Table of Contents
I. Introduction: The Kansas Experience
II. Fighting Drugs
III. Drugs and Crime: Myths and Reality
IV. Why Drug Wars Do Not Work
V. Indirect Costs of Drug Law Enforcement: Rising Crime
VI. Policy Recommendations
A drug war was waged throughout most of the United States by federal, state, and local law enforcement agencies from 1984 to 1989. Relative to the rest of the country, however, it appears that some states were not major battlegrounds in this war. While those states did increase drug enforcement efforts during this period, many started at relatively low levels of drug enforcement and increased their commitment to drug control at relatively low rates. Indeed, while most of the nations law enforcement agencies were increasing their efforts against drug market activity much faster than they were increasing their efforts against other kinds of crime throughout this five-year period, Kansas, for example, actually increased nondrug arrests more rapidly than drug arrests.
In hindsight, it appears that Kansass relatively modest involvement in the nationwide drug war during the 198489 period has resulted in its citizens being relatively safe from crime. This assertion will come as a surprise to many people, but reality mandates that when scarce resources are used to do one thing, they cannot be used to do something else. Legislators face competing demands for the allocation of tax dollars and simultaneous pressure to hold down taxes. As a consequence, substantially fewer resources are available for law enforcement than would be necessary for solving even a majority of crimes reported or imprisoning most of the criminals caught and successfully prosecuted.
Because law enforcement resources are limited, decisions made within the criminal justice system regarding the control of illicit drug markets are part of a general resource allocation problem. This means that when a decision is made to wage a war on drugs, other things that criminal justice resources might do have to be sacrificed. One of the consequences of using a large portion of the criminal justice systems resources to control drugs is that violent and property criminals are not caught until they have committed a relatively large number of crimes, if they are caught at all, and even after they are caught for some crimes, they can be forced out of prison relatively early due to the admission of increasing numbers of drug criminals, freeing them to commit new crimes. Thus, active participation in a drug war often means higher rates of property and violent crime. The fundamental fact of limited criminal justice resources means that getting tough on drugs inevitably translates into getting soft on nondrug crime.
The costs of a war on drugs for many states were very high because many of the law enforcement resources committed to the drug war were diverted from the control of property and violent crime. But what of the benefits? Drugs may create many harmful effects for the individuals who consume them. However, criminalization is rarely justified on the grounds of protecting people from their own actions. Rather, criminalization requires that an action has spillover effects creating costs for others in society. In this regard, there is a widespread belief that drug use causes other crime. Indeed, if drug users are responsible for many other crimes, a strong law enforcement effort against drugs should lower the rates of property and violent crime and simultaneously discourage drug use. The nations experience with cyclical drug wars does not bear this out, however, and Sections III through V of this report explain why.
The trend toward increased law enforcement efforts against drug markets ended after 1989 in much of the United States, but Kansas and other states once again appear to be going against the trend. While drug arrests relative to total arrests fell from 1989 through 1991 in the nation as a whole, they rose in Kansas. If these trends continue in Kansas, the results are likely to be similar to those experienced in other states during the 198489 period: rising crime rates. This report provides a generic overview of the drug/nondrug crime relationship. The fact is that the 198489 drug-war period has inspired several studies exploring various consequences of drug enforcement in some major drug-war states, such as Florida, California, and Illinois, and several nationwide studies can also be brought to bear on this issue. The purpose is to examine in detail what has been learned in order to provide a clearer picture of what lies in store as states continue their increasing focus on drug enforcement.
Section III explores the coincidence of drug use and other crime, explaining that, in sharp contrast to the conventional wisdom, most users of illicit drugs do not commit large numbers of crimes against persons or property. Therefore, increasing enforcement efforts against drug offenders cannot be justified on the ground that these same offenders are inevitably also perpetrators of nondrug crimes. Even if there are other plausible reasons to attempt to control drug markets, restrained use of scarce criminal justice resources for this purpose remains advisable for two reasons. First, it is likely that the criminal justice system will be unable to effectively control these markets. Section IV explores how drug buyers and sellers react to law enforcement efforts in ways that regularly frustrate their intent. Second, even if the evidence in Sections III and IV is denied, implying that control of drug markets is both desirable and possible, alternative uses of scarce criminal justice resources may be more valuable. The scarcity of criminal justice resources means that when policing effort and prison space are devoted to the apprehension and punishment of drug offenders, these resources cannot be used to combat other crimes. The consequences of allocating scarce criminal justice resources to drug control are explored in Section V.
Sections III through V demonstrate that unrestrained use of the criminal justice system to combat illicit drugs does not produce the results that many of its advocates predict, and that the unanticipated costs of a drug war are quite high. Issues of drug policy in Kansas are discussed at the end of Section V and in Section VI, emphasizing that this policy must be analyzed in the context of a broader view of the criminal justice system, recognizing the inevitable trade-offs that arise because criminal justice resources are limited. Before turning to the issues of why drug control has failed to live up to its promises and how a more rational drug policy might be devised, however, a brief overview of drug policy in Kansas and the United States is provided.
Opiates were legal until passage of the Harrison Act in 1914, and marijuana became illegal in 1937, although both of these federal statutes were primarily tax acts rather than criminalization laws per se (some state laws regarding these drugs preceded the federal statutes). The criminal justice system has been an increasingly important tool in the fight against illicit drug consumption since the passage of these statutes. Table 1 shows a dramatic increase in this effort since 1960. Drug arrests per 100,000 population reached a peak level of 538 in 1989, a twentyfold increase from the 1960 level of 26. The drug arrest rate did not increase steadily, however. Indeed, two drug wars have been waged by the criminal justice system since World War II. The first escalation occurred between 1965 and 1970, when the drug arrest rate rose from 34 to 228, a 114 percent average annual increase. Following this burst of antidrug activity, the drug arrest rate grew slowly between 1970 and 1984, averaging about 2.6 percent per year. Renewed escalation of drug enforcement during the 198489 period saw arrests per 100,000 population rise from 312 to 538, an average annual increase of 14.5 percent. Although the growth rate of drug arrests in this latter period is dwarfed by the 114 percent annual growth of the 196570 period, the absolute increase in drug arrests per capita was largest during the 198489 period. Together the 196570 and 198489 periods account for over 80 percent of the increase in the drug arrest rate from 1960 to 1990.
Table 1. Growth of Drug Enforcement in the United States, 1960-1992
Drug Arrests per 100,000 Population
Average Annual Percentage Change
Source: U.S. Department of Justice, Crime in the United States (various years).
The 198489 increase in the drug arrest rate was followed by a period of marked retrenchment in most of the country. Drug arrests per 100,000 population fell 16.5 percent in 1990 and another 8.5 percent in 1991, from 538 per 100,000 population in 1989 to 411 in 1991. The decline in the drug arrest rate in 1990 alone is larger in absolute terms than its increase in the 14 years between 1970 and 1984. However, this retrenchment may be short lived. The national drug arrest rate was back up to 431 per 100,000 population in 1992, a 7 percent increase over the 1991 figure. Furthermore, drug arrests remain very high relative to the predrug war period: the 1992 rate is 57 percent above the 1983 rate.
National data conceal enormous variation in drug enforcement among state and local jurisdictions. The 1984 and 1989 drug arrests per 100,000 population for the states are shown in Table 2. In 1989, the peak year of the recent drug war, they range from a low of 88 (West Virginia) to a high of 1,060 (California). Kansas ranked 37th, with 233 drug arrests per 100,000 population.
Table 2. Drug Arrests per 100,000 Population, by State, 1984 and 1989
|State||Rank||1989||1984||% Change||State||Rank||1989||1984||% Change|
Source: U.S. Department of Justice, Bureau of Justice Statistics (1984 and 1989).
Kansass drug arrest rate increased 66.4 percent during the 198489 period, a figure only slightly below the national average of 72.4 percent. Nonetheless, Kansas started this period 55.1 percent below the national average of 312 drug arrests per 100,000, and because its increase was also slightly below the national average, it ended the period 56.7 percent lower than the average. This is only a modest difference, of course, but it does suggest that while Kansas was an active participant in the 198489 drug war, its drug-war efforts were relatively restrained.
Figure 1 shows that Kansass relative behavior has changed since 1989, however. Most of the nation has reduced drug enforcement efforts, but Kansas has increased its efforts quite markedly. The 1992 drug arrest rate for the nation was almost 20 percent below its 1989 level, while Kansas increased its efforts against drugs by 31 percent. Indeed, Kansass 1992 drug arrest rate of 305 per 100,000 population is only 29.2 percent below the national average of 431. Clearly, Kansas is rapidly gaining on the rest of the nation in terms of its status as a drug-war battleground, for two reasons: (1) most of the rest of the country appears to be withdrawing some resources from drug control and reducing drug arrests, but (2) Kansas appears to be devoting more criminal justice resources to drug control and increasing drug arrests. The consequences of these trends are explored in Section V below, but before discussing this issue, we must consider some of the consequences, real or imagined, of drug use itself.
Figure 1. Drug Arrest Rates for Kansas and the United States, 19841992
Source: Uniform Crime Reports (various years).
The fact that many persons who commit violent and property crimes are also drug users is well documented; for example, during 1988, 72.2 percent of male arrestees in 20 U.S. cities tested positive in a urinalysis for the use of any illicit drug (National Institute of Justice 1990). Similarly, a Bureau of Justice survey of 12,000 inmates indicated that over 75 percent had used drugs, 56 percent had used drugs in the month prior to their incarceration, and one-third admitted to being under the influence of drugs at the time of their offense (Wexler, Galkin, and Lipton 1989). Findings such as these appear to support the claim that drug use is a primary cause of nondrug crime, which in turn has led to increasing emphasis on the control of illicit drugs as a means of general crime prevention. However, by asking Do a substantial portion of known violent and property criminals use drugs? such research misses the mark. To make the drugs-cause-crime link, research must answer a different question: Do a substantial portion of drug users commit non-drug crime?
In this regard, consider a 1992 Bureau of Justice Statistics report on recidivism of felons on probation from 1986 to 1989 (Langan and Cunniff 1992). Table 3 summarizes some of the key results from this study, which found that drug offenders are far more likely to recidivate for a drug offense than for a violent or property offense. Furthermore, violent offenders who are re-arrested tend to recidivate most often for a new violent crime, and property offenders are most likely to recidivate for another property crime. Consider the property offense row, for instance. It suggests that 43.4 percent of the probationers in the sample who had committed property offenses had recidivated at the time of the study; furthermore, 7.4 percent were rearrested for a violent crime, 23.7 percent for another property crime, and 7.3 percent for a drug crime, suggesting that property criminals were over three times as likely to commit another property crime as to commit a drug crime. Similarly, drug criminals apparently were more than two and a half times as likely to commit another drug crime as a property crime.
Table 3. Felony Probationers Who Were Arrested for a Felony Offense While on Probation
|Most Serious Felony Conviction Offense||Percent of Probationers Arrested for:|
|Total[a ]||Violent Offense||Property Offense||Drug Offense|
Source: Langan and Cunniff 1992, table 4.6.
Other research supports this conclusion and suggests that the set of people who are drug offenders only partially overlaps with the set of people who commit Index I crimes (homicide, sexual assault, robbery, burglary, aggravated assault, larceny, and auto theft). The simple diagram in Figure 2 helps visualize the implications. Two distinct types of drug users apparently exist. First, a substantial portion of drug offenders apparently do not commit property or violent crimes. Second, many offenders arrested for violent and property crimes also use drugs. Furthermore, Florida data suggest that there are relatively few habitual offenders who are heavily involved in both drugs and other crime (Rasmussen and Benson 1994:60), so many people arrested and convicted for drug offenses are not hardened criminals.
The apparent fact that many drug consumers are not active in nondrug crime is really not surprising. Indeed, there are fundamental reasons to suggest that most consumers of drugs do not rely on predatory crime as a means of financing drug consumption. Three of them are discussed below.
Addicts and Crime
While the argument that drugs cause crime generally involves an explicit statement to the effect that drug addicts are driven to commit crimes to finance their habits, the fact is that not all drug users are addicts. Indeed, there is a large population of regular heroin users who are not addicted, just as there is a large population of nonaddicted drinkers, and these users bear the same relation to heroin addicts as normal drinkers do to problem drinkers or alcoholics (Kaplan 1983:33). Variation in the pattern and intensity of use is also found among consumers of cocaine. The NIDA National Household Survey of Drug Abuse (National Institute on Drug Abuse 1988) indicates that most cocaine users are not regular consumers of the drug. Among young adults (ages 18 to 25), fewer than one in four lifetime users (that is, individuals who indicate that they have ever consumed the drug) report use in the previous month. More strikingly, among adults over the age of 25, only 1 in 11 lifetime users reported use of cocaine in the previous month. Thus, use-frequency usually declines with age, suggesting that many users are not addicts.
