urban institute nonprofit social and economic policy research

Did Getting Tough on Crime Pay?

Crime Policy Report No. 1

Share on Facebook Share on Twitter Share on LinkedIn Share on Digg Share on Reddit
| Email this pageE-mail
Document date: August 01, 1997
Released online: August 01, 1997

Opinions expressed are those of the authors and do not necessarily reflect the views of the American University or The Urban Institute, its trustees, or sponsors.


Since the early 1980s, sentencing reform has been dominated by the use of mandatory penalties, determinate sentences, and guideline-based sentences. Of the reforms, mandatory prison sentences have been the most popular. These have been aimed primarily at persons who commit violent crimes, who use weapons in the commission of their crimes, who are repeat or habitual offenders, or who are high-profile drug traffickers. These reforms were based on three beliefs: First, that punishments for serious crimes needed to be in-creased, as they were far too lenient; second, that the criminals targeted by the reforms were "dangerous" people who must be incarcerated because lesser sanctions would not be effective in curtailing the proscribed behaviors; and third, that mandatory prison sentences would reduce crime by incapacitating or deterring the dangerous offenders who were targets of the reform. With the implementation of the reforms, there was unprecedented growth in the size of federal and state prison populations. Between 1980 and 1995, prison populations grew from about 330,000 persons to over 1.5 million persons, including those incarcerated in jails.

Now that these sentencing reforms have been in place for more than 15 years, it is clear that the beliefs on which public support rested were largely false. Punishment for serious and violent offenders prior to the reforms was not as lenient as was popularly believed, nor less severe than punishments imposed in other industrialized nations for comparable crimes. The misperception of leniency rested in large measure on erroneous or badly used data and on the media attention that focused on the unusual or atypically short sentences given to violent offenders. In the implementation of the reforms, the definition of who was a dangerous offender was greatly expanded to include not only large numbers of violent offenders, but also large numbers of nonviolent and not-so-dangerous offenders, especially lower level drug offenders who were de-fined as dangerous in the heat of public reaction to the drug crisis of the late 1980s. There is mixed evidence that these reforms, as implemented, have led to reductions in crime. While violent crimes such as murder have decreased recently, imprisonment is but one of the many reasons. Stabilization of drug markets, community policing efforts, and aggressive enforcement of behaviors perceived as threatening also have contributed to the declines in murders and other violent crimes. For drug crimes, however, there is no evidence that the increased use of imprisonment has produced measurable reductions. Imprisonment for drug crimes is not a cost-effective sanction compared to treatment or intermediate sanctions, and its overuse for lower level drug offenders represents a misallocation of scarce prison resources.

Finally, the massive increases in incarceration that resulted from the reforms have been distributed unevenly throughout the population. In particular, there has been a geographical clustering of incarceration. In most states, the vast majority of persons admitted into prisons, particularly for drug crimes, have come from a comparatively few large urban areas within the states. Within many of these areas, in a single year, almost 4 percent of the young black men between the ages of 16 and 34 were removed into prison. This rate exceeded those of comparably aged white men by seven to ten times. The exceedingly high removal rate may contribute to family disruption and social organization of communities, thereby contributing to crime problems. This is especially the case in black, inner city communities.


In the past 15 years, there has been an unprecedented growth of state and federal prison populations (Gilliard and Beck 1994; Farrington and Langan 1992). Between 1980 and 1993, the number of persons incarcerated in these institutions grew from 329,000 to 949,000. By 1994, there were over a million persons in prison. Including those incarcerated in jails, over 1.5 million people were in prison at the end of 1995 (Gilliard and Beck 1996).

These increases in the prison populations were the result of sentencing reforms such as mandatory penalties, determinant sentencing and guideline-based sentences. The reforms were aimed primarily at persons committing violent crimes, those using guns in the commission of their offenses, habitual or repeat offenders, and high-profile drug traffickers. The reforms were (and continue to be) supported by three beliefs: (1) punishments for serious crimes were lenient; (2) criminals targeted by the reforms were "dangerous" people who must be incarcerated because lesser sanctions would not be effective; and (3) imprisonment would reduce crime by incapacitating or deterring these offenders.

Now that these sentencing reforms have been in place for more than 15 years, it is clear that public support for them rested on oversimplified arguments and erroneous data. Punishment of serious and violent offenders prior to the reforms was not as lenient as the public believed, nor was it less severe than the punishment of comparable offenders in other industrialized countries. The implementation of these reforms resulted in the imprisonment not only of large numbers of violent and dangerous offenders, but also of a large number of not-so-dangerous offenders. While violent crimes such as murder have decreased recently, a balanced view of the evidence shows that imprisonment is only one of the many reasons behind this. Other reasons are stabilization of drug markets, community policing, and aggressive enforcement of behaviors perceived as threatening, as well as attention to visible signs of decay. There is no evidence that the increased use of imprisonment produced measurable reductions in drug crimes. Rather, there is growing evidence that the massive increases in incarceration, particularly for drug and other nonviolent crimes, have led to a misallocation of scarce prison resources, and may be undermining families and communities, thereby contributing to crime problems. This is especially true in black, inner city communities.

