pure heroin entered the United States between 1995 and 1998. Because heroin is roughly 80 percent pure when imported into the U.S., the 12 to 13 pure tons represents 15 to 16 bulk tons.

• It was not practical to develop supply-based estimates for methamphetamine and marijuana.

Consumption-based and supply-based estimates do not always agree about the amount of cocaine shipped into the United States. According to consumption-based estimates, Americans used 291 metric tons in 1998; according to the cultivation-based estimates, 204–352 metric tons could have entered the States in 1998. Cultivation-based estimates should be higher than consumption estimates because the former do not fully account for consumption outside the U.S., for quantities seized by State and local authorities, and for amounts otherwise lost in South America. Therefore the cultivation-based estimates should exceed the consumption-based estimates, but that is not always the case.

In contrast, after 1996, the event-based consumption estimates are lower than our consumption estimates. This relationship was expected, because the events understate the flow of cocaine into the United States. Thus, the event-based consumption estimates should provide a lower limit on U.S. consumption.

Consumption-based estimates do not fully agree with supply-based estimates for heroin, but the differences are not great. Colombia seems to produce somewhat less heroin, and Mexico seems to produce somewhat more heroin, than can be accounted for by the consumption-based estimates. This difference might be explained by incorrect information about processing efficiencies in Colombia and Mexico, because estimates of processing efficiencies are based on Southwest and Southeast Asia studies.

Although these estimates are imprecise, they are sufficiently reliable to conclude that the trade in illicit substances was somewhat less than $70 billion per year during the latter part of the 1990s, according to consumption-based estimates (Table 1).² The costs to society from drug consumption, however, exceed the amounts spent on drug abuse. Drug use fosters crime; facilitates the spread of catastrophic health problems, such as hepatitis, endocarditis, and AIDS; and disrupts personal, familial, and legitimate economic relationships. The public bears much of the burden of these indirect costs because it finances the criminal justice response to drug-related crime, a public drug-treatment system, and anti-drug prevention programs.

Table 1—Total U.S. Expenditures on Illicit Drugs, 1988–2000 ($ in billions, 1998 dollar equivalents)

Columns may not add due to rounding. Estimates for 1999 and 2000 are projections.

Sources: See Tables 3 through 10.

Table 2—Supply-Based Estimates of Cocaine and Heroin Available for Consumption in the U.S. (pure metric tons)