Federal laws and policies to control the effects of the use and sale of marijuana are a failure.

Marijuana arrests doubled throughout the 1990s with no discernable impact on use, safety, or availability. Indeed many key indicators portray a situation getting worse rather than stabilization or improvement. Use has increased, potency has been on the rise, availability has improved, and prices are down. The public costs of marijuana arrests have increased as these offenses occupy more and more of police time and resources. The private costs of these policies have escalated as well with some 700,000 people arrested annually. The social costs, though, also include demographic impacts and their effect on society. Marijuana possession and sales arrests disproportionately impact young males between the ages of 15 and 24 as well as black adults, an impact that has likely intensified as marijuana arrests have increased. These disproportionate impacts nurture alienation from the rule of law, a social cost that should not be trivialized.

The fiscal and social costs of marijuana arrests have long been a burden to state and local governments. Many states and municipalities have created significant exceptions to criminal penalties for offenses involving small amounts of marijuana, particularly possession for personal use.

However an examination of the characteristics of marijuana purchases and the demographic characteristics of drug sellers indicate the extent these local policy decisions have influenced the market for and the availability of marijuana. One especially revealing piece of data is an estimate by the National Survey on Drug Use and Health (NSDUH) that there were over 1 million drug sellers in 2002 under the age of 18 - an obvious factor in the continued availability of marijuana to teenagers and adolescents.

Marijuana policy is one in which federal responsibilities are delegated to the states and neither are held accountable for the results. Federal marijuana policy is to prohibit manufacture, distribution, supply, and use of marijuana except for approved research projects. Federal policy relies on state and local law enforcement to deter and prevent marijuana sales to and use by consumers, however states and municipalities can not afford to fully enforce this federal mandate. Federal dependence on local police agencies guarantees inconsistent
enforcement of marijuana laws throughout the country, insufficient to accomplish the policy objective of achieving effective control of the marijuana market. Such control continues to elude federal, state, and local authorities and has for the last generation.

Policy analysis is about results. Does a policy produce its intended result and who does it affect? The role of mathematics in policy analysis is the same as it is in science. The purpose of numbers is to provide measurement, of course, but their real function is to provide certainty. There is a considerable amount of data available in the field of drug policy, particularly with respect to marijuana use and the enforcement of marijuana laws in the United States. There is ample data available to review the performance of marijuana policy over the last few decades. Marijuana arrests increased dramatically in the 1990s. Did this policy produce its intended results of reducing the social cost of marijuana use?

Marijuana laws and more importantly marijuana arrests are instruments of policy. They are used by policy makers at various levels of government in attempts to achieve specific objectives.

The relative harshness of state penalties for marijuana possession and sales is subject to the discretion of state legislatures. The level of enforcement of arrests for marijuana possession and/or sales is subject to the discretion of police and prosecutors. The use of discretion in enforcing marijuana prohibition is not arbitrary, but instead reflects deliberate policy decisions by policy makers.

The policy may be as simple as sending a message to youngsters that marijuana use will not be tolerated, or it may involve a more sophisticated strategy to disrupt local drug markets by increasing arrests for both possession and sales in particular areas of a city. Nonetheless, marijuana arrests have costs and benefits just like any other instrument used to achieve public policy objectives.

Marijuana arrests emerged as a significant law enforcement activity between 1965 and 1970. According to the National Commission on Marihuana and Drug Abuse:

Arrests, prosecutions, convictions and sentences of imprisonment all increased at both the federal and state levels. Marihuana [sic] arrests by the U.S. Bureau of Customs increased approximately 362% from fiscal year 1965 to 1970. Arrests by the Bureau of Narcotics and Dangerous Drugs, an agency which concerns itself primarily with sale, rose 80% from 1965 to 1968. Because major responsibility for enforcing the possession laws lies at the state level, state arrests rose dramatically (1,000%) during the five years from 1965 to 1970. [1]

Most marijuana arrests were accidental in nature, according to the Commission:

There was little formal investigative effort to seek out violators of the possession laws [by 1971]. Instead, 69% of all marihuana arrests arose from spontaneous or accidental situations where there had been no investigation at all. Well over half of these spontaneous arrests occurred when police stopped an automobile and saw or smelled marihuana. The remaining spontaneous arrests occurred when police stopped persons on the street or in a park and discovered marihuana. [2]

This 1972 Commission concluded that:

The salient feature of the present law has become the threat of arrest for indiscretion. The high percentage of cases which, after arrest, are disposed of by dismissal or informal diversion attests to the ambivalence of police officials, prosecutors and judges about the appropriateness of existing law. Anyone processed through the entire system does run a
risk of incarceration, especially when the individual had a prior record and the offense was sale or possession of a significant amount. [3]

During the 1990s marijuana arrests increased substantially, representing a change in the use of marijuana law enforcement as a policy instrument. An increase in arrests logically suggests an increase in the costs of arrests, in terms of both their fiscal and social costs.

The fiscal costs are fairly obvious. More arrests means that police, magistrates, jail guards, prosecutors, judges, and the rest of the criminal justice system have more cases and reports to file, track, update and otherwise occupy their time, all paid for by the general public.

Arrests also have impacts on individual lives beyond the fiscal cost to the public at large. Marijuana arrests make criminals out of otherwise law-abiding citizens. Indeed the primary consequence of marijuana arrests is the introduction of hundreds of thousands of young people into the criminal justice system.

The substantial increase in marijuana arrests in recent years has increased both the fiscal and social costs. However these costs must be measured against any benefits that have resulted from this change in policy.

Marijuana arrests are instruments of a supply-reduction policy with costs and benefits. For example, according to the 2004 National Drug Control Strategy Report:

The drug trade is a profit- making business, one whose necessary balance of costs and rewards can be disrupted, damaged, and even destroyed. The main reason supply reduction matters to drug policy is that it makes drugs more expensive, less potent, and less available. Price, potency, and availability are significant drivers of both addicted use and casual use. [4]

This report will document the increase in arrests, evaluate its potential benefits, and clarify its costs in terms of target populations. Among the benefits to be examined will be the impact marijuana laws have on several widely watched policy indicators monitoring such things as use, potency, and price.

Arrests are the ultimate form of supply reduction. As the National Drug Strategy Report suggests, if arrests go up substantially then other key indicators should go down. This can be portrayed and investigated graphically. The relationships between arrest rates and these indicators can also be investigated statistically. This analytical approach will be discussed and applied to the available data at the end of section 2.

The Uniform Crime Reports (UCR) Program of the Federal Bureau of Investigation (FBI) collects data on arrests from local police departments, including data on the offense and the age, sex, and race of the individual arrested. UCR data also includes the coverage population for the reporting police agency. This local agency data can be summarized at the county, state, and national level.

The Census Bureau publishes annual population estimates, also by age, sex and race. This data is also available at the county, state, and national level. The National Survey on Drug Use and Health (NSDUH) is conducted by the Substance Abuse and Mental Health Services Administration (SAMHSA) of the Department of Health and Human Services (HHS). NSDUH collects national data on a wide range of topics including drug use and sales as well as the age, sex, and race of the individual responding to the survey, as well as data on the prevalence of marijuana use at the state level.

The UCR data provides the population for the police agency's coverage area. But while data is available on the number of arrests according to different demographic characteristics, the population sizes of these sub-groups is not. Census data, though, provides a basis for estimating the sizes of subpopulations in UCR coverage areas on a county, state, and national basis. Census data indicates the composition of a county on both a real and percentage basis. The percentage breakdown of a local region by age, sex, or race can be used to determine the sizes of these sub-groups in the corresponding collection of police agency coverage areas in the same county. Thus arrest rates for the entire population can be calculated using the UCR coverage population data and arrest rates for various sub-populations can be calculated using these coverage populations and sub-population percentages derived from the US Census data.

While UCR and Census data can be combined on the basis of similar regions, UCR and NSDUH data can also be combined on the basis of similar demographic sub- populations. NSDUH provides estimates of the number of annual marijuana users by age and sex. This NSDUH data can be combined with data on the number of arrests by age and sex to calculate arrest rates based on the number of users of a particular age and sex. Arrest rates based on the number of users by race are also obtainable. Similar rates can be calculated using data on the number of drug sellers.