Even for addicts, the drive to maintain a particular level of consumption is much weaker than popularly believed. Consider heroin, long viewed as the most addictive drug. Kaplans (1983:38) examination of studies of heroin indicates that there is a striking analogy between alcohol and heroin use. Many heroin addicts or problem users voluntarily abstain from use for substantial periods for any number of reasons, just as do problem drinkers. Indeed, addicts consumption patterns exhibit considerable variability (Moore 1977; Roumasset and Hadreas 1977). The average addict consumes well above the amount of heroin needed to avoid withdrawals roughly 70 percent of the days in a year and actually suffers withdrawals the other days. Some go without for several days and actually detoxify themselves.
An addicts demand for the drug of choice therefore appears to be flexible, even at the low level needed to avoid withdrawal. This should not be too surprising since the withdrawal pains for addicts are remarkably mild relative to popular perception (Nyswander 1956:12124; Kaplan 1983:35; Moore 1990:114). They are not extremely painful for most addicts, in part because of the low quality of many drugs available in the black market, but even in its classic form, heroin withdrawal is not that serious. Pharmacologists compare it to a bad case of one week flu (Kaplan 1983:35). In the case of cocaine, withdrawal pain is even less severe than for heroin; cravings for cocaine stem from the drugs euphoric effects rather than from avoidance of withdrawals, and on balance, it is less addictive than heroin (Gould 1990:11).
Financing Drug Consumption
Even drug addicts, who may have relatively fixed demands for drugs, may not need to turn to crime to support their habits. Officials often contend that when faced with higher money prices, drug addicts face a choice: either quit the drug and go through withdrawal, or turn to criminal methods of acquiring the necessary money (Canadian Governments Commission of Inquiry 1973:321). But this assumes that the addict has no other source of funds. Indeed, the report just quoted also pointed out that affluent persons with low-risk access to drugs, such as physicians, pharmacists, and perhaps police, generally do not turn to property crime as a consequence of drug consumption, reinforcing the idea that decisions to commit property crimes actually involve economic choices based on prices and income rather than drug cravings.
Drug consumers need not specialize in one method of income generation, and it is apparent that many have legal jobs. Indeed, studies report that use of illicit drugs is associated with higher wages among young adults. Similarly, in a 1986 survey of state prison inmates, almost 50 percent of those who reported daily drug use had been employed full-time in the year prior to their offense. Another 10 percent reported that they had been employed part-time. These incarcerated regular users of drugs also reported income from other sources: 22.8 percent had income from welfare; 30.5 percent, from family and friends; and 47.6 percent, from illegal sources.
Similar findings come from a 1989 Bureau of Justice Statistics survey of jail inmates (Harlow 1991). For example, while 77.8 percent of those who reported never having consumed a major drug listed wages or salaries as one of their sources of income, 65.6 percent of the daily major-drug users also had wages or salaries as a source of income. On the other hand, while 69.1 percent of the nonusers were employed at the time of their incarceration, only 49.8 percent of the daily users were employed. Furthermore, while 4.7 percent of the nonusers reported illegal sources of income, 29.4 percent of the users did so. These results indicate that as the frequency of drug con-sumption rises, the likelihood that an individual can maintain legitimate employment falls and the probability that the individ-ual will turn to criminal activities as a source of income increases. However, the data also illustrate that illegal sources of income are by no means the only source from which drug con-sumers obtain funds. Addiction tends to enhance the possibility that as the price of the drug rises, more effort will be devoted to income generation; but it still does not follow that addicts will commit more property crimes in order to maintain their consumption.
The Temporal Sequencing of Entry Into Nondrug Crime and Drug Use
An important question that those who claim that drugs cause crime have failed to consider is: What happens to drug demand as illegally obtained income rises? In other words, instead of asking, Does drug use lead to crime? we might ask, Does crime lead to drug use?
The fact is that studies of the temporal sequencing of drug abuse and crime suggest that criminal activities generally precede drug use. For example, a Bureau of Justice Statistics survey of prison inmates found that approximately half the inmates who had used a major drug, and about three-fifths of those who used a major drug regularly, did not do so until after their first arrest for a nondrug crime, that is, after their criminal career had begun (Innes 1988:12). Similarly, more than half of local jail inmates who reported they were regular drug users in the survey discussed above said that their first arrest for a crime occurred an average of two years before their drug use (Harlow 1991:7). Indeed, as Chein et al. (1965:6465) concluded, delinquency is not caused by drug abuse, but rather, the varieties of delinquency tend to change to those most functional for drug use; the total amount of delinquency is independent of drug use.
Once an individual has decided to turn to crime as a source of income, he or she may discover that drugs are more easily obtained within the criminal subculture and perhaps that the risks posed by the criminal justice system are not as great as initially anticipated. Furthermore, criminal activity generates the income with which to buy goods that previously were not affordable, including drugs. Thus, crime leads to drug use, not vice versa. Of course, if the individual later becomes addicted, his or her preferences may change and at that point the drugs-cause-crime relationship might well come into play. But the fact is that the portion of the drug-consuming population for which this relationship might hold is quite small, so an indiscriminate war on drugs will not be a very effective general crime control policy.
Drug enforcement policies are based on the proposition that buyers and sellers in drug markets respond to incentives. Greater enforcement against suppliers presumably will reduce the amount of drugs supplied at any given price; efforts against drug users similarly are expected to lower the net benefits of drug use so the quantity of drugs purchased at any price will be lower. Furthermore, evidence suggests that drug offenders do in fact respond to incentives. Kim et al. (1993) report the results of empirical studies of recidivism among drug offenders, using Florida data. After controlling for individual and socioeconomic factors, they find that an increase in the number of police in a jurisdiction reduces the probability of recidivating, supporting the hypothesis that an increase in the probability of arrest reduces drug crime. Furthermore, drug criminals sentenced to probation are more likely to recidivate than drug criminals sentenced to prison. Assuming that prison is a more severe punishment than probation, this result also suggests that drug criminals respond to the severity of punishment. However, given a prison sentence, an increase in the length of that sentence does not impact the likelihood of recidivating. This implies that increasing the length of prison sentences for drug offenders may not be an effective deterrent. While this is a limited sample, and apparently the only sample of drug offenders that has been used to study the deterrent hypothesis, the results are strikingly consistent with the much larger empirical literature on nondrug crime deterrence.
The fact that drug market participants respond to incentives suggests that a strong enforcement effort should reduce the consumption of drugs. In obvious contradiction, however, the drug problem is not demonstrably smaller as a result of the 198489 drug war. It is not the expectation that individual drug users respond to incentives that is flawed. The mistaken expectation in drug policy is the belief that people can respond to increasing enforcement only by curtailing sale and use.
The desires for short-term pleasure and profit are so powerful that many react to the constraints imposed by drug laws and enforcement efforts in ways that allow them to continue their drug market activities. Rather than passively accepting the effects of increasing enforcement as inevitable, drug entrepreneurs attempt to beat the police by adopting new production techniques, offering new products, and developing innovative marketing strategies. Users change their buying habits and use different drugs to maintain their pleasures in the face of rising police interference. The reactions of buyers and sellers of drugs that reduce the effectiveness of enforcement efforts are the focus of this section.
The Drug Market Is a Misnomer
Law enforcement efforts may be relatively more effective against some kinds of drugs compared to others. For example, marijuana is much bulkier and harder to conceal than cocaine or heroin, so it is more difficult to raise, smuggle, and distribute. Furthermore, even frequent users of marijuana are probably more responsive to changes in the risk posed by law enforcement than regular users of cocaine or heroin (e.g., those who are addicted). Thus, it is actually inappropriate to speak of the drug market. Instead, there are several distinct but interrelated markets for illicit drugs, which may be differentially affected by drug enforcement and/or education efforts. They are interrelated on the supply side because drug entrepreneurs often can supply more than one drug and substitute among them depending on availability. They are interrelated on the demand side to the degree that consumers find different drugs to be substitutes or complements. Let us briefly consider the evidence regarding changes in the two largest illicit drug markets.
Increasing drug enforcement can affect the demand for drugs in two ways. First, the added risk of arrest and punishment for possessing drugs directly discourages consumption. Second, if prices rise as a result of efforts to limit drug supply, the amount of illicit substances consumed will fall. In the case of marijuana these two effects have worked together to reduce consumption. The price of marijuana rose continuously from 1974 to 1984, although a simultaneous eightfold increase in potency (THC content) may have actually caused the price per unit of THC to fall over this period. During the 198489 war on drugs, however, marijuana prices rose sharply with relatively small changes in potency. These data, coupled with Reuters (1991) estimate that the severity of punishment per marijuana transaction rose between 1979 and 1988, suggest that the war on drugs may have played a significant role in reducing the demand for and supply of marijuana, the net effect being rising prices.
The story for cocaine is quite different. Estimated use of this drug started rising in 1979, and growth of consumption accelerated after the onset of rising enforcement efforts in 1984. Prices did not rise as a consequence of enforcement, and there is no discernible trend in purity. In fact, 1989 cocaine prices in nominal terms were less than half their 1979 level. Nor did enforcement increase effective punishment for cocaine. Admittedly rough estimates of the user and seller populations led Reuter (1991) to speculate that the risk of arrest for cocaine users declined between 1979 and 1988, while the change in expected punishment of sellers was less certain. Even though total arrests rose sharply, they were rising at a slower rate than the increase in users, and perhaps sellers, so the probability of arrest fell. Furthermore, surveys of high-school seniors indicate an increase in their access to cocaine. While perceptions of the availability of marijuana among high-schoolers was virtually unchanged from 1984 to 1990, the proportion of students saying it was fairly easy or very easy to get cocaine increased by about 20 percent. It appears that the drug war did not have enough impact to reduce cocaine supply. Some of the reasons for the differing consequences of the drug war on these two illegal drugs are discussed below.
Drug suppliers react to increased drug enforcement in order to reduce its impact on their activity. By changing combinations of inputs, these entrepreneurs can mitigate the effects of law enforcement on their profits and the supply of drugs. Production and distribution of a drug may require the combination of several inputs, each of which must be rewarded with an acceptable return.
The price that owners of inputs require to supply their inputs partially depends on the risks involved. Rising law enforcement efforts, and/or more severe punishment of a particular input, make supplying that input more risky and raise its price. Profit-seeking drug suppliers will replace the relatively high-priced inputs, and in the process ameliorate the impact of enforcement. For instance, if a change in enforcement policy makes punishment of adult dealers more severe while leaving punishment of juvenile dealers unchanged, then adult dealers will want higher compensation, and juvenile dealers may be substituted for adults.
Drug entrepreneurs diffuse the risks they face by employing others to make street sales, and the relative wages that must be paid to potential employees determine the makeup of the group. As the risk of arrest increases, the entrepreneur has incentives to lengthen the distribution chain, thereby personally dealing directly with a smaller number of individuals. Thus, drug entrepreneurs will be willing to pay relatively more for a few intermediary brokers who in turn set up their own network of contacts, rather than deal directly with a large number of users. Law enforcement policy may actually make production of drug distribution more labor intensive and encourage the use of juvenile pushers, who face relatively less risk, at least in terms of the severity of punishment. Thus, even if drug enforcement pushes the prices of drugs up and the quantity traded falls, the number of people involved on the supply side of the drug trade can rise. This suggests that increased enforcement can cause one measure of the drug problem, the pool of people to be arrested, to rise even as another measure, the quantity sold, falls. Indeed, as Moore (1990:13738) emphasized, there is no scarcity of human capital prepared to enter the [drug] business....The supply is not limited to those with prior criminal records or with a taste for violence and corruption. Laborers and specialists are easily recruited.