Rather than continuing to pursue expanding the use of imprisonment for more classes of offenders and increasing its use for drug and other nonviolent offenders, punishment policies need to be devised to meet three objectives: (1) punish offenders; (2) ameliorate the negative consequences of past policies; and (3), for drug crimes, to address both the underlying economic motivations and the problems of drug addiction in lower level dealers. To these ends, severe punishments for serious and violent offenders need to be maintained. For less serious property offenders and low-level drug sellers, a mixed strategy of targeted intermediate sanctions in conjunction with expanded economic opportunities is needed.

For these persons, sanctions alone will not remove the underlying economic motivation for engaging in these crimes, but intermediate sanctions will reduce the collateral damage of incarceration, i.e., the negative consequences of imprisonment for individuals — such as reductions in wages and earnings — and for the social organization of the communities from which they come. Finally, given the relatively small number of communities from which the bulk of these nonviolent offenders come, this mixed strategy of sanctions and economic opportunities can be targeted very specifically in order to increase both the feasibility and the impact of such an urban policy.

Misperceptions about Lenient Prison Sentences

The sentencing reforms of the 1980s were driven in part by perceptions that serious and violent criminals were getting off easy. Public perceptions are driven by the sentencing for very serious offenses, particularly murder or homicide (Biderman 1995), and media reports of sentencing are most likely to focus on unusual and lenient sentences for violent offenders. Lighter sentences are more easily recalled in the public’s consciousness than are severe sentences, which actually are more common (Roberts 1992).

Contributing to the public perception of leniency is the absence of clear distinctions between types of violent crimes in media reports of public discussions. Homicide and murder are confused, as are robbery and theft. These crimes are, by law, differently defined and thus sentenced differently. Homicide, for example, refers to a broader class of crimes than murder, including less serious crimes such as negligent homicide and vehicular homicide. These carry shorter sentences than murder. As a result, the time served in prison for homicide is much shorter on average than time served for murder. When time served for homicide is mistaken for time served for murder, as is regularly done when the terms are used interchangeably, the public is appropriately concerned about short prison terms. Similarly, robbery is a violent crime. When it is confused with theft, and the shorter sentences given for theft are perceived as those given for robbery, the public is again appropriately concerned about short sentences.

The perception of leniency in the sentencing of violent offenders also stems from the inaccurate use of information on sentencing practices. The most commonly cited statistic proffered as evidence of the lenient treatment of violent offenders is the time served in prison by a group released from prison during a given year. For example, the Council on Crime in America asserts, as a measure of leniency, that the average time served by murderers released from prison during 1992 was less than six years.

However, the time served by a group released from prison in a given year (an exit cohort) cannot be used to generalize about sentencing policy. The exit cohort is neither representative of persons receiving prison sentences in a given year, nor of persons in prison who received their sentences in earlier years. The exit cohort will yield biased estimates of the time that prisoners can expect to serve when (1) prison populations are growing in size; (2) sentence lengths are increasing; and (3) the number entering prison is growing (Biderman 1995). All of these conditions characterized U.S. prison populations during the 1980s and 1990s. Therefore, the average time served in prison by an exit cohort will substantially understate the severity of punishment for the average or typical offender.

A very different picture of how severely we punish violent offenders is painted by the data on persons who entered prison during 1979 for murder. The majority of these offenders were still in prison in 1996; an estimated more than 60 percent were in prison 12 years after their admission, and an estimated half were still in prison 17 years after their admission — in 1996 (Figure 1). Therefore, the average time served by persons admitted into prison for murder during 1979 is likely to be greater than 17 years (based on the shape of the curves in Figure 1), or almost three times as long as the 6.8-year average based on the exit cohort data and cited as evidence of leniency. Whether 12 or 17 years are long enough for most murderers to serve in prison is a matter for policy debate. But that is a separate issue from the facts concerning how long murderers currently can expect to serve in prison.

Data on the experiences of an entry cohort over time will provide valid information about the impacts of the sentencing policies in effect at the time the offenders were admitted into prison. For example, had the data in Figure 1 been available during the debates over sentencing reforms, it is doubtful that the public would have been left with the impression that sentencing policies were overly lenient. Unfortunately, discussion about sentencing policy cannot wait 12 or 17 years to answer questions about the severity of punishment. However, there are methods for obtaining more current data on the estimated amount of time that offenders can expect to serve in prison. Standard demographic techniques such as life tables (Shryock and Seigel 1973) provide a more accurate gauge than estimating time served by the current prison population from the time served by exiting cohorts. These techniques already are used in familiar contexts, such as calculating life insurance premiums and estimating life expectancy based on the experience of the living population at a given point in time.

