There are three ways to estimate the amount of marijuana available for sale in the United States. The first method is based on federal seizures of marijuana. Another method is based on intelligence of foreign production, observations about domestic production, and other data used in federal inter-agency studies of the drug supply and their subsequent published reports. The third approach is to generate an estimate of supply based on consumption models based on survey and other data on marijuana usage. Consumption models are problematic because of under-reporting of use and a general lack of data on marijuana consumption.

a) Estimates Based on Seizures

The first approach is to examine federal seizures of marijuana. According to a Library of Congress report, "an estimated 50 percent of the marijuana available in the United States is imported."(28) "There seems to be general agreement among law enforcement officials that only a maximum of 10 percent of the marijuana being smuggled into the United States is intercepted."(29) According to this report:

"Calculating the total amount of marijuana available in a given year based on the amount seized during that year necessarily provides only a rough estimate. If only 10 percent of illicit drugs are seized in any given year, then, based on the figure of 2,412,365 pounds of marijuana seized in 2002, one could estimate that in 2002 the total amount of marijuana that traffickers succeeded in smuggling into the country was roughly 24 million pounds, or about 10,889 metric tons. If one doubles that amount to take into account the domestic production of marijuana that was not seized, then the total amount would be closer to 22,000 metric tons."(30)

From 1998 to 2003 federal drug seizures averaged 2,410,571 lbs per year, On this basis one can estimate that on average traffickers succeeded in smuggling into the United States roughly 24.1 million lbs of marijuana annually, or 10,932 mt per year. As in the example above, taking domestic production into account this suggests that there is a supply of marijuana in the United States of 21,865 metric tons annually. (See Table 12.)

b) Estimated Based on Government Study Groups

The next approach is taken by inter-agency government study groups. During the 1980s and 1990s the primary source of supply data on marijuana and other drugs was the National Narcotics Intelligence Consumers Committee (NNICC), which issued an annual report on the supply of illicit drugs to the United States. The NNICC Report was:

"the product of a cooperative effort involving Federal Agencies with drug-related law enforcement, foreign and domestic policy, treatment, research, and intelligence responsibilities . . . In 1989, membership consisted of the Central Intelligence Agency, U.S. Coast Guard, U.S. Customs Service, Department of Defense, Drug Enforcement Administration, Federal Bureau of Investigation, Immigration and Naturalization Service, Internal Revenue Service, National Institute on Drug Abuse, Department of State, and the Department of the Treasury. The Office of National Drug Control Policy (ONDCP) was an observer. The Deputy Assistant Administrator for Intelligence of the Drug Enforcement Administration served as Chairman."(31)

In 1986 NNICC estimated that 8,050 metric tons (mt) of marijuana was available for consumption in the United States.(32) About one fourth (2,100 mt) was domestically produced; the rest was produced in Columbia, Mexico, Jamaica, Belize, and other foreign countries. According to NNICC, domestic production of marijuana increased considerably in the late 1980s, to 3,500 mt in 1987(33), 4,600 mt in 1988(34), and 5,500 mt in 1989 and 1990(35). Foreign marijuana available for US consumption also increased from 5,950 in 1986 to 10,070 mt in 1988.(36)

By 2002 the government's estimate of marijuana available to consumers in the United States had increased to 17,000 mt. The 2003 Library of Congress report on "Marijuana Availability in the United States and Its Associated Territories"(37) cited above relied, in part, on a 2002 ONDCP report on "Drug Availability Estimates in the United States."(38) Based on the ONDCP report and other data, the Federal Research Service concluded that:

"Using its two estimates derived for foreign- and domestically produced marijuana . . . ONDCP estimated the street availability of marijuana in 2001 to be between 10,000 and 24,000 pure metric tons."(39)

The 2002 ONDCP report indicated that "yields estimates for the availability of foreign-produced marijuana as high as 7,135 metric tons . . . [and that their] estimate for the availability of domestic marijuana rang[ed] between 5,577 and 16,731 metric tons."(40)

On the basis of these reports, domestic marijuana production was cited in reports on international drug production by the US Department of State at the level of 10,000 mt in 2002(41), 2003(42), and 2005(43).

