Note: Subscriptions to the FAS Drug Policy Analysis Bulletin may be requested by sending email to email@example.com, please be sure to reference the publication title
Appraisals of the Adverse Health Effects of Cannabis Use: Ideology and Evidence
by Wayne Hall
The cannabis policy debate in the USA and many other Western countries has often been represented as a forced choice between two positions: Doves who argue that cannabis use is harmless, and hence it should be legalized; and Hawks who argue that cannabis use is harmful to health, and hence should continue to be prohibited. This false antithesis has prevented a realistic appraisal of the adverse health effects of cannabis (Hall, 1997). It has meant that the public have been exposed to two polarized views of the adverse health effects of cannabis dictated by their proponents' views on the legal status of cannabis. The Doves focus on the modest health risks of intermittent cannabis use; the Hawks emphasize the worst case interpretation of the evidence on the risks of chronic cannabis use. There seems to be an implicit agreement between Doves and Hawks that the acute health effects of intermittent cannabis use provide at best a weak justification for prohibition. The Doves stress that there is no risk of overdose from cannabis. The Hawks respond by pointing to the possibility of death or serious injury in a motor vehicle accident if cannabis users drive, and to the social consequences of engaging in risky sexual and other behavior while intoxicated by cannabis.
Points of Dispute
Hawks and Doves both accept that there is some respiratory risk from sustained heavy cannabis smoking but they disagree about its magnitude because they use different bases to assess it. Hawks stress the fact that on a puff-for-puff basis marijuana smoke contains more carcinogens and toxins than tobacco smoke (Tashkin, 1993). Doves (e.g. Zimmer and Morgan, 1997) point out that this comparison ignores two facts: that there are many fewer regular marijuana than tobacco smokers under current policies, and the average marijuana user smokes many fewer joints in a day than a daily tobacco user smokes cigarettes. Doves also discount the respiratory effects (e.g. Zimmer and Morgan, 1997), arguing that on current patterns of use the attributable risk of cannabis smoking to respiratory disease is very small by comparison with that of tobacco smoking; they ignore the amplification of risk among the minority of regular heavy marijuana smokers, many of whom are also daily tobacco smokers (Hall, 1998).
Is Cannabis a Drug of Dependence
One of the most contested issues is whether cannabis is a drug of dependence. Doves argue that it is not because it does not have a clearly defined withdrawal syndrome. Hawks cite animal evidence of the development of tolerance to the effects of THC and the occurrence of withdrawal symptoms. Both arguments depend upon a narrow view of dependence that makes withdrawal and tolerance sine qua non for dependence. Modern concepts of dependence place greater emphasis on impaired control over use and continued use despite problems caused or exacerbated by drug use. In the latter sense, there is no doubt that some cannabis users want to stop or cut down, and find it very difficult to do so without assistance and support (Hall et al, 1994). Epidemiological studies (e.g. Anthony et al, 1994) indicate that impaired control over cannabis use is the most common form of drug dependence in the community after tobacco and alcohol. Doves contest these prevalence estimates and reports of people requesting assistance to stop using cannabis (e.g Zimmer and Morgan, 1997). They argue that the apparent increase in persons complaining of cannabis dependence is a by-product of drug testing and the promotional activities of a "cannabis treatment industry". Yet, population surveys suggest that one in ten of those who have used cannabis in their lifetimes have met criteria for dependence at some time, and 80% of these people have not sought treatment (Hall et al, 1994). In Australia, moreover, cannabis use is highly prevalent, drug testing is still rare and there has not been a cannabis treatment industry. Yet treatment services that traditionally treat people who are alcohol and opiate dependent have seen a steady increase in the numbers requesting help to stop using cannabis (Hall et al, 1994). The most contentious issue of all is the explanation of the association between heavy adolescent cannabis use and the risk of using harder drugs (MacCoun, 1997). Hawks see the association as evidence of the deleterious effects of cannabis on the development of young people, and hence, as a potent reason for continued prohibition. Doves, by contrast, see it as a consequence of two things: (1) the types of troubled adolescents who begin cannabis use early and become heavy cannabis users; and (2) the shared illicit drug markets for cannabis and harder drugs. There is evidence in favor of both (1) and (2): adolescents who initiate cannabis use early, and who become heavy users, are independently at higher risk of using other drugs (e.g. Fergusson and Horwood, 1997); and cannabis users are also more likely to keep company with other heavy drug using peers. But the association is not wholly explained by pre-existing risk and peer group affiliations so Hawks can still legitimately argue that heavy cannabis use by adolescents predicts an increased risk of harder drug use (MacCoun,1997). Hawks also contend that heavy cannabis use produces an "amotivational" syndrome. There is reasonable self-report data that cannabis intoxication can affect motivation but it seems unnecessary to invoke an "amotivational syndrome" to explain the narrowed interests, loss of motivation and achievement seen in some chronic heavy cannabis users. It is simpler to regard these as symptoms of chronic cannabis intoxication (Hall et al, 1994). A more realistic understanding of the health effects of cannabis demands more appraisals that are not driven by the appraisers' views on the legal status of cannabis (e.g. Hall and Solowij, 1998). Debates about cannabis, and drug policy more generally, are too important to be left to the Hawks and Doves. The challenge is finding the institutional arrangements and professional incentives that will encourage non-partisans from a variety of relevant disciplines to provide fairer appraisals of the health and social consequences of cannabis use and predictions about the likely effects of changes in the legal status of cannabis use.