Other substitution effects are also likely. As drug enforcement efforts become effective in one geographic area, drugs will be shipped to destinations where the risks are lower. Similarly, one way to avoid the risks of shipping a drug across national borders is to increase domestic production. The success of interdiction efforts with regard to marijuana has created strong incentives to develop domestic supplies. As a result, it is now estimated that marijuana is the largest cash crop in California, the largest agricultural state in the United States in terms of the value of output. Substantial marijuana crops are grown in many other parts of the country as well. The increase in domestic sources may not have completely offset the impact of interdicted international supplies, given the price trends already noted, but it did reduce the effectiveness of law enforcement efforts by diversifying the sources of supply. A long history of drug enforcement efforts suggests that elimination of supplies coming from one area will soon lead to increased cultivation elsewhere (Reuter 1985:1316). The fact is that the total U.S. demand for drugs can be supplied by crops grown on a very small amount of the total world acreage that is suitable for growing opium poppies, coca shrubs, and marijuana plants (Reuter and Kleiman 1986:30615).
Californias effort to thwart marijuana growers with aerial surveillance shows how the industry can respond to changing constraints. As the risks of outdoor cultivation increased, growers started indoor cultivation using marijuana strains that are particularly well suited for high performance under artificial light. Automated hydroponics are used to feed plants, diesel generators are required to conceal high energy use, and thick concrete walls mask the heat buildup that might be detected by infrared sensors on aircraft. The capital-intensive production techniques are placed beneath structures that look like ordinary houses and produce four crops a year that are more potent than the strains previously grown outdoors. Despite the higher costs of production, it was estimated in 1990 that a $1 million investment could generate $75 million in profits when the wholesale price is $3,000 per pound.
Output Substitution: Increasing the Supply of One Drug in the Face of Effective Control of Another
Much of the law enforcement success against marijuana was due to interdiction efforts, which were considerably more successful than similar efforts against cocaine and heroin. For instance, in 1984 a drug task force dramatically increased its efforts to intercept drugs in the Miami area and virtually eliminated the incoming supply of marijuana. The marijuana smugglers responded by converting to cocaine smuggling because it was much more difficult to detect. So the local supply of cocaine increased, pushing the price down (Thornton 1991:109). Thus, while the supply of one drug declined, the supply of another increased. Indeed, the most important explanation of the increasing availability and use of cocaine may be the relative success that law enforcement has had against marijuana (Thornton 1991; Nadelmann 1993:45). After all, it has been estimated that as much as a third of the marijuana shipped into the United States in the early 1980s was being intercepted (Kleiman 1985, chapter 3).
Another reason to expect an increase in the supply of a drug is the potential increase in demand as consumers substitute among drugs, which tends to push price and profits up in the short run. This induces entry of new suppliers. The success of interdiction efforts with regard to marijuana means that its price is considerably higher in the United States than in source countries (Reuter and Kleiman 1986). Thus, a substitution effect led to increases in demand for cocaine and created incentives to develop more sources of cocaine supply even in the absence of any change in the relative costs of producing cocaine.
Increasing Supply by Technological Change
Entrepreneurs always face strong incentives to find ways to produce or distribute existing products at lower costs and to offer new products that will attract consumer demand. The result of such developments are broadly described as technological change, and virtually all legal markets regularly exhibit at least some technological advances. Drug entrepreneurs have even stronger incentives to look for technological change than entrepreneurs in legal markets because they have an added cost to consider and try to avoid: the arrest and punishment cost associated with illegal activity. If a drug entrepreneur can find a way to lower either production or distribution costs or to lower the probability of arrest, the business will be more profitable. Since reduced costs, including reduced risks of arrest and punishment, lead to an increase in output for a firm, a technological advance on any front will lead to an increase in market supply.
Perhaps the most dramatic change in recent years was the introduction of crack cocaine, which allowed drug suppliers to produce and sell cocaine at a much lower price than before the innovation. The increase in the supply and drop in the price of a high from consuming cocaine that resulted from this technological change may even be comparable with some of the more remarkable technological advances in computer or communications technology. Since the crack technology can be adopted fairly easily by new entrants into the cocaine market, the profitability of crack (due both to falling costs and perhaps to increasing demand, for reasons suggested above) apparently has attracted new sources of supply. Introduction of this innovation was also partially motivated by law enforcement successes against marijuana. Producers had incentives to provide a new drug at the low-price end of the drug markets, and crack served that purpose.
Other important innovations in supply are improved methods of transporting and marketing cocaine. These innovations could easily be swamping all the increased law enforcement efforts against cocaine. Increasing availability and falling prices of cocaine when law enforcement efforts to combat this drug are increasing provide strong evidence of the entrepreneurial abilities of drug suppliers.
Increasing Potency: Technological Change Arising Because of Consumer and Producer Substitution Effects
Drug statutes generally consist of three parts. First, the commodities that are declared to be illegal are described in terms of minimum potency levels. For example, a product containing any detectable amount of heroin is generally illegal. Second, given that a product contains at least the minimum statutorily defined potency, penalties are generally levied on the basis of weight. Third, different penalties are set for production, distribution, and possession, all based on weight, not potency. Smugglers caught with relatively heavy shipments face stiffer penalties, for example. Thus, punishment is clearly a function of the weight of the commodity possessed, sold, or transported. The probability of being caught is also likely to be a function of the physical volume that an individual is trying to conceal. This law enforcement focus on weight creates incentives to avoid holding heavy bundles of a drug; it also creates incentives to increase the potency of illicit drugs.
For consumers, punishment based on weight leads to stronger demand for a more potent drug variant relative to demand for a less potent variant, because smaller quantities of a high-potency drug variant are required to achieve desired effects and because they should be easier to conceal than the larger quantities of lower-potency variants, implying that the probabilities of arrest could be lower for the high-potency drug. An analogous argument applies to drug suppliers and drug smugglers. Illegality lowers the expected costs of dealing in high-potency drugs relative to the costs of dealing in low-potency drugs, due to the relative risks of detection. Furthermore, inasmuch as greater potency means that consumers will pay more for smaller weights, sellers and smugglers can reduce the sizes of their holdings and reduce their expected costs even more. Thus, as law enforcement efforts against a drug market increase, the production and transportation of a relatively potent variety of a drug become increasingly attractive at the same time that the demand for it tends to increase.
One way to increase potency is simply not to cut the drug at as early a stage in the distribution. Pure heroin and cocaine are both cut by adding nondrug substances that reduce potency before consumption. Cutting may be done at virtually any stage in the distribution chain, and several different cuts may actually be made (Thornton 1991:96), so incentives to reduce weight and increase potency may simply mean that the drugs are not cut as early in the chain. But beyond this, considerably more potent uncut drugs can also be produced. The average potency of marijuana increased by a factor of eight between 1974 and 1984, for instance, and methods currently exist that can increase the present average potency of marijuana by at least another five times (Thornton 1991:108).
Reliable data on cocaine and heroin potency are not available, although there is some evidence that the retail purity of both drugs has increased during the 1980s (Nadelmann 1993:45), and the introduction of highly potent crack cocaine suggests that this drug was made available at a lower price per dose. Furthermore, there is another piece of supporting evidence for the law enforcement causes increased potency hypothesis: during Prohibition a product containing more than 0.5 percent alcohol was illegal, and it is a well known fact that Prohibition [was] more effective at suppressing the drinking of beer, than of whiskey (Fisher 1927:29). Prohibition did more than simply create incentives to shift from low-potency beer to high-potency spirits, however. The spirits available also increased in average potency.
The evidence about increasing THC in marijuana and the consequences of liquor prohibition, coupled with the fact that most increases in potency simply involve implementation of known technology, led Thornton (1991:110) to conclude that the availability of increasingly potent illicit drugs is primarily the result of their illegality and increasing law enforcement efforts rather than the discovery of new technology. Supporting this interpretation is the fact that trends in potency are in the opposite direction in legal markets, where consumers do not have incentives to demand increased potency: the tar and nicotine content of cigarettes, the caffeine content of coffee and soft drinks, and the alcohol content of liquor consumed since the repeal of Prohibition have all tended to decline over time. Increasing drug potency is clearly a result of law enforcement efforts and entrepreneurial adjustments to avoid their consequences. Indeed, even if the amount of drugs supplied as measured by weight does not increase due to law enforcement efforts, the quantity of drug highs supplied could still increase with rising potency.
It should not be surprising to find that the overall results of drug enforcement policy are far different from the results policy makers expect. These expectations are formed in a political environment and are based on a less than complete understanding of the consequences of allocating scarce criminal justice resources to the incentives affecting individual choice and entrepreneurial behavior. Unanticipated consequences of drug enforcement policy also spill over into other criminal activities, further undermining the effectiveness of the criminal justice system.
During the 1984-89 war on drugs, police resources in most of the country were reallocated to combat drug market activities. The relative effort against drug crime is reflected in drug arrests as a proportion of all arrests, which is shown for each of the states in Table 4. Drug arrests on average accounted for 5.8 percent of all arrests in 1984 and 9.7 percent in 1989, an increase of 67 percent. A cursory glance at the data suggests great variability, however. If the states are ranked in order of their increase in relative drug enforcement effort, the range is from a decline of 45 percent in South Dakota to a 268 percent increase in Montana. The states with relatively large increases in drug enforcement efforts, such as New Jersey (55 percent), New York and California (66 percent), Florida and Ohio (67 percent), Michigan (72 percent), Pennsylvania (95 percent), and Illinois (156 percent), account for more than half the nations population. The 198489 period saw increasing emphasis on drug enforcement in most states, a tendency that was most pronounced in the biggest states.
Drug arrests as a proportion of total arrests for Kansas and the United States for 198492 are shown in Figure 3. Kansas went against the trend during the 198489 drug war, reducing its drug-arrest/total-arrest ratio by 21.4 percent. By 1989, drug arrests accounted for only 5.5 percent of total arrests, compared to 9.7 percent for the nation. As noted above, however, since 1989, Kansas appears to be getting relatively more active in drug law enforcement, while the rest of the nation is generally moving in the opposite direction. Drug arrests as a percentage of total arrests fell 25.8 percent in the U.S. from 1989 to 1991, while this ratio fell only 12.7 percent in Kansas over the same period, to 4.8. Kansass drug-arrest/total-arrest ratio has not fallen since 1990, however. Indeed, it rose to 5.2 in 1992, an 8.3 percent increase over its 1990 level. The trend has continued, with another 17.1 percent increase in 1993, to 6.1. These data, coupled with the continually increasing drug arrest rate shown in Figure 1, suggest that Kansas may bear some unexpected costs from their increasing emphasis on drug law enforcement.
Table 4. Drug Arrests as a Percentage of Total Arrests, by State, 1984 and 1989
|State||1984||1989||% Change||State||1984||1989||% Change|
|D. Columbia||19.6||20.9||6.6||North Dakota||4.1||3.4||17.1|
Source: U.S. Department of Justice, Bureau of Justice Statistics (1984 and 1989).
Reallocating police resources toward drug law enforcement diverts resources from fighting nondrug crimes. If, as popularly believed, the population of drug users was a subset of individuals who commit crimes against people and property, the arrest of a drug offender would simultaneously remove a nondrug offender from the street. This view is flawed, however, as shown in Section III above: the population of drug users only partially overlaps that of Index I offenders.
Figure 3. Drug Arrests/Total Arrests for Kansas and the United States, 1984-1992
Source: Uniform Crime Reports (various years).
Drug law enforcement can affect other crime rates because law enforcement resources are scarce. For instance, police departments must allocate their resources among competing uses, ranging from Officer Friendly programs and traffic control to the solution of robberies and murders. One thing is clear: when police resources are used for one purpose, they are not available for another.
Drug Law Enforcement and Property Crime
Increasing efforts against drugs can be accomplished by new police resources, of course, but in most jurisdictions, some of the required resources come from a reallocation of existing police resources. Kansas may be unusual in the sense that it increased drug arrests throughout the 198493 period but managed to increase nondrug arrests even faster until 1990. However, since 1991, it appears that police resources have been reallocated toward drug arrests. Kleiman (1992:153) notes that much of the increase in local drug enforcement during the 1980s came at the expense of other law enforcement efforts.... As a result, certain kinds of property crime are treated as unworthy of investigation or prosecution. This may have occurred relatively late in Kansas, but the consequences of these reallocations can be predicted by looking at what happened elsewhere.
Empirical studies using Florida data reveal the inevitable trade-off: increasing police effort against drug crimes relative to the effort against Index I crimes results in a lower probability of arrest for property crimes. Studies estimate that a 1 percent increase in drug law enforcement in Florida relative to Index I enforcement, as measured by arrests, leads to an approximately 0.20 percent to 0.34 percent decrease in the probability of arrest for property crime. Shifting police resources to drug enforcement results in more property crime for two reasons: (1) the lower probability of arrest means that the existing stock of property criminals will commit more offenses before being apprehended, and (2) the lower probability of arrest is a decline in deterrence that may stimulate property crime among individuals who were not previously engaged in this activity.