After the 10,000 mt estimate of US domestic marijuana production was publicized in a December 2006 report in the Bulletin of Cannabis Reform(44) that received extensive media attention in the United States and other countries, the US Government lowered their official estimate. The 2006 State Department report places foreign marijuana production marketed to US consumers at 5,000 mt(45), and the 2007 State Department report estimates that 4,000 mt of foreign grown marijuana is marketed to the United States.(46) The 2007 National Drug Threat Assessment by the National Drug Intelligence Committee (NDIC) estimates domestic marijuana production to be 4,708 mt in 2006, based on the premise that law enforcement was able to seize and destroy 40% of the total crop.(47)

The claim that law enforcement seized 40% of the crop is suspect for several reasons. A 1982 report by DEA, for example, indicated that in most states eradication efforts seized 10 to 20% of marijuana grown there.(48) A 1994 report by ONDCP suggested that marijuana eradication programs on average eradicated 20% of all marijuana grown in the US.(49) As indicated above, the 2003 Federal Research Division report noted that it is widely recognized that law enforcement is only able to seize about 10% of the drugs reaching the US market.(50) The most recent State Department and NDIC reports provide a combined estimate that in 2006 at least 8,700 mt of marijuana was available for sale in the United States.

c) Estimates Based on Consumption Models

The most recent government produced report on marijuana consumption is a highly flawed 2001 ONDCP report titled "What America's Users Spend on Illegal Drugs, 1988 - 2000."(51) The most conspicuous problem with this report is its failure to account for the consumption of the supply of marijuana reported above. This 2001 ONDCP report estimates that Americans only consumed 1,047 mt of marijuana in 2000 and 927 mt in 1999.

The estimation of consumption used in this 2001 ONDCP report is based on the number of monthly marijuana users as estimated by the 2000 National Household Survey, a calculation of the average consumption of marijuana expressed in joints according to early 1990s surveys, and the assumption that a joint consists of .4 gr of marijuana. This approach is flawed for several reasons.

1) The report's estimate of total consumption is inconsistent with government reports of the total supply of marijuana available on the US market. The 2002 ONDCP report on the availability of marijuana in the United States discussed the inconsistency of this consumption estimate with contemporary marijuana seizures:

"The result of the above calculations-that 927 metric tons of marijuana were consumed in the United States in 2000-must be regarded with some skepticism when marijuana seizure data for 2000 are acknowledged. According to the Federal-wide Drug Seizure System, in 2000, approximately 1,200 metric tons of marijuana were seized in the United States, and a large portion of the seized marijuana was from foreign sources. Thus, according to these estimates the amount of marijuana seized exceeded the amount of marijuana consumed in the United States . . . it seems unlikely that marijuana growers would continue to export into the United States when the probability of detection and seizure of product was as high as is implied by the combination of the consumption and seizure estimates."(52)

2) The report only relies on monthly marijuana users and does not include estimates of the consumption of annual marijuana users. This was also observed in the 2002 ONDCP Availability Report:

"The failure to include . . . individuals who used marijuana in the past year (but not in the past month) probably has resulted in a much lower final consumption estimate . . . the marijuana consumption estimates yielded by these calculations are likely still underestimates, in part because the NHSDA data upon which the estimates are based rely on information self-reported by users themselves. This may render the estimates considerably lower as users likely underreport the amount of marijuana they consume . . . The Full Market Model provides a much higher, alternative estimate for the amount of marijuana consumed in the United States. DEA's Statistical Services Section yielded a marijuana consumption estimate of 4,270 metric tons for 2000."(53)

3) The report relies on an estimate of the weight of a "joint" that is inconsistent with other data. (See discussion above in Section 2.) A standard of .75/gr is a more realistic parameter for a consumption estimation model.

4) The consumption estimate was based on survey data that was obtained before the survey method was improved, providing more accurate and larger estimates of both monthly and annual marijuana use from 2002 on.

5) The consumption model was based on the assumption that the average monthly marijuana user consumed 18.7 joints per month. This parameter is based on the flawed assumption that the statistical distribution of monthly consumption amounts is a normal distribution and, consequently, that an average consumption is an accurate representation of all monthly users. More detailed survey data on marijuana consumption is available from the same survey the report relied on for this figure.