Anthony, J.C., Warner, L.A., and Kessler, R.C. (1994) Comparative epidemiology of dependence on tobacco, alcohol, controlled substances and inhalants: basic findings from the National Comorbidity Study. Clinical and Experimental Psychopharmacology, 2, 244-268.
Fergusson, D. and Horwood, (1997) Early onset cannabis use and psychosocial adjustment in young adults. Addiction, 92, 279-296.
Hall, W. (1997) The recent Australian debate about the prohibition on cannabis use. Addiction, 92, 1109-1115.
Hall, W. (1998) The respiratory risks of cannabis smoking. Addiction, 93, 1461-1463.
Hall, W. and Solowij, N. (1998) The adverse effects of cannabis use. Lancet, 352, 1611-616.
Hall, W., Solowij, N. & Lemon, J. (1994) The Health and Psychological Effects of Cannabis Use. National Drug Strategy Monograph Series No. 25. Canberra, Australian Government Publication Service.
MacCoun, R. (1997) In what sense (if any) is marijuana a gateway drug? FAS Drug Policy Analysis Bulletin, Number 2., 6-10.
Tashkin, D.P. (1993) Is frequent marijuana smoking harmful to health? Western Journal of Medicine, 158, 635-637.
Zimmer, L. and Morgan, J.P. (1997) Marijuana Myths: Marijuana Facts. Lindesmith Center, New York.
Can Supply Factors Suppress Marijuana Use By Youth?
by Jonathan Caulkins
Monitoring the Future has consistently measured high school seniors' self-reports of drug-use and related variables for many years, and these time series show some striking correlations. One of the most widely reported is the strong negative correlation between prevalence of drug use and perceptions of the riskiness of that use (see, e.g., Bachman et al., 1988). For example, between 1981 and 1995 the correlation between four measures of high school seniors' marijuana use and their perceptions of the risk of smoking marijuana at three different intensities almost all ranged between -0.75 and -0.97, with an average correlation of -0.88. This is sometimes contrasted with the fact that while there were substantial changes in self-reported use, there were only modest variations in reports of marijuana's availability. For example, past-year prevalence of marijuana smoking fell from a peak of 50.8% for the class of 1979 to a trough of 21.9% in 1992. The fraction of 12th graders reporting that it would be "fairly easy" or "very easy" to get marijuana also reached its peak and trough in 1979 and 1992, respectively. However, the absolute magnitude of the decline was much smaller, falling only from 90.1% to 82.7%.
Some see in these data evidence that supply reduction has limited capacity to influence use. For example, Johnston et al. (1996, p.276) write that "Overall, it is important to note that supply reduction does not appear to have played a major role in perhaps the two most important downturns in drug use which have occurred to date, namely, those for marijuana and cocaine. . . In the case of cocaine, perceived availability actually rose during much of the period of downturn in use. These data are corroborated by data from the Drug Enforcement Administration on trends in the price and purity of cocaine on the streets. In the case of marijuana, availability remained almost universal to this age group over the last 18 years, while use dropped substantially until 1993. . . .What did change dramatically are young peoples' beliefs about the dangers of using marijuana and cocaine" (emphasis in original in all cases).