Primarily as a consequence of this reduced probability of arrest, the property crime rate in Florida rose 16.3 percent, from 6,892.4 offenses per 100,000 population in 1983 to 8,019.1 in 1989. Since 1989, Florida has reduced its drug enforcement efforts, and its property crime rate has fallen.
Drug Law Enforcement and Violent Crime
Increasing violence is often cited as a direct consequence of increasing drug law enforcement. Kleiman (1992:20) argues that enforcement takes its greatest toll on relatively benign drug dealers, leaving the trade with better armed and more violent organizations. Independent of this potential selective enforcement effect on violence, viewing drug market competition in its geographic context suggests that the drug war may generate unintended consequences with respect to violent crime just as these effects have been shown for property crime. In particular, drug dealers relocate some of their operations in response to differential policing. For instance, the Florida police department in Tampa, like most large city departments, formed a special drug task force to shut down street dealing locations. The result, according to Kennedy (1993:4), was as follows:
The task force would typically shut down one spot, only to find the same dealers in business around the corner shortly afterward or dispersed to several new locations....It was all short term, [Police Captain] Sollazzo says, the problem in fact escalated and spread throughout the community. The task force apparently made things hot enough in predominantly black neighborhoods that dealers, for the first time, moved heavily into more affluent white parts of town.
But not all of the movement was into previously untapped markets. Tampa also experienced violent battles over turf (Kennedy 1993:8). Such results are commonplace. Indeed, along with pushing the drug problem from one neighborhood to another and producing violent confrontations among drug dealers, these battles also increased the number of dead and wounded police officers (Stutmann and Esposito 1992:70).
Relocation, and the resulting entry into an established market in a neighboring jurisdiction, disrupts the local drug market. In order to establish a niche in this geographic area, the new entrants must necessarily tread on the turf of existing sellers. As competition intensifies and predatory practices are employed to establish market share, the probability of violent confrontations increases. Therefore, violent crime will increase in one jurisdiction as a consequence of more intense drug law enforcement in neighboring jurisdictions.
Consumers also face a higher risk of arrest and conviction in the jurisdiction that is getting tougher on drugs. They have an incentive to buy their drugs in locations with relatively less diligent enforcement. The result is an increase in demand in nearby areas that reinforces the rising supply of drugs, increasing the size of the drug market in the jurisdiction that allocates relatively few police resources to drug enforcement. Neighborhoods in which drugs are marketed experience a relatively high rate of violent robbery because drug users and sellers are carrying either cash or drugs, and when victimized they are not prone to report the theft to the police (Goldstein 1989:35). This makes drug market participants attractive targets for robbery. These robberies are likely to become crime statistics only if they involve sufficient violence to require medical treatment, in which case the incident will probably be characterized as an assault. Thus, a growing drug market is likely to generate an increase in violent crime independent of that caused by competitors fighting over market share.
Are There Similar Trade-offs in Kansas?
An examination of crime rates in Kansas compared to the states trends in drug arrests suggests that some of the same trade-offs may be relevant that have been revealed in more detailed studies of other states, but with differences in timing. Kansas reduced its drug-arrest/total-arrest ratio through the 1980s, and its crime rate in 1989 was lower than its crime rate in 1981. In fact, over this eight-year period, Kansass crime rate rose and fell four times, in contrast to other states such as Florida, which witnessed steady increases in its crime rates once it entered the drug war. However, the Florida crime rate fell 5.3 percent from the 1989 peak of 8,755.9 to 8,289.0 in 1992, as drug arrests were reduced relative to total arrests and police have backed off from the drug war. In contrast, Kansass 1992 crime rate of 5,319.9 is 6.8 percent higher than its 1989 rate of 4,982.8, perhaps reflecting Kansass reallocation of police resources toward drug control over this period. Of course, many factors influence crime rates, so the actual trade-offs in Kansas cannot be determined without a more detailed study, but there is no doubt that law enforcement resources used to combat drug use must come at the cost of reducing other police services. Both property crime and violent crime increase as a direct consequence of the policing efforts to control drugs. This reflects the fundamental economic facts that resources used for one purpose cannot be used for another, and that criminals respond to the incentives that they face.
Trade-offs arise in the allocation of all criminal justice resources, not just police. A rising tide of drug convictions swamped many state prison systems, which caused the adoption of early-release and other programs and diminished the severity of punishment. Overcrowding of Florida prisons provides one of the most dramatic examples of how rising drug law enforcement can compromise the punishment of other criminals. Incarceration of drug offenders rose rapidly during the 198489 war on drugs. During fiscal year 198384 there were 1,620 admissions to Floridas prisons for drug offenses, accounting for 12.9 percent of all admissions. By 198990 this figure had risen to 15,802 admissions, or 36.4 percent of the total. Getting tough on drug offenders resulted in leniency for others. Prior to 1984, prisoners in Florida typically served about 50 percent of their sentences. In 1987, Florida began addressing acute prison overcrowding by increasing gain time awards. At the end of 1989, the average prisoner served only 33 percent of his or her sentence; in fact, about 37 percent of the prisoners released in December 1989 had served less than 25 percent of their sentences, and some served less than 15 percent.
Statistical studies of the impact of early release on overall crime rates have not been performed, but to the degree that severity of expected punishment is a deterrent, such early-release programs should raise crime rates. In addition, there are growing numbers of examples of violent felons being released early and committing more violent crimes. Political reactions to the consequences of early release and rising crime rates have forced the Florida criminal justice system to adjust its treatment of drug offenders. There were 2,304 drug admissions to Florida prisons during the last quarter of 1992, down from 4,014 in the second quarter of 1989. The Florida legislature also revoked or reduced several mandatory minimum sentences for drug offenders. Drug admissions still accounted for 25 percent of total admissions in 1993, but this reduction, along with reform of mandatory sentencing and some prison construction, was sufficient to increase the portion of sentences served back up to 43 percent.
Kansass pattern of drug admissions and prison crowding is in some ways quite comparable to Floridas, although the timing is somewhat different. The percentage of total court admis-sions in Kansas with a drug offense as the primary offense fell from 10.5 in 1985 to 9.3 in 1986, but then rose steadily until 1990, when it reached 24 percent. Over this period, Kansas prisoners were also serving increasingly long periods in prison. The average number of months from admission to parole rose from 18.2 in 1984 to 28.3 in 1989. With the growth in drug admissions and months served by prisoners, Kansass prison population exceeded capacity between 1984 and 1990. Kansas apparently attempted to build its way out of this crowding with a 91.6 percent increase in prison capacity (from 2,953 in 1984 to 5,657 in 1989). In addition, parole board decisions to grant parole as a percentage of total decisions rose from 43 percent in 1987 to 56 percent in 1990 and 58 percent in 1991. However, the additions to capacity and the increased parole rate were not suf-ficient to avoid overcrowding due to the combination of a growing prison population (which increased by 53 percent, from 4,033 to 6,172) and an increase in the average time served by each prisoner (which rose by 55.5 percent, from 18.2 months to 28.3 months). As a result legislated changes in good time and parole eligibility determination were instituted in 1988 and 1989 in order to reduce the prison population (as indicated below, this means that the increased parole rate was occurring during a period when offenders were becoming eligible for parole after serving smaller portions of their sentences).
In 1988, the rate at which good time could be earned was increased from its previous level of one day credit for every three days served plus 30 days annually if an entire year was served. The result was that parole eligibility could be achieved after serving half the minimum sentence plus six months. The change was made retroactively to affect all offenses committed prior to July 1988 as well (class A felons still had to serve 15 years, however). Then in 1989, the good-time rate was changed again, allowing day-to-day credit and applying it retroactively; that is, Kansas prison inmates could become eligible for parole after serving one-half of their minimum sentences. With these changes, both the total prison population and the average stay in prison for Kansas felons began to fall. The population dropped to 5,677 in 1990 and 5,619 in 1991, when population finally fell below capacity. The average number of months served before parole dropped to 25.1 in 1990, 21.5 in 1991, 20.3 in 1992, and 19.1 in 1993.
The year 1991 also witnessed a reduction in the percentage of total court admissions accounted for by drug offenses to 21.5, but after 1991 drug admissions as a portion of court admission rose, reaching 25.8 percent in 1993 and 28.4 percent thus far in 1994. Therefore, part of the reason behind Kansass increases in good-time awards to alleviate prison crowding has apparently been the almost steady increase in drug admissions since 1986.
There is no magic policy bullet to kill the drug problem. Drug use has been a continuous feature of American life, and is best viewed as a chronic condition (Zimring and Hawkins 1992:191). Furthermore, the drug problem inevitably changes over time, both in terms of the substances consumed and the social costs associated with their use. There are a number of policy recommendations, however, that clearly can be made in light of the reality of scarce law enforcement resources and the absence of a strong causal relationship between drugs and crime.
Recall that several studies of the temporal sequencing of drug use and nondrug crime among people who commit both types of offenses suggest that for the majority, nondrug crime occurs before drug use starts. And this is not simply a matter of hours or days in which a person desiring drugs can accumulate money to make a buy. The implication is that as individuals, especially juveniles, get involved in the subculture of crime, they find that the criminal justice system actually does not appear to pose any real threat to them. Juveniles, in particular, are apparently very myopic. They consider only the immediate costs and benefits of their actions. When they learn that there are no real costs associated with committing crimes, even when they are arrested (because as first-time and even multiple-time offenders they get off with no real punishment) they tend to focus only on the immediate benefits of the crime. Many also eventually begin to experiment with drugs.
A policy of punishing youthful offenders for their first or second offense might teach them that they will have to pay for their crimes. This might not only divert them from further crime, but also move them out of the crime subculture where they are likely to find their way into drug use. By letting young people commit several offenses before they are punished, a criminal cohort is created that has few options other than crime. Shifting criminal justice resources away from a focus on drugs toward a focus on nondrug crime, including early intervention and treatment of such crimes (perhaps traditional punishment, but perhaps alternatives such as work programs to earn money for mandated restitution to victims, as well as efforts to enhance literacy), could actually prevent the development both of more serious criminal activities and of drug use among such criminals.
The number of juveniles under custody is one measure of the punishment youthful offenders face. In this regard, Kansas follows a strategy that results in a relatively high level of youthful incarceration. Figure 4 shows that Kansas had 329 delinquent offenders per 100,000 juveniles in custody in 1989, compared to a national average of 259a 27 percent differential. Kansas is also more likely to have youthful status offenders in custody58 per 100,000 juveniles in Kansas compared to 36 for the nation. If the old adage a stitch in time saves nine is relevant to the criminal justice system, it would appear that the relatively more custodial resources Kansas spends on youthful offenders are relatively well spent.
Figure 4. Juveniles under Custody per 100,000 Juvenile Population in Kansas and the United States, 1989
Source: Sourcebook of Criminal Justice Statistics, 1992, table 6.25.
Resources to combat drug crime are limited, so the cost effectiveness of law enforcement efforts relative to drug treatment is an important policy issue. A recent RAND study concluded that a significant portion of current expenditures on drug abatement would be better spent on treatment than on law enforcement (Rydell and Everingham 1994). The RAND study estimates that an additional $34 million expenditure on treatment would reduce U.S. cocaine consumption by 1 percent, while a similar reduction would be gained by a $246 million expenditure on domestic law enforcement. Thus, treatment is estimated to combat cocaine use about seven times more effectively than is law enforcement.
Treatment is relatively more effective because it cuts consumption directly, while law enforcement works by raising the price of drug offenses. Cutting drug use may reduce other crime directly to the degree that some property and violent offenders commit more crimes when their drug consumption is highest. Putting these users in supervised treatment can reduce their criminal activity while in treatment, and if treatment helps reduce their dependence on illicit drugs, a long-term effect may also arise. Perhaps just as important is the fact that treatment reduces the unintended consequences of enforcement. Raising the cost of using cocaine through more intensive law enforcement efforts may cause users to substitute some other illicit substance, with the result being, for example, higher heroin use. Since the demand for mind-altering substances is not quenched by enforcement, users have an incentive to seek lower-cost alternatives to the substances that are most effectively discouraged by police efforts. In contrast, cocaine consumption is cut during the treatment process while at the same time undermining the demand for substitute drugs.