As shown in Table 1, the life table estimates illustrate that the expected time to be served for murder or homicide is longer than the most frequently cited estimates of time served that are based on exiting cohorts. The life table estimate of the expected time served for persons sentenced for murder in 1991 is 12.8 years, compared to the 7.7 years served by the 1991 exit cohort. This life table estimate of expected time to be served is about 60 percent longer than the exit cohort estimate. As expected, time served for murder is substantially more than that served for the more heterogeneous category of homicide. In 1991, the life table estimate for expected time to be served for homicide was 9.3 years, which is 73 percent of the estimate for murder. A similar set of results apply to the data for 1986.

Particularly unfortunate for the sentencing reform debates is the fact that the disjuncture between perception and practice vis-à-vis sentences for murder and homicide was discernible during the 1980s, when "get tough" sentencing policies were being debated and implemented. This misperception of leniency in sentencing for murder and homicide helped feed the general view that the criminal justice system was soft on crime. This, in turn, led to demands to get tough not only on violent crime, but also on drug offenses and other nonviolent crimes. Ironically, even before this crackdown on nonviolent crime, in the case of property and drug crimes, the U.S. was already substantially more likely to imprison and for a longer period of time than other industrialized nations (Lynch 1988; Lynch 1993; and Lynch et al. 1994). In the mid-1980s, for example, the average time served for theft in the U.S. was 1.5 times that served in England and 3.5 times that served in Canada. Ironically, despite the frequent claims by the media and public opinion-makers that sentences are too short and judges are too lenient, a recent study of public attitudes toward punishment found that the public viewed the federal sentences prescribed for drug traffickers — sentences determined largely by the mandatory minimum penalties for these offenses and passed during the frenzy of the drug war — as much harsher than the public deemed appropriate (Rossi and Berk 1997).

The Impacts of Sentencing Reform on the Prison Population:
On Whom Did We "Get Tough?"

The sentencing reforms of the 1980s aimed to increase the severity of sentences for offenses and offenders deemed to be "dangerous." This was to be done by increasing the likelihood that a convicted offender would go to prison, increasing the length of prison terms, or both. The reforms defined "dangerous persons" as offenders with a history of violent criminal activity and drunk drivers, and "dangerous behaviors" as murder, aggravated rape, drug offenses, felonies involving firearms, and sex offenses (Tonry 1992; BJA 1996). Among the states reforming their sentencing structure, there was broad consensus that habitual and violent offenders were dangerous, as were crimes involving weapons. More states (40), for example, passed mandatory minimum penalties for offenders with long criminal histories than for any other type of behavior, except possession of weapons (BJA 1996). On the other hand, there was less consensus among states about the specific types of drug offenses and offenders that were defined as dangerous. Over 30 states passed mandatory penalties for drug offenses, but the types of drugs, the quantities required for the mandatory minimum, and the penalties varied widely enough to preclude typifying proscribed behaviors, with the possible exception of high-profile, violent drug dealers (Tonry 1992).

Following the implementation of the sentencing reforms, the composition of federal and state prison populations changed. In some respects, the changes reflected the priority the reforms gave to dangerous offenders and behaviors. For example, between 1979 and 1991, the number of violent offenders in prison doubled (Beck et al. 1994). In other respects, the prison population changed in ways that did not reflect the reforms. For example, the number of prisoners who were convicted of nonviolent offenses and who had comparatively minor criminal histories increased more rapidly than the number of violent offenders. The increase in the number of drug offenders in prison was due to the incarceration of lower level dealers, not to the incarceration of high-profile dealers who led large organizations or dangerous gangs. The share of the prison population having comparatively strong connections to mainstream social institutions, such as education and labor markets, increased. Prior to the sentencing reforms, these offenders had been generally considered to be relatively good risks for non-prison sentences such as probation. Finally, punishment practices resulting from the reforms had the unintentional effect of creating habitual offenders, as an estimated 150,000 persons per year entered prison for the first time; this amounted to more than one-third of all state prison admissions.

Changes in the Criminality and Dangerousness of Prisoners

As shown in Table 2, the state prison population increased from about 275,000 in 1979 to about 712,000 in 1991. The number of offenders incarcerated by states for violent crimes more than doubled over this period, from about 159,000 to more than 335,000. The number of state prisoners convicted of violent crimes or having violent prior incarceration histories also more than doubled over this period, from almost 177,000 to more than 406,000, as shown in Table 3. However, the proportion of state prisoners incarcerated for violent crimes decreased from about 58 percent to about 47 percent (Table 2), and the proportion of state prisoners convicted of violent crimes or having violent prior incarceration histories also decreased from about 64 percent to about 57 percent (Table 3).