There is no doubt that cocaine prices fell and availability increased during the 1980s, a time when cocaine use was declining, or that those trends raise serious questions about the ability of supply control efforts to control cocaine use, particularly given that the level and intensity of cocaine control efforts increased dramatically during that period. It is not clear, though, that the case is as strong regarding marijuana.
Consider first the availability of marijuana. That the indicator of availability was maxed out during this period does not necessarily mean that availability or lack thereof never affects use. It may simply mean that the measuring scale was not calibrated properly. When repairing an electrical device, if the voltmeter reading goes off the scale, one just switches to a more appropriate scale. One doesn't conclude that electricity has no influence on the operation of the machine.
A theoretical argument for this possibility is the observation that availability is not a binary condition, so availability can vary meaningfully over time even if the drug is readily "available" throughout the period in question. Regular heroin users would probably report that heroin was "fairly easy" or "very easy" to obtain, but we know they often spend 30 to 40 minutes per purchase trying to obtain it (Rocheleau and Boyum, 1994). An empirical argument for this possibility is the observation that the seniors' reports of availability were actually quite highly correlated with their self-reported levels of use. The specific correlation depends on the measure of use and the range of years, but it is generally between 0.75 and 0.85.
Of course correlation does not imply causality, and causality could run in the opposite direction. High prevalence could cause marijuana to be more available because it gives respondents a reason to be in regular contact with sellers. Nevertheless, it is perhaps safer to remain agnostic in the face of Monitoring the Future's availability data than to conclude that availability does not influence use.
The story is similar with respect to prices. Figure 1 shows a strong negative correlation between median national marijuana prices and seniors' self-reported use between 1981 and 1997. Unfortunately price data are not available before 1981, and the series is noisy because there are relatively few observations per year and because we have no information on the quality or potency of the marijuana. Furthermore, these are prices paid by enforcement agents and are not restricted to the market for youth. Nevertheless, the negative correlation is clear. (The correlation coefficient is between -0.79 and -0.95 depending on the measure of use.)
Again correlation does not imply causality, and causal relations could run in the opposite direction. Price and measures related to consumption are determined simultaneously in a market. Without a good independent measure of supply or demand it is hard to interpret their relationship. For example, increased use could help drive down prices through enforcement swamping (Kleiman, 1993), although it does not seem likely that this could affect national prices. High school seniors account for a modest portion of all consumption, and overall prevalence did not follow exactly the same pattern as that seen in the seniors' data. More generally, there may be other factors, such as demographic trends, that drive changes in both prevalence and price. The number of 15-19 year olds was highly positively correlated with prevalence (average correlation over four prevalence measures was 0.9) and price (correlation of about -0.85). 
Hence, the point of this essay is not to argue that changes in supply or supply control efforts substantially influence use by youth. One cannot draw such sweeping conclusions from such short time series, at least not without geographically disaggregated data. However, it would likewise be premature to dismiss the importance of supply factors on the basis of the Monitoring the Future data. Only more refined analysis can tease apart causal relationships in a highly endogenous system of variables.
 The levels of self-reported use were lifetime prevalence, annual prevalence, past-month use, and past-month daily use. The levels of use in the riskiness questions were trying marijuana once or twice, smoking occasionally, and smoking regularly. The one relatively low correlation was between lifetime prevalence and perceived risk of regular use, which was only -0.59.
 The marijuana price series was constructed from the Drug Enforcement Administration's STRIDE data set as part of work on two unrelated research projects. The methods are similar to those described in Caulkins (1994). For more information on STRIDE, see Frank (1987).
 Price correlation computed from 1981 to 1997 for mean (-0.84) and median (-0.86) price. Prevalence correlation computed from 1975-1997; restricting to 1981-1997 increases the average correlation to 0.94.
Bachman, Jerald G., Lloyd D. Johnston, Patrick M. O'Malley, and RH Humphrey. 1988. "Explaining the Recent Decline in Marijuana Use: Differentiating the Effects of Perceived Risks, Disapproval, and General Lifestyle Factors." Journal of Health and Social Behavior. Vol. 29, pp.92-112.
Caulkins, Jonathan P. 1994. Developing Price Series for Cocaine. MR-317-DPRC, RAND, Santa Monica, CA.
Frank, Richard S. 1987. "Drugs of Abuse: Data Collection Systems of DEA and Recent Trends." Journal of Analytical Toxicology. Vol. 11, Nov./Dec. pp.237-241.