The RAND study does not suggest that law enforcement efforts necessarily should be abandoned in favor of treatment, however. Many heavy users may refuse treatment, and law enforcement can provide an effective inducement to seek treatment. About 30 percent of law enforcement dollars spent on U.S. drug control could profitably be shifted to treatment, according to RAND estimates. Because drug problems vary greatly among jurisdictions, this specific estimate should not be uncriti-cally applied in Kansas. An increasing role for treatment alternatives in Kansas drug policy should nonetheless be explored.
A promising way of combining the cost effectiveness of treatment with the compulsion of the criminal justice system is the innovative Drug Court in Dade County, Florida. This program is described by Finn and Newlyn (1993) and was evaluated by the Crime and Justice Research Institute (Goldkamp and Weiland 1993). The Drug Court was founded in 1989 by Floridas Eleventh Judicial Circuit in order to reduce the circuits rising caseload generated by the large volume of drug offenders. This was to be accomplished by providing a range of services and support to shut the revolving door through which drug offenders repeatedly encounter the criminal justice system. The premise behind this program is that investing a year of comprehensive treatment coupled with close surveillance in these casesinstead of a few hourswould pay off in the long run with reduced costs to the police, courts, and jail as more and more drug users kicked the habit (Finn and Newlyn 1993:11). Treatment and intensive support are designed to move defendants through detoxification to eventual job placement.
Three fundamental aspects of the Drug Court are designed to facilitate its goals. First, the traditional adversarial relations of the courtroom are put aside, and all players in the system are theoretically on the same side: that of the defendant in pursuit of a drug- and crime-free life. Defendants talk directly with the judge. The public defenders role is not to exact a better deal for the defendant; like all the others, it is to keep the defendant out of the courts and off of drugs. Second, no program participant can be dismissed from treatment without the judges approval. This means that program staff cannot lighten their work loads by merely dumping recidivists back into the courts and prisons. Finally, the program expects intermittent lapses as defendants return to drug use, and gives them multiple chances. The programs philosophy seems to be do what it takes rather than a specific course of treatment to be followed by all participants with failures being punished by a prison sentence. Lapses are regarded not as program failures, but as part of the normal detoxification process. Indeed, two-week stints in jail (motivational jail) may be required, often more than once, when the offender is unable to stay drug-free in the first phase of the program. Since the usual practice outside the Drug Court was to impose no jail time for first offenders, this practice tends to increase short-term costs in anticipation of long-term savings.
It was suggested above that the undesirable consequences of a drug war are quite significant. The question becomes: Why do law enforcement policy makers wage such wars? One inducement to wage a drug war can be traced to taxpayers reluctance to fund law enforcement agencies at a level that allows them to pursue all types of crime. This has turned police into entrepreneurial agencies as they focus on the pursuit of asset seizures to expand their budgets. In this regard, the U.S. attorney general has argued that asset forfeiture is a valuable double-barreled weapon against crime. One weapon presumably is the deterrence effect of taking away ill-gotten gains; the second is the reinvestment of the proceeds into the law enforcement effort. Research reported by Benson, Rasmussen, and Sollars (1995); Rasmussen, Benson, and Mast (1994); and Rasmussen and Benson (1994, chapter 6) suggests a very different interpretation, however. Police agencies have changed their priorities because of the incentives created by federal and state changes in asset seizure laws, resulting in increased drug law enforcement with a possible decline in public safety.
A section of the Comprehensive Crime Act of 1984 established a system whereby any local police bureau that cooperated with federal drug enforcement authorities in a drug investigation would share in the money and/or property confiscated as part of that investigation. As a result, police in many states whose own laws or constitutions limited confiscation possibilities began to circumvent state laws by having federal authorities adopt their seizures. Thus, under the 1984 federal statute, a substantial percentage of these seized properties went to the agencies that seized them, even if the states laws mandated that confiscations go someplace other than to law enforcement. Benson, Rasmussen, and Sollars (1995) contend that this legisla-tion was the primary stimulus for the nationwide upsurge in drug control effort that began in 1984.
Since there is significant variation in drug enforcement activity across states and cities, as well as through time, it follows that if this argument actually provides a strong explanation of drug enforcement policy, it should help explain cross-sectional variation in enforcement policy as well as time series variation. Since federally adopted seizures are only partially turned over to the local police (the federal authorities extract a 20 percent handling charge), police in states whose own laws allow them to retain seized assets are able to obtain even greater benefits from seizures than police who must involve federal authorities in the process. Rasmussen, Benson, and Mast (1994) explored this issue using a statistical model of the demand for and supply of drug law enforcement in a sample of large U.S. cities. Included in the model are variables that control for the extent of drug use, the opportunity cost of police resources, and socioeconomic factors affecting the demand for drug law enforcement. They report that the level of drug use, as measured by the percentage of arrestees testing positive for any illicit drug use, is a highly significant determinant of drug arrests. More important, they find that a state law that allows the police to keep any portion of seized assets raises the drug-arrest/total-arrest ratio between 35 and 50 percent. Allowing police to profit from the confiscation of assets from alleged drug offenders apparently provides a powerful incentive to law enforcement agencies which, as expected, changes agency behavior.
In light of this evidence, the inclusion of asset forfeiture as a desirable strategy in a U.S. Department of Justice (1992) report titled Combating Violent Crime was inappropriate. The incentives created by asset forfeitures do not encourage law enforcement agencies to focus on violent crime, but rather, they focus relatively more effort against drug markets because of the large amounts of money involved in drug markets, undermining efforts to fight crime against persons in two ways. First, as police seek revenues via confiscations, leading to a reallocation of policing effort to-ward drug crimes relative to other crimes, the impacts described in Section V above are the inevitable consequence. Second, since juveniles have few assets, efforts to expand seizures lead to a focus on adults in the drug market, producing adult drug arrests (which are relatively easily prosecuted) and contributing to the prison overcrowding that has resulted in the early release of some violent criminals in many states.
Rather than being a double-barreled weapon against crime, asset forfeiture might at best be described as a single-barreled weapon with a very strong kick. Deterrence effects of forfeitures may be strong (although evidence of even this effect is not available), and thus a significant weapon against drug offenses if offenders believe that the probability of a seizure occurring is significant. The kick from the asset forfeiture weapon, at least when police are allowed to keep the proceeds, is that enforcement efforts are distorted toward drug offenses, away from crimes in which some person is, at least potentially, explicitly victimized.
As of 1991, Kansas law dictated that seized assets go to the law enforcement trust fund in the political subdivision (e.g., state, county, local) whose agency made the seizure. One partial solution to the apparent reallocation of policing resources in Kansas toward a growing emphasis on drug crime is straightforward, and still on the books in several other states: change the seizure law to mandate that proceeds must go to general revenue rather than to the arresting agency. This retains the deterrent weapon while eliminating the distorting effect of asset forfeiture on the allocation of police resources among crimes (a similar reform is needed for the federal enforcement agencies as well, and the federal adoption of state and local seizures should be stopped). Of course, local budget-setting authorities may want to reward police for their efforts to enhance general revenues through seizures, but at least police will have to compete for the revenue by convincing an oversight authority that their focus on seizures is an appropriate one; and if they cannot do so, they have incentives to redirect their efforts.
Public Safety versus Arrests and Seizures
The nature of bargaining over police budgets in the public sector influences the allocation of police resources. For police, an appeal of keeping the proceeds of seized assets is that discretionary budgets can be enhanced without having to justify their need in the face of competing demands from other community agencies. However, putting seizures into the general fund so that police must compete for them does not solve all of the problems of police resource allocation. Statistical studies of police budgets suggest that budgets rise with higher measures of productivity (as measured by outputs such as arrests and seizures) as well as with increased need (as measured by crime rates).
Budget bargaining over these measures of productivity and need has unfortunate consequences. For example, police agencies tend to be better off from a budget perspective if they arrest offenders after crimes are committed, rather than defending communities by preventing crime from occurring in the first place (Sherman 1983). Arrests for possession of small quantities of marijuana may be as effective in these negotiations as arrests for a burglary, but searching for a burglar may be far more time consuming than finding drug offenders in open-air markets. The best police strategy to compete for public dollars might be to make as many arrests as possible and make sure that highly publicized crimes are solved. Indeed, by focusing on drug enforcement, both measures of output (e.g., arrests, drugs seized, assets seized) and measures of need (crime rates) can rise, as explained in Section V, because with more police resources devoted to drug control, other crimes are less effectively deterred. Therefore, the budget-setting process itself also perverts police incentives.
Changing the budget negotiating process at the local level requires political action at the local level. A focus on local patrol (perhaps by officers on foot) and community involvementto prevent crime rather than waiting to respond and arrestrequires that the budgetary evaluation of police performance change to include more qualitative evidence and less reliance on quantitative measures. Political action by communities that demand safer streets can encourage police to alter their strate-gies, concentrating their efforts to ensure public safety and putting a somewhat lower priority on making arrests and seizures, or fighting drugs per se. For instance, Tampas drug problem had become so bad by the end of the 1980safter several years of drug warthat community leaders began to put increasing pressure on the equally frustrated police. The result was an innovative program that was focused on community involvement for the purpose of making neighborhoods safer. Police officials knew that drug use might not be significantly changed by such a program. Indeed, they expected that the best that might happen to the drug trade would be that it moved indoors. That seemed, nonetheless, a worthy goal (Kennedy 1993:18).
Legislation of Sentencing Policy
Legislators at the federal and state levels often are enthusiastic supporters of legislation that symbolically gets tough on crime, while being understandably reluctant to finance more criminal justice resources through higher taxes. Their abuse of the system comes in many forms. Legislators aversion to having constituents pay for tougher enforcement raises the appeal of the asset forfeiture legislation just discussed, for instance, because it appears to make incremental police activities self-financing. The actual results, however, are greater police efforts against drugs relative to other crimes, with the likely consequence of making the community less safe. Legislatures have further abused the criminal justice system by passing mandatory minimum sentences for drug crimes, making it easy for prosecutors to get convictions for lesser crimes via plea or charge bargaining. The result is that more drug offenders are easily prosecuted, producing prison overcrowding and, in many states, the early release of relatively dangerous offenders.
Legislatures actually have the power to codify, by statute, a mechanism that would move the allocation of scarce prison space toward its highest and best use. Sentencing guidelines are well suited to this task, in theory at least, although in fact they emerged in the United States after 1975 largely in an attempt to limit judicial discretion, and later alterations have reflected symbolic actions to demonstrate get-tough approaches to crime. If guidelines were to be used as a mechanism for allocating prison space, they would have been written in a way that determined the relative seriousness of crimes and criminal histories in light of prison capacity restrictions. If guidelines are promulgated without reference to prison capacity, they become a symbolic wish list, unrelated to the reality of punishment. The result is prison crowding, so desired sentences are subverted by early release and other methods of controlling the prison population.
No state has had the political will to provide such a system. Minnesota, a state with admirable intentions in this regard, set up its guidelines to allow for a constant prison population, although they could adjust over time with population growth. Tonry (1988) reports that Minnesotas matching of penalties with prison resources recognized the inevitable trade-offs in incarceration policy, but the original intent was undermined by the desire of subsequent legislatures to increase penalties for drug and sex offenders without regard to prison capacity. This is not surprising. Legislators can avoid the impasse in corrections by being able to promise constituents tougher penalties for some crimes without having to confront voters with the stark reality: pay more taxes to build additional prisons, give up other public services, or abuse the prison system, which could include the early release of some prisoners.
Kansas instituted a guidelines system in 1993. How successful it will be can be determined only by observing legislative changes in the guidelines and their relationship with prison capacity in the future. But there is an alternative solution. Rather than attempting to develop and maintain a statewide mechanism for rationing the states prisons, the allocation of prison space might be localized by making prosecutors and/or judges responsible for rationing scarce prison resources.
Prosecutors and judges have what is perhaps the clearest incentive to abuse the prison system. Consider an example. A person is charged with writing a bad check. The prosecutor, seeing that the person has had a lengthy record of such offenses, decides to prosecute. The judge recognizes that this offense is not very serious, but is so frustrated with the continued unlawful behavior that the offender is sent to prison. Because the state prison system is under court order to limit its prison population, an early-release mechanism releases an inmate in order to make room for the bad-check writer. A violent criminal who has served a substantial period of time is released, but to a county in another part of the state. Neither the prosecutor nor the judge has to consider the consequences of the decision in this case because another community is most likely to suffer the consequences. This example, unfortunately, closely approximates an actual series of events in Florida involving Charlie Street. Within several days of his release in 1988, he killed two policemen.