This decrease in the proportion of prisoners incarcerated for violent crimes corresponded with an increase in the number of offenders incarcerated for drug offenses. The number of drug offenders increased eightfold, from about 18,000 to more than 150,000, so that by 1991, drug offenders comprised close to 21 percent of state prison populations — up from about 6 percent in 1979. This increase reflected a change in the priorities of law enforcement agencies more than a change in the incidence of violent crimes. From the late 1970s through the early 1990s, the number of victimizations of violent crimes increased slightly, as did the rate at which they were reported to the police (Zawitz et al. 1993). However, arrests and convictions of drug offenders increased dramatically throughout the 1980s and 1990s. Arrests for drug abuse violations increased from 560,000 in 1981 to over 1.4 million in 1994 (Austin and Cohen 1996). Prison admissions of drug offenders increased tenfold between 1980 and 1993 (Brown et al. 1996). By 1993, drug offenders comprised 30 percent of persons admitted to state prisons, up from 7 percent in 1980.

Moreover, the drug offenders incarcerated in state prisons typically did not conform to the image of the high-profile dealers, leaders of large operations, or violent gang leaders that were the focus of sentencing reform. Of the more than 150,000 drug offenders in prison in 1991, almost 127,000 (about 84%) had no history of prior incarceration for violent crimes (Table 3). Not shown in the table, half of these offenders had no prior incarcerations at all. One-third of the drug offenders sentenced to state prison were convicted of less serious drug possession offenses (Perkins 1994) — further diluting the image of the typical drug prisoner as a high-profile dealer — and only about 6 percent of all state prison inmates belonged to a gang before entering prison (Beck et al. 1993).

A comparison with the federal criminal justice system underscores the point that drug offenders incarcerated in state prisons typically were not high-profile dealers. Persons convicted of drug crimes in state courts were less likely to be convicted of drug trafficking than those convicted in the federal courts (50 percent versus 89 percent), and those convicted of drug trafficking in state courts were less likely to be sent to prison than those convicted in federal courts (41 percent versus 91 percent). And, as Harer (1993) points out, 36 percent of the persons incarcerated for drug trafficking in the federal system were "mules" — low-level dealers who had no history of violent crime, no involvement in sophisticated criminal activity, and no prior prison term.

To be sure, many prisoners were involved in both violent crimes and drugs. For example, in 1991, about 23,000 of the more than 150,000 offenders (16%) had prior histories of violent crimes. And, many of the state prisoners sentenced for violent crimes in 1991 also were involved with drugs or were under the influence of drugs at the time they committed their offense (Beck et al. 1993). Nevertheless, the drug offenders incarcerated in state prisons typically were not high-profile dealers. The huge increase in the number of these offenders in state prisons is consistent with the zero-tolerance approach to drugs, but not consistent with the emphasis of sentencing reform on high-profile dealers.

The sentencing reforms also paid special attention to violent and habitual offenders, and the number of prisoners with either a current admission for a violent offense or a prior incarceration for a violent offense increased between 1979 and 1991 from more than 176,000 to about 406,000 (Table 3). However, there were even larger increases in the number of prisoners who fell outside of these categories. The number of offenders who were incarcerated for nonviolent offenses and had no criminal history of violence increased from about 71,000 to 305,000 between 1979 and 1991. Consequently, the proportion of state prisoners incarcerated for violent offenses or having violent criminal histories actually decreased from about 64 percent in 1979 to about 57 percent in 1991 (Table 3). Finally, the proportion of persons in prison for a violent crime or prisoners with a prior incarceration for any reason fell; the proportion of prisoners outside of this target of sentencing reform, therefore, increased from 16 percent in 1979 to 21 percent in 1991. This 21 percent translated into about 150,000 persons on any given day in 1991.

Who Was Entering Prison?

Data on the stock of persons in prisons on any given day overstate the number of dangerous offenders affected by sentencing reform and, thus, overstate the degree to which sentencing practices coincided with the intentions of sentencing reform. The prison population on any given day is overrepresented by persons who have been sentenced to longer prison terms, because the probability of being in the prison stock is a function of the length of sentence imposed, and persons with longer sentences will be overrepresented in the stock population. Given that prisoners with longer sentences generally have committed more violent offenses or have longer or more serious criminal histories, the prison stock data will understate the degree to which prison was used for nonviolent offenders and for persons with limited criminal histories.