Johnston, Lloyd D., Patrick M. O'Malley, Jerald G. Bachman. 1996. National Survey Results on Drug Use from The Monitoring the Future Study, 1975-1995. US Department of Health and Human Services, Washington, DC.
Kleiman, M.A.R. 1993. Enforcement Swamping: a Positive-Feedback Mechanism in Rates of Illicit Activity. Mathematical and Computer Modeling, 17, 65-75.
Rocheleau, A.M. and D. Boyum. 1994. Measuring Heroin Availability in Three Cities. Office of National Drug Control Policy, Washington, DC.
Marijuana Arrests and Incarceration in the United States
by Chuck Thomas
There were more than 700,000 marijuana arrests in the United States in 1997. This was the largest number in U.S. history. Of these arrests, 87% were for possession rather than sale or manufacture. The percentage of possession arrests has been at least 80% for more than a decade, and it has been rising throughout the 1990s. The total number of annual marijuana arrests, having dipped in the 1980s, has been rising sharply since 1992.
It is often asserted that these arrests rarely lead to any substantial penalty, and that therefore the costs of the current high-arrest policy, both to those arrested and to the correctional system, are modest. Some recent figures from the Justice Department's Bureau of Justice Statistics (BJS) cast doubt on that assertion.
Calculations based on recent BJS reports suggest that, at any one time, 59,300 prisoners charged with or convicted of violating marijuana laws (3.3% of the total incarcerated population) are behind bars, at a total cost to taxpayers of some $1.2 billion per year. They represent almost 12% of the total federal prison population and about 2.7% of the state prison population. Of the people incarcerated in federal and state prison and in local jails, 37,500 were charged with marijuana offenses only and an additional 21,800 with both marijuana offenses and other controlled-substance offenses. Of the marijuana-only offenders, 15,400 are incarcerated for possession, not trafficking.
The Estimates: State and Federal Prisons
The BJS report provides data from the 1997 Survey of Inmates in State and Federal Correctional Facilities. According to the report, 12.9% of the drug prisoners in state prison and 18.9% of those in federal prison were incarcerated for marijuana/hashish offenses. The report uses the numbers 216,254 and 55,069 as the total numbers of state and federal inmates, respectively, for all drug offenses. Using these numbers, the total number of people incarcerated for marijuana offenses would be 27,900 in state prison and 10,400 in federal prison, for a total of 38,300 marijuana prisoners.
However, this estimate of the number of marijuana prisoners is too low, as it is based on an estimated total number of all prisoners which the BJS report notes is an underestimate.
The BJS report's estimates of the total number of drug prisoners represents 20.7% and 62.6% of the total estimated state and federal inmate populations, respectively. But the report notes that its estimated 1,046,705 state inmates and 88,018 federal inmates represent undercounts. Excluded from the BJS estimate of federal inmates were unsentenced inmates and those prisoners under federal jurisdiction but housed in state and private contract facilities. Those prisoners who were under state jurisdiction, yet held in local jails or private facilities, were excluded from the estimated number of state prisoners.
An even newer BJS report provides accurate prisoner counts as of 30 June 1998--a total of 1,102,653 state prisoners and 107,381 federal prisoners.  In the Survey of Inmates, marijuana prisoners composed 2.7% of the state prison population and 11.8% of the federal population. Assuming that the proportions of drug prisoners to all prisoners--and of marijuana prisoners to all drug prisoner--was the same in the total prison population in June 1998 as in the population subject to the Survey of Inmates, there would be 29,800 marijuana prisoners presently incarcerated in state prisons and 12,700 marijuana prisoners presently incarcerated in the federal prison system, for a total of 42,500 marijuana prisoners.
This number is surprisingly high; the only recent published estimate, in a report by the Marijuana Policy Project issued in November 1998, put the figure at 29,300. 
Additional data obtained from a BJS official distinguish between offenses involving "marijuana only" and "marijuana and other drugs" (usually cocaine/crack).  According to the 1997 BJS prisoner surveys, 16,435 state prisoners and 8,150 federal prisoners were incarcerated for "marijuana only" offenses. Adjusting for the June 30, 1998 prisoner counts yields estimates of 17,600 state prisoners and 10,000 federal prisoners incarcerated for offenses involving only marijuana, for a total of 27,600 "marijuana only" prisoners.