An improved consumption model has been prepared in conjunction with this report based on the following assumptions and parameters:

1) The consumption model should incorporate all 25 million annual users of marijuana reported in the most recent (2005) National Survey on Drug Use and Health.

2) The model should incorporate data on the number of days marijuana was used; NSDUH asks respondents how many days they used marijuana during the year.

3) The model should incorporate an estimated weight of the marijuana "joint" as .75 gr.

4) The model should incorporate survey data on the frequency and amount of marijuana consumed, and differences between the consumption practices of males and females.

5) Individuals who used marijuana between 1 and 11 times per year should be assigned a consumption level of .5 joints per usage day. This is an arbitrary assignment based on the assumption that they either used a small amount by themselves or shared one or two joints with one or more additional individuals. The amount of marijuana consumed by these individuals is relatively minor.

6) The model should inflate the number of annual users by two thirds to account for non-reporting. (See section 2 above)(54)

7) The model should use a parameter of 15% in classifying non-consumable bulk product such as seeds and stems included in the purchase price of marijuana but not consumed in end use. (55)

The National Household Survey on Drug Abuse collected data from 1991 to 1993 on the amount of marijuana consumed (in terms of the number of joints) by different categories of monthly users of marijuana (in terms of 1-2 days per month, 3-4 days per month, 5-19 days per month, or 20-30 days per month.) Using the data from all three years, and expanding the categories into consumption days per year produced the consumption levels reported for males in Table 13 and females in Table 14 below. Unfortunately, the questions soliciting this data on consumption amounts and frequency were discontinued after the 1993 survey.

In utilizing this data in a consumption model, it is assumed that the amount of marijuana consumed increases as the number of days used per year increases.

The 2005 National Survey on Drug Use and Health does include data on how many days per year the respondents used marijuana. This data, along with the consumption levels provided by the 1991 - 1993 survey data, provides another means of estimating marijuana consumption for each segment of annual marijuana users consuming marijuana for a specific number of days per year. Multiplying the number of days used per year by the number of joints consumed, the size of a joint, and the number of people using marijuana with the same frequency (days per year) is the basic method of estimating consumption. Using the consumption levels in Tables 13 and 14 along with the usage data from the 2005 survey produces evidence that Americans consumed at least 9,830 mt of marijuana in 2005. (When allowing for a larger size marijuana cigarette and increases in estimates of the number of marijuana users due to improved survey techniques, this estimate is similar to the "full market model" estimate reported above.)

d) Consolidated Estimate of Supply

The data discussed above provides four credible estimates of the annual supply of marijuana to the United States over the last several years. An estimate of 8,700 mt is contained in the most recent government reports and is contradicted by earlier government reports and the evaluation of consumption related data. The highest estimate of 21,865 mt is based on conventional wisdom among law enforcement and is based on anecdotal interpretation of data on seizures. The data reported in the 2002 ONDCP study of marijuana's availability, especially considering its review and republication by the 2003 report by the Library of Congress, provides the most credible and consistent estimates of marijuana supply, especially in light of the additional data on usage, availability and price provided earlier in this report.

Marijuana usage has remained consistent throughout the last several years, with at least 25 million Americans using the drug on an annual basis. All surveys indicate that the drug is easily available. The vast majority of the marijuana available to Americans each year is eventually bought and consumed. If this were not the case, the price would drop dramatically in response to over-supply. While the price of the drug remains high, this has not had an impact on use but instead appears, as a long term trend, to have served as an incentive for increased domestic production. Despite the evidence from the 2002 ONDCP availability report and the 2003 Library of Congress report, the consumption model presented above cannot account for their reported levels of supply.

Consequently, all the data presented above suggests the most reliable estimate of annual supply is one that takes each of four most prominent estimates into consideration: 1) the 21,865 mt estimate based on seizures and domestic production; 2) the 17,000 mt estimate reported by the Library of Congress; 3) the 8,700 mt estimate generated by combining State Department and NDIC reports; and 4) the 9,830 mt consumption estimate above derived from National Survey data. The average of these four estimates of supply is 14,349 mt of marijuana available in the US on an annual basis.