Looking at a cohort of persons entering prison helps us obtain a better picture of the scope of the impact of sentencing reforms by showing how many nonviolent offenders were incarcerated. We derived estimates of the criminal histories of persons admitted into prison during 1990 by using official statistics from several national and subnational data sources. 1 These data show that about 24 percent of persons sentenced to incarceration2 during 1990 were not "dangerous offenders," in that they were not sentenced for a violent crime, nor did they have prior convictions in adult court (Figure 2). If we look at that group of nonviolent offenders without prior convictions and add to it those with only misdemeanor convictions, the proportion of persons sentenced to incarceration who were not "dangerous" increases to almost 39 percent. By going even further and including those with one prior conviction for a felony, then about 54 percent of persons sentenced to incarceration in 1990 could be considered lower risk, relative to the dangerous offenders who were the primary target of the stricter sentencing policies introduced in the 1980s.

Using these data, we estimated that 20 percent of the persons sentenced to state prison were sentenced for a nonviolent offense and had no prior convictions. While an estimated 29 percent who had only prior misdemeanor convictions were sentenced for a nonviolent offense, an estimated 46 percent who had one prior felony conviction were incarcerated for a nonviolent offense. These percentages translate into an estimated 100,000 persons admitted to state prison in 1991 for a nonviolent offense who had no prior convictions, and about 150,000 persons admitted in 1991 for nonviolent offenses who had no prior felony convictions. Including the number of sentenced prisoners admitted into jails would raise this number even higher. For example, about half, or 200,000 of the roughly 400,000 prisoners in local jails, were sentenced prisoners (Lynch et al. 1994). By comparison, in 1991, 711,000 sentenced inmates were in state prisons.

The changes in sentencing practices begun in the 1980s have resulted in an increase in the incarceration of persons convicted of violent offenses who have prior histories of violence and other crimes. At the same time, these reforms have led to a proportionately larger increase in the incarceration of nonviolent offenders with little or no criminal history. More than 125,000 persons in prison in 1991 (18 percent of the prison population on any given day) were neither violent nor classified as dangerous offenders. From the perspective of flows into prison, between 100,000 and 150,000 persons admitted into prison during 1991 were neither violent, nor classified as dangerous offenders. At these rates, within several years, close to 500,000 nonviolent offenders with little or no criminal history could expect to enter prison. These offenders bear little resemblance to the serious, violent, habitual, or dangerous offenders that were the major focus of sentencing reform.

Increases in "Socially Integrated" Offenders Imprisoned

Along with the changes in the offense and criminal history composition of prison populations was an increase in the incarceration of "socially integrated offenders." Social integration refers to the degree to which offenders have ties to legitimate institutions such as families, education, and labor markets. The presence of these ties has been shown to reduce the risk of criminal behavior because of the socializing influences of these institutions and the informal sanctions they impose for proscribed behaviors. The incarceration of socially integrated offenders may be both unnecessary and unwise. It may be unnecessary because prior experience has shown that socially integrated people are less likely to re-offend. Such people can, therefore, be punished by means other than incarceration without putting the public at undue risk. It may be unwise because incarcerating socially integrated people is likely to reduce their ties to legitimate institutions, thereby making them more prone to subsequent criminal involvement. Moreover, as we will argue next, taking these individuals out of their communities further weakens the ability of these communities to prevent criminal behavior.

The number of inmates who had a high school education or above, were employed prior to their arrest, and reported as ever having been married increased between 1979 and 1991, as shown in Table 4. Moreover, the number who did not commit a violent crime or have a violent criminal history and who had social ties also increased over this period (Table 5). Overall, the number with at least a high school education (or equivalent) increased from about 78,000 to almost 415,000. The number with some college education rose from about 10,000 to 44,000 persons; the number ever married more than doubled, from about 132,000 to almost 315,000 persons; and the number employed increased from about 192,000 to more than 476,000 (Table 4).

Looking only at inmates incarcerated for nonviolent offenses, the number with social ties also increased over this period. The number with a high school education or more increased from almost 34,000 to about 211,000; the number with some employment prior to prison tripled, from about 81,000 to more than 243,000; and the number ever married almost tripled, from about 56,000 to almost 159,000.

The Impacts of Sentencing Reform on Crime

The sentencing reforms of the 1980s aimed to control and reduce crime by using prison sentences to deter and incapacitate offenders. Deterrence occurs when the threat of prison sanctions prevents people not in prison from committing crimes. Incapacitation reduces crime by removing from the street people who commit crimes at high rates over long periods of time — the so-called career criminals. By removing and incarcerating these criminals, their crimes are prevented and crime-reduction benefits accrue. These two types of crime control effects interact in complex ways, and empirical research has not satisfactorily separated them (Blumstein et al. 1978). For example, estimates of the deterrent effects of increases in imprisonment on changes in crime rates combine deterrent and incapacitation effects, while estimates of incapacitation effects do not take into account any deterrent effects that might have reduced the individual crime rates to levels below what they would have been in the absence of deterrent effects. In addition to sanctions, many social, economic, and demographic factors — such as age distribution, unemployment rates, wages, welfare rates, etc. — are directly and indirectly associated with changes in crime rates (Austin and Cohen 1996). Disentangling the contribution of these effects to changes in crime rates also presents difficult methodological problems.