Would the prosecutor and the judge have behaved differently if they knew that the person to be released would be returned to their own community, where the people who voted them into office would be directly affected? The answer would surely be yes. What prosecutor would like to seek reelection when the opponent could argue that the incumbent is willing to let out a murderer to punish a bad-check writer? Judges would likewise consider that bad press. Stopping the abuse of the prison system requires that prosecutors and judges consider the consequences of their actions: what are the costs of incarceration and who might be let out early when a person is convicted to a prison term?
Zero tolerance may be appealing politically, but it is an unattainable goal under any conceivable condition. Furthermore, at the margin at least, drug use itself is apparently more effectively reduced through treatment than through law enforcement (Rydell and Everingham 1994). To the degree that making the possession and consumption of drugs felonies diverts users from treatment programs into the prison system, decriminalization clearly should be considered as a method of decreasing drug use. Decriminalization would have the added benefit of freeing up scarce criminal justice resources in order to focus on violent and property crimes.
The inevitable reactions of drug entrepreneurs to increased enforcement suggest that at the very least, criminal justice drug enforcement efforts might best be aimed at the drugs that are most objectionable. Trends in price and use suggest that law enforcement has had considerable effect on the marijuana market over the past decade, but this very success probably led to increases in both the demand for and supply of cocaine (and alcohol). The opposite can also be expected: reducing the probability of arrest and/or severity of punishment for marijuana should lead to reductions in the use of other drugs. Indeed, Model (1993) found that decriminalization of marijuana, which had occurred in some states during the 1970s, was associated with lower numbers of emergency room episodes related to other kinds of illegal drugs. The available evidence from decriminalization in these states and in parts of Europe suggests that there are few adverse consequences that would accompany an experiment with decriminalizing marijuana.
Getting tough on drugs has had the inevitable consequence of getting soft on nondrug crimes. But because of the nature of criminal justice resources, part of the costs of a jurisdictions drug war are shifted onto citizens in other jurisdictions. In this political and institutional environment, suggestions like those made above are likely to fall on deaf ears. The fact is that effective reform of drug policy is likely to require fundamental institutional reforms. Indeed, the current trend toward centralization of criminal justice, with more laws and funds coming from Washington, and from state capitals, will have to be reversed.
Criminal justice problems in general and drug problems in particular are ubiquitous, but they vary considerably across the nation, and the impacts of many specific problems are, in fact, highly localized. Therefore, state-level decisions on criminal justice are likely to conform to the characteristics of a states crime problems much more accurately than are congressional mandates. A highly urbanized state like New York or New Jersey has a very different set of problems to deal with than a very rural state like North Dakota or Wyoming. A state like Kansas, with both large rural areas and significant urban development, faces yet a different mix of problems. Why should we expect that a homogeneous policy emanating from Washington, D.C. would suit any of these kinds of states very well? And state-level actions may be undesirable for the same reasons that federal actions are undesirable. The solution lies not in state policy per se, but in decentralization: increased reliance on local government.
A policy of localization means that the officials who are elected locally are the ones who make both the taxing and the spending decisions that affect local citizens. Clearly, writing a letter to a county commissioner or a city council member is far more likely to stimulate a real response than writing to a state legislator or a member of Congress. The reason is that local officials can be held accountable for their actions, at least much more effectively than Congress or state legislators can. In this light, a policy that forces a local court to decide which of its prisoners to release when another prisoner is added to a crowded prison will induce the local judge and prosecutor to recognize the consequences of their decisions. They will no longer have incentives to simply maximize the number of convicts that they can ship off to a shared state prison. Similarly, a police department that gets both its funding and its mandates regarding policing priorities exclusively through local channels is more likely to be responsive to the wishes of local citizens.
The fact is that it is difficult to determine whether a policy is effective or not when the costs of some resources are localized while the costs of others are dispersed and involve common-access resources. Thus, experimentation is limited and there are fewer alternatives against which to compare any one policy. Decentralization means that alternative policies are much more likely to be tried, so information about the consequences of various policy options is more likely to be available. If experiments with treatment alternatives to prison tend to reduce drug and crime problems more effectively than a lock-them-up approach, for instance, then given that the lock-them-up jurisdiction cannot export the costs of its approach, citizens are much more likely to find out about successful polices and demand similar ones.
Today, many states face the legacy of the 198489 drug war that continues to plague their criminal justice systems: mandatory sentences for drug crimes, federal and state seizure laws, and other legislative actions produced in an effort to appear tough on drugs mean that drug control activity is not likely to return to its pre-1984 level, let alone to a level that might be an appropriate reflection of the true costs and benefits of drug enforcement. Indeed, unlike most states, Kansas has continued to increase its drug control efforts since the 1980s, so the states costs are probably continuing to mount. Significant changes in the institutionalized incentives facing criminal justice decision makers, from legislators down to patrol officers, are required to rationalize the system.
 Kansas appears to be different from the average in other ways as well. For instance, unlike in most states, drug enforcement in Kansas is not particularly concentrated in larger cities: drug arrest rates in Topeka (229) and Wichita (217) were below the state average in 1989, while that of Kansas City (430) was higher.
 Drug arrests in Kansas rose 8.7 percent in 1993, raising the rate to an estimated 328 per 100,000 population. Comparable national data are not available at this time.
 This section draws from Rasmussen and Benson (1994, chapter 3).
 There is a large academic literature that addresses the issue of a drugcrime causal connection. This literature does not suggest the simple connection between high rates of drug use and criminal activity that many policy makers perceive. See Wilson and Herrnstein (1985, chapter 14); Gottfredson and Hirschi (1990); Chaiken and Chaiken (1990); and Rasmussen and Benson (1994, chapter 3) for overviews on the literature.
 See Rasmussen and Benson (1994, 5865) for details regarding studies of recidivism and of arrest histories of Florida drug criminals that strongly support the general implications of this national study.
 The suggestion that an addicts demand may not be as fixed as is generally believed is reinforced by the fact that there are obvious substitutes for drug consumption. For the most severely addicted persons, treatment programs and drugs that block withdrawal are available and are used when heroin prices rise (Gould et al. 1974; DuPont and Greene 1972; Levine, Stoloff, and Spruill 1976; Moore 1977; Silberman 1978; and White and Luksetich 1983). In addition, research suggests that most drug users stop drug consumption eventually, whether they use a treatment program or not to facilitate the change in behavior (Vaillant 1969; Silberman 1978; Baridon 1976; and Gould 1990:34). The suggestion is further reinforced by the fact that drug consumers respond to the incentives created by law enforcement, as explained below.
 There is a large literature on the economic analysis of crime that attempts to identify variables that affect the costs and returns to criminal activity in order to predict the incidence of crimethe literature stems from a pathbreaking article by Becker (1968). The underlying rationale for this approach is that criminals respond consistently to incentives. In the economists view, people are less likely to commit a crime when society raises the expected price of committing it. The probability of being caught, the likelihood of conviction, and the severity of sentencing are components of price that can be directly influenced by public policy.
After reviewing a great number of studies, using different methods and varied data, Rasmussen and Benson (1994:4143) concluded, as had early surveys of empirical literature by Elliot (1977) and Silver (1974), the following: (1) Relatively poor legal income opportunities appear to result in more crime. (This appears to hold for drug dealers as well. Myers  used a sample of California, Michigan, and Texas jail and prison inmates who were self-admitted drug dealers and found that the dominant factor contributing to entry into drug selling, particularly among black males, was unattractive market opportunities. Furthermore, racial differences in returns to legitimate employment explained most of the gap between blacks and whites drug-dealing activities.) (2) Potential criminals are apparently deterred by higher probabilities of punishment. (As noted below, Kim, et al.  support this conclusion for drug criminals as well.) (3) The severity of punishment also may deter crime, but the evidence for this effect is not as convincing as that for higher probabilities of punishment or for the impact of such factors as unemployment.
 Gill and Michaels (1990) and Kaestner (1991) use the National Longitudinal Survey of the Work Experience of Youth in their analyses of the relationship between wages and drug consumption.
 See Innes (1988). Among those inmates who had used drugs at all in the month prior to the offense, 35 percent had income from illegal sources, while 27.9 percent of those using drugs regularly in the past reported illegal income.
 While some drug addicts clearly appear to be involved in illegal activity, property crime does not seem to account for most of the illegal income generated by these addicts. They probably raise almost as much money through such crimes as prostitution, pimping, and drug sales as they do from property crimes (Kaplan 1983:54; Reuter, MacCoun, and Murphy 1990). Moore (1977) estimated that heroin addicts obtained about 45 percent of their funds for purchasing heroin from working within the drug distribution system, with more than a third of the remaining funds coming from legitimate income or victimless crimes. Similarly, Gould et al. (1974) reported that profit margins for drug sales exceeded 100 percent, and suggested that half or more of the money spent on heroin at the time could have been generated through the sale of that drug. In one study (Waldorf 1973:50), one-third of the addicts in the sample supported their habits exclusively by selling drugs, and 48 percent partially supported themselves in this fashion. In this instance over 80 percent of the addicts were involved in the drug trade. Among another sample of addicts (Baridon 1976:45), more than 60 percent sold drugs and most also worked at legitimate jobs. So the implication is that the impact of drugs on the level of any particular crime, e.g., burglary, robbery, drug sales is not theoretically predictable.
 See Reuter, MacCoun, and Murphy (1990); Gandossy et al. (1980); Greenberg and Alder (1974); Innes (1988); and Harlow (1991).
 This section draws from Rasmussen and Benson (1994; chapter 4).
 See note 7.
 Estimates of money prices are found in Moore 1990, and potency in Thornton (1991). Price data for illegal drugs are at best impressionistic, however, so these conclusions must be viewed with caution.
 Narcotics Control Digest (May 10, 1989); and U.S. Drug Enforcement Administration, Illegal Drug Price/Purity Report, United States 19871990, March 1991. THC content apparently did increase some in the 1980s, however, and has continued to since. The increase in THC is explained below.
 See Moore (1990) and Thornton (1991).
 Kleiman (1992:284) suggests that some marijuana growing shifted from Northern California to the mountains of Kentucky as a result of increasing enforcement. An unintended consequence of this change was to involve people whose cultural roots include moonshining and a history of violence, making the trade rougher than it had been before.
 Economist, December 15, 1990, p. 24.
 Another drug that is being substituted for marijuana is alcohol. Recent and as yet unpublished statistical studies by DiNardo and Lemieux (1992), Chaloupka and Laixuthai (1992), and Model (1991), as well as a published paper by Model (1993), have confirmed previous perceptions that alcohol and marijuana are substitutes. DiNardo and Lemieux (1992) found that a higher drinking age led to reduced alcohol consumption among high-school seniors between 1980 and 1989, but it also led to an almost one-for-one increase in marijuana consumption. Model examined the effect of marijuana decriminalization on hospital emergency room episodes (1993) and violent crime (1991), finding that decriminalization reduced episodes related to alcohol and other illicit drugs and reduced violent crime, noting that a higher percentage of violent crime is alcohol related. Chaloupka and Laixuthai (1992) found that either lower money prices of marijuana or reduced legal sanctions against marijuana possession and use led to a substitution of marijuana for alcohol and to a significant reduction in nonfatal and fatal accidents from driving under the influence among youths. Similarly, higher prices of beer and increased drinking age led to more marijuana consumption and a reduction in traffic accidents associated with driving under the influence. Clearly, then, one unintended consequence of marijuana illegality is a greater level of alcohol consumption by youths, along with greater levels of traffic accidents and fatalities.
 Note that crack is not a new technology in the sense of being a new discovery. The knowledge necessary for producing crack has existed for some time. Indeed, virtually all of the technological changes that occur in the drug market involve the implementation of existing technology rather than new technology (Thornton 1991).
 This subsection draws on Thornton (1991; chapter 4).
 See Thornton (1991:100105) for a detailed analysis of the effects of law enforcement on liquor potency during the Prohibition era. Also see Warburton (1932:14866) and Fisher (1927).
 There are other reasons to expect that drug control efforts will have relatively little long-term impact on overall drug use. See Rasmussen and Benson (1994, chapter 4) for details and references.