These three primary data sets are used in other ways. One important approach will be to compare the composition of the group of people arrested for marijuana offenses with the composition of both the group of people who use marijuana, for example, and the general population. Sub-groups in which marijuana use is more popular than the general population will have greater representation in the group of users then in the general population. Also, sub-groups that are arrested more often than others will have greater representation in the group of people arrested than in the group of users or in the general population. The UCR data can also be used to examine trends over time in arrests and arrest rates at the national, state, and local levels. The NSDUH data characterizes both the population of marijuana users and sellers as well as provides data on the characteristics of marijuana purchases.

Three forms of UCR program data will be used. The master file provides raw local agency level data on the age, sex, and race of arrests. The county file includes estimates of arrests for many areas with incomplete agency level reporting. The annual report Crime in the United States (CIUS) provides a national estimate for arrests. The CIUS data will be used to review national trends in arrests and arrest rates over time. The master file data will be used to examine the extent of arrests for marijuana possession and sales in population sub-groups, as well as marijuana arrest rates at the local level. The county file will be used to provide multi-year summaries of arrests and arrest rates at various regional levels. Arrest rates from each of these UCR sources will reflect the differences in the original files.

The National Drug Intelligence Center (NDIC) is part of the U.S. Department of Justice and a member of the U.S. intelligence community. NDIC was established in 2000 as the principle center for strategic drug counter drug intelligence. Excerpts from NDIC's 2004 National Drug Threat Assessment will be presented in relevant sections of this report in order to provide the reader with a balanced perspective on the characterizationof recent trends in important indicators. [5]

Section 1 will present NSDUH data on the demographics of marijuana use, the characteristics of marijuana purchases, and the demographics of drug selling. This data describes the behavior and economic decisions of the individuals national policy seeks to influence and is essential to evaluating its success of failure. Table 1 summarizes data on the number of marijuana users and drug sellers in each of selected age groups. Not only do a million under-18-year- olds sell drugs, presumably to members of their own age group as well as to younger
customers, but Table 1 also indicates that there is a drug seller between the ages of 13 and 17 for every 3.6 marijuana users between the ages of 13 and 17, the highest ratio of any age group reviewed.

Table 1. Marijuana Users and Sellers by Sex and Age

Male Users
Male Sellers
Female Users
Female Sellers
Total Users
Total Sellers
Ratio Users:
Age 13 to 17
Age 18 to 20
Age 21 to 23
Age 24 to 34
Age 35 to 49
Age 50+

Source: National Survey on Drug Use and Health, 2002.

National trends in marijuana arrests, arrest rates, and annual marijuana use will be examined in section 2. The CIUS trend data will be compared to trends in other important indicators used to evaluate policy at the federal level. This evaluation of marijuana arrests as a tool of national policy will be presented in Section 2. These comparisons will initially take the form of two-scale graphs, followed up with an examination of correlation coefficients that describe performance and provide a basis for both comparison and policy evaluation. The doubling of marijuana arrests has produced the opposite of the intended effect in every major indicator considered. For example an increase in arrests should produce a reduction in use and the availability of marijuana, however during the 1990s both the use and availability of marijuana increased.

Section 3 of this report will present data on marijuana possession arrests, annual marijuana use, a comparison of population, use, and arrest composition percentages, arrest rates per 100,000 annual users, and arrest rates per 100,000 general population. Five year age groups for both males and females will define age/sex categories. Adult and juvenile categories will sub-divide race categories of White, Black, Indian (Native American), and Asian. Single year age categories for males and females from age 15 to age 24 will also be examined. The primary results of this review are presented in Table 2 in which selected demographic groups are ranked according to the arrest rate for marijuana possession per 100,000 annual users. This rate controls for differences in the prevalence of marijuana use in different demographic use. For example adult blacks are 8.8% of the general population, 11.9% of annual marijuana users, and 23.1% of marijuana possession arrests. The arrest rate per 100,000 population for adult blacks is 524 per 100,000 compared to 200 for the general population using. The arrest rate per 100,000 marijuana users for adult blacks is 4,576 compared to 2,685 for the general population. This is the basis for the conclusion that marijuana law enforcement impacts adult blacks disproportionately. Some of the key findings of this report are summarized in Table 2, which indicates that marijuana law enforcement has its strongest impact on young males and black adults while its weakest impact is on females, whites, and older males.