Nevertheless, crime control benefits of imprisonment can occur by increasing the certainty of punishment, increasing the severity of punishment, or both. During the 1980s, both of these increased for violent and drug offenders, as the likelihood of imprisonment given arrest and the length of time spent in prison increased.

However, the substantial increases in imprisonment were not associated with substantially large estimated reductions in violent crime. The tripling in the number of violent offenders in prison during the 1980s resulted in only an estimated additional 9 percent decrease in violent crimes above the decrease that would have occurred had imprisonment not grown (Cohen and Canela-Cacho 1994).

Levitt (1996) estimates that the deterrent impact on violent crimes of adding one additional prisoner amounts to a reduction of slightly more than two violent crimes. He argues that these reductions, along with the estimated reductions in property crimes, have been an effective approach to reducing crime, due largely to the costs to society of violent crimes such as murder and rape. He suggests, however, that if the amount of incarceration is to be increased, keeping the current pool of violent offenders behind bars for a longer period of time is a more advisable policy approach than that of vastly expanding incarceration to include minor offenders. Why? Two reasons: First, the current stock of prisoners are more criminally active than offenders who come into contact with the criminal justice system. In technical terms, this has been described as "stochastic selectivity," a concept meaning that offenders who commit crimes at high rates have an overall higher likelihood of being caught, convicted, and imprisoned than do offenders who commit crimes at lower rates, even if the chances of capture and conviction are the same per criminal event. Second, the length of criminal careers of new offenders is likely to be shorter than the length of the careers of offenders already in prison, thereby reducing the marginal crime reduction benefits of incarceration (Cohen and Canela-Cacho 1994).

For drug offenders, the prison stock increased eightfold during the 1980s, but there has been no demonstrable effect on reducing drug crimes (Blumstein 1995a, 1995b). Further, Caulkins et al. (1997) found that the mandatory minimum sentences responsible for the increase in the number of drug offenders in prison "are not justifiable on the basis of cost-effectiveness at reducing cocaine consumption, cocaine expenditures, or drug-related crimes" (p. xvi). Rather, these mandatory minimum sentences for drug crimes wasted scarce prison resources and often forced prison administrators to release serious offenders in order to relieve stress caused by prison overcrowding.

Why didn't the longer sentences for drug offenders reduce drug crimes? There are a few reasons, but fundamentally, drug crimes are economically motivated. Either users commit crimes to buy drugs, or dealers sell drugs to obtain money. Most drug sellers are small-time operators, relatively young, and have short criminal careers. In the absence of incarceration, they would have stopped their drug activities after a short period of time. If they are incarcerated for a mandatory term, much of the time they spend in prison is time during which they would not have committed the drug crimes anyway. On the other hand, big-time dealers simply incorporate the costs of imprisonment into their costs of doing business, so the risk of imprisonment may cause them to raise the price of drugs, effectively decreasing the demand for their product. However, at the higher prices, their total revenue may remain constant or even increase, depending on the price elasticity of demand. Therefore, a decrease in demand for drugs may not necessarily drive dealers out of the market. Removing some big-time dealers by incarcerating them simply opens up the market for other dealers. Conversely, removing small-time dealers does not necessarily interrupt the drug markets because big-time dealers have little trouble recruiting small-time dealers. Studies of drug markets show that small-time dealers tend to be young or to perceive that the economic gain of small-time dealing is greater than it actually is (Sullivan 1989; McGahey 1986). They tend to desist from drug selling activity after a relatively short period of time because they recognize that the gains were not as large as expected, they fear the violence, or they are incarcerated.

Using incarceration to control crime may be an effective strategy to combat some types of crimes, particularly those involving violent offenders and offenders with long and serious criminal careers. Ironically, as imprisonment increases, its marginal crime reduction potential decreases. This is because at any point in time, offenders who commit more crimes are more likely to end up in prison than are the less serious offenders, and expanding incarceration to include these lower rate offenders will produce smaller reductions in crime. The exception to this rule — that at any point in time, offenders who commit more crimes are more likely to end up in prison — occurs when criminal justice policy selectively targets a group of less serious offenders, such as drug offenders. The zero-tolerance approach to drug crimes has led to the massive increase in the incarceration of less serious offenders who typically have short criminal careers. For many of these offenders currently in prison or who will enter prison, the longer prison sentences stemming from the reforms could yield zero crime control benefits.

