"Existing estimates about drug production and consumption are cryptic, inconsistent, and often impossible to verify. Apart from the series of studies titled What America’s Users Spend on Illegal Drugs that was produced in the 1990s under ONDCP’s auspices (see Rhodes, 1995, and Abt Associates, 2001) and the 1990s work of the Drug Availability Steering Committee (2002), many of the most-quoted estimates are not documented in a manner that allows others to assess their credibility, let alone replicate them. The large year-to-year changes in official estimates of consumption and particularly of production reduce their credibility, given the stable data on marijuana use in the U.S. population over the past decade.
"While a number of estimates are described as being 'intelligence based' or are released by intelligence agencies, this does not mean we should automatically give them high credibility. This paper identifies a number of these estimates from national and international sources that are simply implausible. Drug-market estimation is a field plagued by a lack of data and heavily dependent on assumptions; thus, estimates from both intelligence and nonintelligence agencies need to be scrutinized. Policymakers would be well served by preventing the publication of figures without peer review. If the truth is that these figures are estimated imprecisely, that fact should be noted."

Source

Kilmer, Beau; Caulkins, Jonathan P.; Bond, Brittany M.; and Reuter, Peter H., "Reducing Drug Trafficking Revenues and Violence in Mexico: Would Legalizing Marijuana in California Help?" International Programs and Drug Policy Research Center (Santa Monica, CA: RAND Corporation, October 2010), p. 44.
http://www.rand.org/pubs/occa…