Table 2. Marijuana Possession Arrests Among Selected Groups

Percentage of Population
Percentage of Annual Users
Percentage of Possession Arrests
Arrest Rate per 100,000 Population
Arrest Rate per 100,000 Annual Users
Male Age 18
Black Juvenile
Black Adult
Male Age 21
White Juvenile
White Adult
Males Age 35-49

Sources: United States Census Bureau Population Estimates-State Characteristics [Aggregated] (2002); National Survey on Drug Use and Health (2002); Uniform Crime Reports (2002).

Marijuana sales arrests will be examined in Section 4. Marijuana sales arrests, a comparison of population, sellers, and arrest composition percentages, arrest rates per 100,000 sellers, and arrest rates per 100,000 general population will be reviewed. Age, sex, and race sub-populations will also be examined.

The penalties and levels of enforcement in different states and local jurisdictions will be compared in Section 5. Data on penalties for marijuana possession and sales at the state level were obtained from the National Organization for the Reform of Marijuana Laws (NORML). Penalties for marijuana possession and sales will be based on one ounce quantities, and maximum penalties will be used for comparisons. Special attention will be paid to the degree of variation in arrests rates among U.S. counties with similar population sizes; the standard deviation from the average rate per population group will be considered. Appendix 1 contains data on the cost of marijuana arrests, an estimate based marijuana arrests as a percentage of all arrests and total law enforcement costs. Appendix 2 contains tables listing thecounties and local agencies with the highest marijuana arrest rates for each of 9 population size categories. (Missing data from states that do not report to the UCR program prevent these lists from providing comprehensive rankings.) Appendix 4 lists the minimum penalty for marijuana possession at the state level.

A framework for policy analysis is provided in the final section of this report. Section 6 introduces some fundamental aspects of cost-benefit analysis for evaluating drug control system originally presented by John Kaplan in 1972. Kaplan describes an analytical approach that compares the ability of a control system to increase the benefits available from a drug while reducing the harm of the drug as well as the fiscal and social cost of the control system. Kaplan participated in a 1982 analysis of marijuana policy by the National Research Council (NRC)that is introduced in section 6 in order to provide a context for assessing the data presented in this report.

The original 1982 NRC analysis is provided in Appendix 5. The NRC analysis included a review of recent research findings on marijuana and health at that time and concluded that the effects of marijuana use were not dangerous enough to over-ride other policy considerations. The NRC committee recommended further study and debate over replacing the current prohibition policy with a regulatory approach. Appendix 6 contains a contemporary review of research findings on marijuana and health that supports renewing the basic recommendation of the NRC report.

Marijuana prohibition was ineffective as a drug control policy in the 1960s. Marijuana prohibition was ineffective in the 1970s. Marijuana prohibition was ineffective in the 1980s. Throughout these three decades marijuana use became and remained widespread throughout American society. Marijuana arrests were doubled in the 1990s and marijuana prohibition has remained just as ineffective as ever. Based on the data presented in this report and predicated on a well-established framework for analysis, this report recommends serious national debate over replacing the current prohibition policy of marijuana control with a regulatory policy that provides legal access to marijuana for adults and removes the profit incentive for sales among teenage users.

[1] National Commission on Marihuana and Drug Abuse. (1972) Marijuana: A Signal of Misunderstanding. Washington, D.C.: Government Printing Office. [Reprinted as a Signet Special. New York: New American Library.]
[2] ibid
[3] ibid
[4]Office of National Drug Control Programs. National Drug Control Strategy - 2004. Washington, D.C.: Office of National Drug Control Programs. Page 31.
[5] The web site for the National Drug Intelligence Center (NDIC) is: index.htm