- savings of $7.7 billion per year in government expenditure on enforcement of prohibition, most of it at the state and local level
- tax revenue of $2.4 billion annually (or $6.2 billion if we tax pot at the same rate as alcohol and tobacco).
The $10 billion figure could be reasonable if the model accounts for all the "large" effects and if the ceteris paribus assumption (that nothing else changes) actually holds. The toughest thing to model, in my opinion, is whether (and to what extent) legalizing and taxing marijuana would increase use (or decrease it, I guess, depending on tax rates!). If use rates are unaffected, no problem. If use changes by 5% or 10%, then the analysis gets complicated. It would have to put a number on the costs (or benefits?) to long-term health, increased Doritos sales, and myriad other factors that are tough to identify, let alone measure. Lifetime marijuana prevalence rates for high school seniors have fallen since 1997 to about 46% in the 2004 Monitoring the Future data series. I don't hear many people making the "gateway drug" argument these days, but I'm sure that (socially) conservative critics will raise the spectre of increased cocaine, meth, and heroin as well. Though supported by the pro-reform Marijuana Policy Project, I think this study probably understates the costs of the drug war. In addition to the impossible position of those prescribed medical marijuana who remain subject to prosecution, a detailed analysis would need to estimate the intergenerational costs of incarceration to families and communities. It seems as though a pilot experiment (or quasi-experimental comparison of laws across states and time) would help provide some additional traction here. What would Friedman say about banning Sudafed?