 This section draws from Rasmussen and Benson (1994, chapter 5). The focus is on the unintended consequences of the 198489 drug war that emanate from changes in the allocation of criminal justice resources. There are other important consequences of rising drug enforcement during the 1980s that, in the minds of some observers, may be more important than those examined here. These include constitutional issues relating to the relaxation of the standards for reasonable search and seizure and invasions of individual privacy. Legal scholars, who have a comparative advantage in examining these issues, have done so in detail. The importance they attach to the drug wars intrusion in these areas is captured in the titles of their articles: for example, The Incredible Shrinking Fourth Amendment (Wassertrom 1983); Another Victim of Illegal Narcotics: The Fourth Amendment (Saltzburg 1986); and Crackdown: The Emerging Drug Exception to the Bill of Rights (Wisotsky 1987). While these are important issues, we focus on other issues to which the tools of economics can be fruitfully applied. However, there are important areas of overlap between these constitutional issues and the allocation of criminal justice resources, notable examples being provisions for asset forfeiture and issues of excessive punishment (Rasmussen and Benson 1994, chapters 6 and 7). Broader constitutional issues are left to legal scholars. See Wisotsky (1992) for a summary of these issues.
 Several studies have examined how offenders respond to changes in enforcement efforts; it appears that a relatively strong policing effort against one crime type or in one police jurisdiction induces some existing criminals to shift to a different criminal activity or to commit crimes in another jurisdiction. For example, see Furlong and Mehay (1981); Hakim, Speigel, and Weinblatt (1984); and Sollars, Benson, and Rasmussen (1994).
 Benson and Rasmussen (1991); Benson et al. (1992); and Sollars, Benson, and Rasmussen (1994).
 See Kleiman (1992:62 and 140); Reuter (1991); and Miron (1992).
 These and other unintended consequences of the Tampa Drug Task Force approach led to considerable dissatisfaction on the part of both citizens and police and ultimately to a new experimental program that changed the focus of policing in the city. This program is discussed in Section VI below.
 While this discussion focuses on adjacent jurisdictions, it readily applies to different precincts within a city.
 Goldstein (1989) discusses three sources of systemic violence in drug markets: (1) that resulting from competition among sellers, (2) that committed during robberies of drug market participants, and (3) violent acts between drug users, resulting from disputes over drugs. The latter is not likely to be affected by changes in the spatial distribution of drug markets so long as drug users do not change their place of residence as a result of changing drug prices. Goldstein is skeptical of the notion that much violence is spawned by the pharmacological effects of drugs. He also reports that there is little research evidence supporting the proposition that some drug users are compelled by economic necessity to engage in violence-prone property crime (i.e., robbery).
 Using data from 279 jurisdictions in Florida, Rasmussen, Benson, and Sollars (1993) found, after controlling for other factors, that the violent crime rate in a community is significantly and positively correlated with the drug arrest rate in adjacent jurisdictions. Furthermore, the changes in violent crime in light of these enforcement differentials are much larger than the spillover effects commonly reported in studies of property crime (Sollars, Benson, and Rasmussen 1994; Mehay 1977; Kennett 1982; and McPheters and Stronge 1981). Police no doubt respond with greater enforcement when faced with a large violent drug market, but in Florida at least, greater police enforcement appears to generate increased competition in nearby drug markets, which in turn results in a higher violent crime rate.
 Increasing drug arrests do not necessarily come at the expense of Index I crime enforcement, of course. The decision as to what to sacrifice is at least partly up to police officials. A different trade-off apparently emerged in Illinoiss war on drugs (Benson and Rasmussen 1992). Drug arrests in Illinois accounted for 4.77 percent of all arrests in 1984, and rose to 7.01 percent in 1989, an increase of 47 percent. But Illinois did not sacrifice the control of Index I crime relative to what was being done prior to the drug war. Examination of arrest statistics suggests that the most dramatic reallocation of resources for increasing drug enforcement came from traffic control, as reflected by the DUI drunk driving arrest category. DUI arrests in Illinois fell from 53,038 in 1984 to 41,093 in 1989, a decrease of 22.5 percent. DUI arrests in the United States fell much less during the same period, from 1.347 million in 1984 to 1.333 million in 1989, a decline of less than 1 percent. The fall of DUI arrests in Illinois appears to represent a decline in general traffic enforcement relative to the nation. This may be a cause of the relatively faster growth of traffic fatalities in the state compared to the rest of the country. Nationally, traffic fatalities rose 0.8 percent between 1984 and 1989. Traffic fatalities in Illinois rose over 4 percent over the same period. (Nationally, about one-third of all automotive fatalities are caused by drunk drivers [Kleiman 1992:224].) Since the number of motor vehicle registrations and licensed drivers rose more slowly in Illinois than in the nation as a whole, this comparison probably understates the relative growth of fatalities in Illinois.
 There is yet another type of crime that has risen sharply due to criminal justice system efforts to control drugs: corruption among criminal justice officials. A 2,325 percent increase in federal convictions of public officials for corruption, from 44 in 1979 to 1,067 in 1988 (Thornton 1991), most of which is connected with the drug trade, suggests that public officials have considerable opportunities to serve the illicit drug industry. Convictions for corruption may represent only the tip of illegal behavior by public officials.
Corruption is most likely to occur when the potential payoff is high relative to the risks of detection and the opportunities for legal income. The illicit drug market apparently is the most lucrative source of police corruption that has ever existed in the United States, including the period of liquor prohibition. Exacerbating these risks is the undercover work that is typical of drug enforcement activities. Long-term association with criminal elements facilitates corruption, in part because of a gradual erosion of the officers value system, increased sympathy for the criminal elements, and added exposure to opportunities for illegal behavior (Girodo 1991). As Moore and Kleiman (1989:2) note, the police executive knows from bitter experience that in committing his force to attack drug trafficking and drug use, he risks corruption and abuse of authority. Informants and undercover operationsso essential to effective drug enforcementinevitably draw police officers into close, potentially corruption relationships with the offenders they are pledged to control. Risks of detection would appear to be modest since there are not strong institutional incentives to expose corruption. Indeed, exposing corruption undermines citizen respect and support for police. See Benson (1981, 1988a, 1988b, and 1990); and Benson and Baden (1985) for explanations of the incentives underlying corruption and for additional reasons to expect that public officials are very reluctant to report corruption among their ranks, and see Rasmussen and Benson (1994:10718) for details on police corruption in the context of drug policy.
 Rasmussen and Benson (1994, chapters 5 and 7) reported that in Florida it appears that the courts followed the police lead in the war on drugs and meted out relatively tougher sentences to drug offenders than to nondrug offenders during this same period.
 The information on drug admissions and total admissions, as well as on average stays and changes in the Kansas prison systems and changes in Kansas laws regarding determination of parole, was provided by Richard Koerner, Correctional Manager, Policy, Planning and Data Analysis, Kansas Department of Corrections, in response to a request for information by the authors made to Gary Stotts, Secretary, Kansas Department of Corrections.
 Between 1974 and 1979, the conditional release date of prisoners in Kansas was the maximum sentence for a crime minus good-time credits. This determination of conditional release remained in place until 1993, but as explained below, the rate of good-time credit accumulation has increased significantly. Parole eligibility was based on certification by the Department of Corrections. In 1979 there was a change in the determination of parole eligibility for Kansas inmates convicted of B or C felonies (parole eligibility for class A felons required serving 15 years, and for D and E felons, eligibility was based upon certification by the Department of Corrections). Good-time credits became the criteria: if all good time was earned, a class B or C felony inmate would be eligible for parole consideration after serving half of the minimum sentence for the crime plus six months except for an offense involving a gun. Additional changes in this law were made in 1981 and 1982. Good time became the determinant of parole eligibility for all felons in 1981 until determinantal sentences were imposed in 1993; class A felons still had to serve 15 years before parole eligibility, however. Then in 1982, good-time credits could be earned by all new felons, including mandatory gun felons, at one day of credit for every three days served plus an additional 30 days annually if an entire year was served (the 1979 and 1981 rules applied for offenses committed prior to July 1, 1982). No further changes in good-time accumulation rates were made until 1988.
 Kansas adopted a system of sentencing guidelines, or determinantal sentencing, in 1993. Inmates are released after serving their court-imposed sentences under the guidelines, although the sentences for all crimes except first-degree murder and treason, which are still eligible for parole after 15 years, can be reduced by up to 20 percent for good time. The amount by which a prisoner is released early by good-time credits is added to the period of postrelease supervision, however.
 U.S. Department of Justice (1992:22).
 Many states mandated that confiscated assets be turned over to a general government authority, while others required that some or all seized assets be used for specific purposes, such as drug treatment or education. Various states also limited the kinds of assets that can be seized. For instance, in 1984, only seven states allowed seizure of real estate used for illegal drug activities. The federal statute had no such limitation. Since 1984, several states have changed their asset seizure laws, making them more consistent with the federal law.
 There are additional reasons to rethink seizure laws as well, such as constitutional questions of due process: seizures are treated as civil issues rather than criminal issues, and citizens wrongly victimized by seizures must prove their innocence before they can get their assets back. We leave constitutional issues to legal scholars; for example, see Reed (1992) in this regard.
 See Benson et al. (1992); Benson, Rasmussen, and Sollars forthcoming; and Rasmussen and Benson (1994, chapter 6).
 Lawrence (1991) describes the breakdown of the Minnesota system. Florida, another innovator in using guidelines, created a system that ignored prison capacity. Over time, guidelines were eroded in a variety of ways, including the legislative approval of mandatory minimum sentences and habitual-offenders statutes. See Rasmussen, Benson, and Kim (1991) for an evaluation of the impact of these policy innovations on the allocation of prison space.
 Similarly, Frank Potts was released from the Florida prison system in 1988, after serving 6 years of a 15-year sentence for molesting an 11-year-old girl, despite a parole examiners report suggesting that there was a very high probability of recidivism if Potts were to be released. He is now being held in Alabama on charges of molesting another 11-year-old girl; in addition, an intense investigation is under way regarding allegations that Potts has killed as many as 13 people in Alabama, Florida, Georgia, Kentucky, Pennsylvania, and New York. This example is particularly relevant to the issue of drugs and crime because, as a Florida Department of Corrections spokesperson explained, the early release of Frank Potts can be attributed to the fact that the agency is bound by mandates from the courts and the legislature. In the mid-1980s, the prison system was inundated with inmates carrying minimum-mandatory sentences during the countrys initial skirmishes in the war on drugs (Associated Press, Probe: Potts Granted Early Release, Tallahassee Democrat, May 10, 1994, p. 5B). Some law enforcement officials have clearly begun to realize the trade-offs arising from a war on drugs.
 Decentralizing the prison system would have to involve major institutional reform, of course. Benson and Wollan (1989) and Rasmussen and Benson (1994, chapter 8) discuss an array of possible institutional changes to achieve this goal, along with the potential problems and benefits that would accompany such changes.
 See note 19 in this regard.
Baridon, P. C. 1976. Addiction, Crime, and Social Policy. Lexington, Mass.: D. C. Heath.
Becker, Gary S. 1968. Crime and Punishment: an Economic Approach, Journal of Political Economy (March/April): 169217.
Benson, Bruce L. 1981. Corruption of Public Officials: The Black Market for Property Rights. Journal of Libertarian Studies 5 (Summer): 30511.
. 1988a. An Institutional Explanation for Corruption of Criminal Justice Officials. Cato Journal 8 (Spring/Summer): 13963.
. 1988b. Corruption in Law Enforcement: One Consequence of The Tragedy of the Commons Arising with Public Allocation Processes. Int. Review of Law and Economics 8 (June): 7384.
. 1990. The Enterprise of Law: Justice Without the State. San Francisco: Pacific Research Institute for Public Policy.
Benson, Bruce L., and John Baden. 1985. The Political Economy of Government Corruption: The Logic of Underground Government. Journal of Legal Studies 14 (June): 391410.
Benson, Bruce L., Iljoong Kim, David W. Rasmussen, and Thomas W. Zuehlke. 1992. Is Property Crime Caused by Drug Use or Drug Enforcement Policy? Applied Economics 24 (July): 67992.
Benson, Bruce L., and David W. Rasmussen. 1991. The Relationship Between Illicit Drug Enforcement Policy and Property Crimes. Contemporary Policy Issues 9 (October): 10615.
. 1992. Illinois War on Drugs: Some Unintended Consequences. Heartland Policy Study no. 48, April.
Benson, Bruce L., David W. Rasmussen, and David L. Sollars. n.d. Police Bureaucracies, Their Incentives, and the War on Drugs. Public Choice 83 (April 1995): 2145.
Benson, Bruce L., and Laurin A. Wollan, Jr. 1989. Prison Overcrowding and Judicial Incentives. Madison Paper Series 3 (May): 121.