BJS officials also estimated that 42% of state "marijuana only" prisoners and 23% of federal "marijuana only" prisoners were incarcerated for possession, not "trafficking."  ("Trafficking" includes "possession with intent to distribute.") Applied to the previously calculated estimates, as adjusted for the June 1998 prisoner counts, there would be 7,400 state prisoners and 2,300 federal prisoners incarcerated for marijuana possession only, for a total of 9,700 prisoners.
As opposed to prisons, which primarily hold persons convicted and sentenced to a year or more, jails generally hold persons awaiting trial or sentenced to less than a year. As of June 30, 1998, there were a total of 592,462 local jail inmates in the United States. An estimated 22% were incarcerated for drug law violations in 1996. Assuming that the same percentage applied in 1998, this amounts to 130,300 people in jail for violating drug laws. Assuming that the percentage of drug offenders in jail incarcerated for marijuana offenses is 12.9% (the same as the percentage of drug offenders incarcerated for marijuana offenses in state prison), there would be 16,800 people in jail for marijuana offenses. Assuming that 59% of those offenders were "marijuana only" offenders (the same percentage as in state prison), there would be 9,900 people in jail for marijuana only.
Both assumptions are very conservative: the proportions of people in jail for marijuana offenses, and for "marijuana only" offenses, are probably greater than those proportions among state prison inmates, because the penalties for marijuana offenses, compared to other drug offenses, are more likely to fall in the lower (jail) range rather than in the higher (prison) range.
Finally, an estimated 58% of drug law violators in jail are locked up for possession, not trafficking.  Assuming that the same percentage applies to "marijuana only" offenses, there would be 5,700 people in jail for possessing only marijuana.
Totals and Costs
Adding the jail and prison estimates gives a total of 59,300 people incarcerated for marijuana offenses. Using the adjusted estimates for "marijuana only" gives a jail-plus-prison total of 37,500 people incarcerated for marijuana without any other drugs involved. (To be even more precise, this figure ought to be adjusted to reflect the fact that the "lead charge" reflected in the Survey of Inmates may not be the only, or even the primary, reason a person is in prison; the data does not tell us whether this adjustment would, on balance, be up or down.) Finally, using the adjusted estimates for possession gives a jail-plus-prison total of 15,400 people incarcerated for possessing only marijuana.
At an average annual cost per prisoner-year of more than $20,000, the total cost to taxpayers of marijuana-related incarceration reaches more than $1.2 billion per year. (This does not include the cost of investigating, arresting, and prosecuting the hundreds of thousands of marijuana users arrested every year.)
The benefits of marijuana prohibition and its enforcement have long been the subject of debate. For example, a National Academy of Sciences report recently concluded that "there is little evidence that decriminalization of marijuana use necessarily leads to a substantial increase in marijuana use."  However one judges the merits of that debate, the latest figures cast serious doubt on the argument that marijuana incarceration costs are low enough to be ignored.
 Crime in the United States: 1997, FBI Division of Uniform Crime Reports; Washington, D.C.: U.S. Government Printing Office, 1998; pages 221-222.
 Data print-outs from FBI Division of Uniform Crime Reports (e.g., "Estimated Drug Abuse Violations Arrests, 1979-1993") and Crime in the United States volumes published in 1975-1979 and 1992-1998.
 "Substance Abuse and Treatment, State and Federal Prisoners, 1997," Bureau of Justice Statistics; December 1998.
 "Prison and Jail Inmates at Midyear 1998," Bureau of Justice Statistics; March 1999.
 "Marijuana Arrests and Incarceration in the United States: Preliminary Report," Marijuana Policy Project; November 1998.
 Unpublished data from the 1997 Survey of Inmates in State Correctional Facilities and the 1997 Survey of Inmates in Federal Correctional Facilities, conducted by the Bureau of Justice Statistics, from June through October, 1997.
 "Profile of Jail Inmates, 1996," Bureau of Justice Statistics; April 1998.
 The 1997 Corrections Yearbook, Criminal Justice Institute; September 1997.
"Marijuana and Medicine: Assessing the Science Base," Institute of Medicine; Washington, D.C.: National Academy Press, 1999; page 3.26 (prepublication copy).
c/o Federation of American Scientists
307 Massachusetts Ave., NE
Washington, DC 20002