The reliance on incarceration for crime control has come about at a substantial cost to society. In 1990, for example, $25 billion was spent on corrections (Freeman 1996). With more than 1 million persons in prison that year, the average cost per prisoner was $22,000. With about 20 percent of the prison population in 1990 consisting of drug offenders, the amount spent on incarcerating them was about $5 billion. Given that the bulk of these offenders were lower level offenders with minor or no criminal histories, a sizable portion of the money spent locking them up produced no crime reduction benefits.

Impacts of Sentencing Reform on Communities

Research and speculation suggest that the massive increases in the use of incarceration have had negative consequences for community organization, particularly within black communities. Within black communities, men with comparatively strong social ties appear to have borne the brunt of the drug enforcement effort. For example, between 1986 and 1991, the incarceration rate for socially integrated black men who had incomes above the poverty level increased from 21 to 37 per 1,000 of these men in the general population. These rates are eight times those for a comparable group of white men. The incarceration rate for black men who did not have strong ties to legitimate social institutions decreased slightly, from 32 to 29 per 1,000. More striking, however, are the offense-specific increases in incarceration rates. For socially integrated black men, the drug incarceration rate increased from less than 2 per 1,000 in 1986 to over 9 per 1,000 in 1991; for property offenses, it increased from 6 to 8 per 1,000; and for violent offenses, from 13 to 17 per 1,000 (Lynch and Sabol 1992). Two-thirds of the growth in the incarceration rate for socially integrated black men over this period was attributable to the increase in their incarceration rate for nonviolent crimes: 50 percent was due to the increase in their incarceration rates for drug offenses and 17 percent to the increase in their incarceration rates for property offenses.

What these national-level data don't reveal is that the bulk of prison admissions for drug and nonviolent crimes come from a relatively few number of communities. For example, Los Angeles County provided more than 37 percent of all persons admitted into state prison in California during 1985 and 1992, but it comprised less than 12 percent of the state's total population. Similarly, in these same years Baltimore City contributed more than 50 percent of all persons admitted into prison in Maryland, but it comprised only about 15 percent of the state's population. Among the states for which these data were available in 1992, this pattern of a few counties providing the majority or a disproportionately (relative to population) large share of prison admissions held steady. For example, in Illinois, one county (Cook) provided 69 percent of prison admissions; in Minnesota and Oklahoma, two counties provided 50 percent; in New York, four provided 68 percent; in Oregon, three provided 58 percent; and in Pennsylvania, two provided 52 percent. In each case, these counties provided a share of prison admissions that exceeded their share of the state’s population.

These concentration effects are even greater for prime-age males between the ages of 16 and 34, the ages when men complete school, start careers, and begin families. Men in this age group had the highest prison admission rates. During 1992, for example, an estimated 3.7 percent of black men between the ages of 16 and 34 in the counties including, and immediately surrounding, core cities of large metropolitan areas were admitted into state prisons. This rate was seven times higher than the rate for whites in the same age group. For black men in the core metropolitan areas between the ages of 16 and 34, an estimated 1.3 percent were arrested and sentenced to prison for drug offenses during 1992. This figure was 10.5 times that for comparably aged white men.

The concentration of drug-related incarceration of black men between the ages of 16 and 34 that accelerated in the mid-1980s coincided with the collapse of the job market and the declines in wages for less skilled men in major metropolitan areas (Freeman 1996; Grogger 1997). Incarcerating these men at this stage in their lives can lead to several negative consequences. It can decrease the post-release employability of men in areas with comparatively limited opportunities. It can also lead to a decline in wages post-release (Lott 1993; Woldford 1993). In areas that already suffer from wage declines, imprisonment furthers the decline. In turn, these incarcerated men become less attractive as marriage partners or as fathers, thereby affecting family structure (Darity and Myers 1994; Sampson 1995). The breadth of the potential impacts of incarceration reach deeply into black communities, as an estimated 29 percent of black men are likely to be admitted into prison during their lifetime (Bonczar and Beck 1997, forthcoming).

Further, incarceration may contribute to the incidence of out-of-wedlock births and to an increase in the percentage of families headed by females (Lynch and Sabol 1992; Darity and Myers 1994). Imprisonment may impose additional burdens on the children of parents who are incarcerated, as Hagan (1996), Moore (1996), and Nightingale and Watts (1996) point out. The interplay of these factors may result, paradoxically, in a rise in juvenile violent crime. Incarcerating a large portion of a community’s young men can weaken its ability to protect itself against other crimes — either by weakening informal social controls (e.g., neighbors noting strangers or assuming responsibility for juveniles), or by increasing the number of targets for criminal opportunities (Sampson 1987).