Canadian Governments Commission of Inquiry. 1973. Interim Report: The Non-Medical Use of Drugs. New York: Penguin.
Chaiken, Jan M., and Marcia R. Chaiken. 1990. Drugs and Predatory Crime. In Drugs and Crime, edited by Michael Tonry and James Q. Wilson. Chicago: University of Chicago Press.
Chaloupka, Frank J., and Adit Laixuthai. 1992. Do Youths Substitute Alcohol and Marijuana? Some Econometric Evidence. Mimeo, November.
Chein, Isidor, Donald L. Gerard, Robert S. Lee, and Eva Rosenfeld. 1965. The Road to H: Narcotics, Delinquency, and Social Policy. New York: Basic Books.
DiNardo, John. 1992. A Critical Review of the Estimates of the Cost of Alcohol and Drug Use. Irvine Economic Papers no. 90-92-18, March.
DuPont, Robert, and Mark H. Greene. 1972. The Dynamics of a Heroin Addiction Epidemic. Science 181 (August 24.): 71622.
Elliot, Nicholas. 1977. Economic Analysis of Crime and the Criminal Justice System. In Public Law and Public Policy, edited by John Gardiner, 6889. New York: Praeger.
Finn, Peter and Andrea K. Newlyn. 1993. Dade County Diverts Drug Defendants to Court-Run Rehabilitation Program. National Institute of Justice: Program Focus (June).
Fisher, Irving. 1927. Prohibition at Its Worst. rev. ed. New York: Houghton Mifflin.
Furlong, W. J., and S. L. Mehay. 1981. Urban Law Enforcement in Canada: An Empirical Analysis. Canadian Journal of Economics 14: 4457.
Gandossy, R. P., J. R. Williams, J. Cohen, and H. J. Harwood. 1980. Drugs and Crime: A Survey and Analysis of the Literature. Washington, D.C.: National Institute of Justice, GPO.
Gill, Andrew M., and Robert J. Michaels. 1990. Drug Use and Earnings: Accounting for the Self-Selection of Drug Users. Working Paper no. 11-90, Department of Economics, California State University, Fullerton.
Girodo, Michel. 1991. Drug Corruption of Undercover Agents: Measuring the Risk. Behavioral Sciences and the Law 9: 36170.
Goldkamp, John S., and Doris Weiland. 1993. Assessing the Impact of Dade Countys Felony Drug Court. Philadelphia, PA: Crime and Justice Research Institute, Augus.
Goldstein, Paul J. 1989. Drugs and Violent Crime. In Pathways to Criminal Violence, edited by Neil A. Weiner and Marvin E. Wolfgang. Newbury Park, Calif.: Sage.
Gottfredson, Michael R., and Travis Hirschi. 1990. A General Theory of Crime. Stanford, Calif.: Stanford University Press.
Gould, Leroy C. 1990. Criminal Episodes: An Interactionist Approach to Crime, Criminality and Criminal Behavior. Manuscript, Florida State University, Tallahassee.
Gould, Leroy C., Andrew L. Walker, Lansing E. Crane, and Charles W. Lidz. 1974. Connections: Notes from the Heroin World. New Haven: Yale University Press.
Greenberg, S. W., and F. Alder. 1974. Crime and Addiction: An Empirical Analysis of the Literature, 19201973. Contemporary Drug Problems 3.
Hakim, Simon, Uriel Speigel, and J. Weinblatt. 1984. Substitution, Size Effects, and the Composition of Property Crime. Social Science Quarterly 65 (September): 71734.
Harlow, Caroline Wolf. 1991. Drugs and Jail Inmates, 1989. In Bureau of Justice Statistics: Special Report. Washington, D.C.: U.S. Department of Justice, Bureau of Justice Statistics, August.
Innes, Christopher A. 1988. Drug Use and Crime. In Bureau of Justice Statistics Special Bulletin. Washington, D.C.: U.S. Department of Justice.
Kaestner, Robert. 1991. The Effects of Illicit Drug Use on the Wages of Young Adults. Journal of Labor Economics 9 (4): 381412.
Kaplan, John. 1983. The Hardest Drug: Heroin and Public Policy. Chicago: University of Chicago Press.
Kennedy, David M. 1993. Closing the Market: Controlling the Drug Trade in Tampa, Florida. In Program Focus. Washington, D.C.: National Institute of Justice, March.
Kennett, David. 1982. Standards and Procedures for the Distribution of a Public Service: Shoup Revisited. Public Finance 37 (1): 8097.
Kim, Iljoong, Bruce L. Benson, David W. Rasmussen, and Thomas W. Zuehlke. 1993. An Economic Analysis of Recidivism among Drug Offenders. Southern Economic Journal 60 (July): 16983.
Kleiman, Mark A. R. 1985. Marijuana: Costs of Abuse, Costs of Control. New York: Greenwood.
. 1992. Against Excess: Drug Policy for Results. Champaign, Ill.: Basic Books.
Langan, Patrick A., and Mark Cunniff. 1992. Recidivism of Felons on Probation, 198689. In Bureau of Justice Statistics: Special Reports. Washington, D.C.: U.S. Department of Justice, Bureau of Justice Statistics, February.
Lawrence, Richard. 1991. The Impact of Sentencing Guidelines on Corrections. Paper presented at the annual meeting of the Academy of Criminal Justice Sciences.
Levine, Daniel, Peter Stoloff, and Nancy Spruill. 1976. Public Drug Treatment and Addict Crime. Journal of Legal Studies 5 (June): 43562.
McPheters, Lee R., and William Stronge. 1981. Crime Spillover in the Boston Area. In Crime Spillover, edited by Simon Hakim and G. F. Rengert, 8395. Beverly Hills, Calif.: Sage, 1981.
Mehay, Stephen L. 1977. Interjurisdictional Spillovers of Urban Police Services. Southern Economic Journal 43: 135259.
Miron, Jeffery A. 1992. An Economic Analysis of Drug Prohibition. España Económica (April): 6272.
Model, Karyn E. 1991. The Relationship Between Drug and Alcohol Use and Income: An Econometric Model. Mimeo., Harvard University, August.
. 1993. The Effect of Marijuana Decriminalization on Hospital Emergency Drug Episodes: 197578. Journal of the American Statistical Association, pp. 73747.
Moore, Mark H. 1977. Buy and Bust: The Effective Regulation of an Illicit Market in Heroin. Lexington Mass.: Lexington Books.
. 1990. Supply Reduction and Drug Law Enforcement. In Drugs and Crime, edited by Michael Tonry and James Q. Wilson. Chicago: University of Chicago Press.
Moore, Mark H., and Mark A. R. Kleiman. 1989. The Police and Drugs, Perspectives on Policing no. 11. Washington, D.C.: National Institute of Justice, September.
Myers, Samuel L. Jr. 1992. Crime, Entrepreneurship and Labor Force Withdrawal. Contemporary Policy Issues 10 (April): 8498.
Nadelmann, Ethan A. 1993. Should We Legalize Drugs? History Answers Yes. American Heritage (February/March): 4248.
National Institute of Justice. 1990. Drug Use Forecasting. Washington, D.C.: U.S. Department of Justice, National Institute of Justice.
National Institute on Drug Abuse (NIDA). 1988. National Household Survey of Drug Abuse. Washington, D.C.: Department of Health and Human Services, GPO.
Nyswander, Marie. 1956. The Drug Addict as Patient. New York: Grune and Stratton.
Rasmussen, David W., and Bruce L. Benson. 1994. The Economic Anatomy of a Drug War: Criminal Justice in the Commons. Lanham, Md.: Rowman and Littlefield.
Rasmussen, David W., Bruce L. Benson, and Iljoong Kim. 1991. A Benefit-Cost Study of Alternative Sentencing Guidelines. Report prepared under contract with the Florida House Appropriations Committee, July, Tallahassee.
Rasmussen, David W., Bruce L. Benson, and Brent D. Mast. 1994. Entrepreneurial Police and Drug Enforcement Policy. Working paper, Department of Economics, Florida State University, Tallahassee.
Rasmussen, David W., Bruce L. Benson, and David L. Sollars. 1993. Spatial Competition in Illicit Drug Markets: The Consequences of Increased Drug Enforcement. Review of Regional Studies 23 (Winter): 21936.
Reed, Terrance G. 1992. American Forfeiture Law: Property Owners Meet the Prosecutor, Cato Institute Policy Analysis no. 179, September 29, Washington, D.C.
Reuter, Peter. 1985. Eternal Hope: Americas International Narcotics Efforts. Public Interest 79 (Spring): 1394.
. 1991. On the Consequences of Toughness. In Searching for Alternatives: Drug Control Policy in the United States, edited by Edward Lazear and Melvyn Krauss. Palo Alto, Calif.: Hoover Institution Press.
Reuter, Peter, and Mark A. R. Kleiman. 1986. Risks and Prices. In Crime and Justice: An Annual Review of Research, vol. 7, edited by Michael Tonry and Norval Morris. Chicago: University of Chicago Press.
Reuter, Peter, R. MacCoun, and P. Murphy. 1990. Money from Crime: A Study of the Economics of Drug Dealing in Washington, D.C. Santa Monica, Calif.: The RAND Corporation.
Roumasset, J., and John Hadreas. 1977. Addicts, Fences, and the Market for Stolen Goods. Public Finance Quarterly (April): 24772.
Rydell, C. Peter, and Susan S. Everingham. 1994. Modeling the Demand for Cocaine. Santa Monica, Calif.: The RAND Corporation.
Saltzburg, Stephen A. 1986. Another Victim of Illegal Narcotics: The Fourth Amendment. University of Pittsburgh Law Review 48: 126.
Sherman, Lawrence W. 1983. Patrol Strategies for Police. In Crime and Public Policy, edited by James Q. Wilson. San Francisco: Institute for Contemporary Studies Press.
Silberman, C. E. 1978. Criminal Violence, Criminal Justice. New York: Random House.
Silver, S. 1974. Punishment, Deterrence, and Police Effectiveness: A Survey and Critical Interpretation of Recent Econometric Literature. New York: Crime Deterrence and Offender Career Project.
Sollars, David L., Bruce L. Benson, and David W. Rasmussen. 1994. Drug Enforcement and Deterrence of Property Crime Among Local Jurisdictions. Public Finance Quarterly 22: 2245.
Stutmann, Robert M., and Richard Esposito. 1992. Dead on Delivery: Inside the Drug Wars, Straight from the Street. New York: Warner Books.
Thornton, Mark. 1991. The Economics of Prohibition. Salt Lake City: University of Utah Press.
Tonry, Michael. 1988. Structuring Sentencing. In Crime and Justice: A Review of Research, edited by Michael Tonry and Norval Morris. Chicago: University of Chicago Press.
U.S. Department of Justice. Office of the Attorney General. 1992. Combating Violent Crime: 24 Recommendations to Strengthen Criminal Justice. Washington, D.C.: U.S. GPO, July.
Vaillant, George E. 1969. The Natural History of Urban Narcotic Drug Addiction. In The Scientific Basis of Drug Addiction, edited by Hannah Steinberg. New York: Grune and Stratton.
Waldorf, D. 1973. Careers in Dope. Englewood Cliffs, N.J.: Prentice-Hall.
Warburton, Clark. 1932. The Economic Results of Prohibition. New York: Columbia University Press.
Wassertrom, Silas J. 1983. The Incredible Shrinking Fourth Amendment. American Criminal Law Review 21: 257402.
Wexler, H., G. Galkin, and D. Lipton. 1989. Outcome Evaluation of a Prison Therapeutic Community for Substance-Abuse Treatment. New York: Narcotic and Drug Research Inc.
White, Michael D., and William A. Luksetich. 1983. Heroin: Price Elasticity and Enforcement Strategies. Economic Inquiry 21 (October):55764.
Wilson, James Q., and Richard J. Herrnstein. 1985. Crime and Human Nature: The Definitive Study of the Causes of Crime. New York: Simon and Schuster.
Wisotsky, Steven. 1987. Crackdown: The Emerging Drug Exception to the Bill of Rights. Hastings Law Journal 3F: 889926.
1992. A Society of Suspects: The War on Drugs and Civil Liberties. Cato Institute Policy Analysis no. 180. October 2, Washington, D.C.
Zimring, Franklin E., and Gordon Hawkins. 1992. The Search for Rational Drug Control. Cambridge, Mass.: Cambridge University Press.