This is not to say that offenders should never be removed from communities by incarceration. Violent and repeat offenders may pose too great a risk to serve their sentences within the community. Incapacitating them makes some sense, but it is unrealistic to expect huge reductions in violent crime from increasing the length of their incarceration. For property offenders and low-level drug offenders, incarceration does not address the economic motivation for engaging in these crimes in the first place, and by diminishing an offender’s capacity to engage in legitimate work after release, incarceration merely increases the likelihood that an offender will return to drug and property crimes. Conversely, punishing these offenders in the community will do less to weaken their attachments to the community, and it will not weaken their earnings potential as severely as would incarceration.

Policy Implications

The sentencing reforms of the 1980s led to a vastly expanded number of violent offenders in prison. They also resulted in increases in the number of nonviolent offenders incarcerated, particularly lower level drug offenders. Although the massive increases in incarceration of violent offenders are associated with minor reductions in violent crimes, there is no evidence that more widespread imprisonment of drug offenders has reduced drug crimes. Moreover, incarcerating less serious offenders who have not only minor or no criminal histories, but also have ties to families, schools, churches, and labor markets, may adversely affect communities.

These effects of the sentencing reforms raise the question of how we should use scarce prison resources. The evidence that prison is cost beneficial in reducing crime is mixed. The research shows that prison may be cost beneficial for violent crimes, but it also shows that it is unrealistic to expect huge reductions in violent crime with large increases in imprisonment. The high value placed on human life ensures that incarcerating violent offenders is cost beneficial.

The research also shows that imprisonment is not cost effective in reducing drug crimes, and may only be marginally cost beneficial in reducing property crimes. Continuing to incarcerate large numbers of low-level drug dealers and minor property offenders makes little sense for these crimes. Rather, these crimes are caused by income deficiencies, and require an income policy to address their cause. In addition, as research on drug courts is beginning to show, many lower level drug dealers also are drug addicts. Their addictions require drug treatment programs (Brooklyn Treatment Court Program 1997).

What sanctions can be implemented that punish, produce cost-effective reductions in crime, and do not damage communities? Alternative sanctions such as community corrections or intermediate punishments such as intensive supervision, work release, community service, house arrest with electronic monitoring, and even boot camps may also prove to be more cost effective than incarceration for less serious offenders. Evaluations of these alternative approaches are rare but instructive. To operate as effective alternatives to prison, intermediate sanctions need to be used for persons who, in their absence, would have been sent to prison. At present, such sanctions tend to be used for persons who otherwise would have received probation. When used as an alternative to prison, intensive supervision programs appear to be attractive alternatives for first-time drug offenders (Petersilia and Turner 1993). To be managed successfully, however, these sanctions require the capabilities that allow for sufficient supervision (Blumstein 1995). Otherwise, judges can only choose between probation with no monitoring or prison, which is excessively severe for less serious offenders.

Even if alternatives to prison such as intermediate sanctions were to function well, sanctions alone cannot address the problems of drug crimes in urban areas. This is because such crimes are primarily economic. With the collapse of the labor market for low-skill workers in the 1980s (Wilson 1987; Freeman 1996), legitimate sources of economic gain were eliminated in many communities. Stiffer criminal penalties designed to make legitimate work more attractive have not reduced drug crimes, in large part because of the lack of legitimate alternatives. Raising the costs of doing crime (by increasing penalties) will only work if changes in the level of punishment are sufficiently high to cause a reduction in criminal involvement. This will occur only if there are alternative and legitimate opportunities that make the costs of the sanctions something worth avoiding.

To make sanctions effective, economic opportunities as well as returns to these opportunities need to be expanded. For nonviolent offenders and drug addicts, a more varied and graduated system of sanctions and incentives with the goal of increasing social integration should be developed. For instance, to address the situation confronting drug sellers, a mixed strategy of intermediate punishments in conjunction with expanded long-term economic opportunities is required (Kleiman 1997). For drug users, probation with escalating prison sanctions for repeated failures of drug tests is one way to minimize costs, maximize public safety, and permit individuals to choose to behave responsibly. Incarceration would be used as a last resort when lesser sanctions did not change behavior. However, none of these sanctions alone can be expected to alter the relative returns to crime in environments in which legitimate opportunities are lacking. Given that the areas providing the bulk of young, male, nonviolent, and low-level drug offenders are the same ones that have suffered wage declines and the collapse of the labor market for low-skilled workers, the mixed strategy of sanctions and economic opportunities can be targeted very specifically. This will enhance the feasibility and impact of such an urban policy.

Figures and Charts


1. National data on the criminal histories of persons admitted into prison do not exist. By using data from the National Pre-trial Reporting Program (NPRP), which reports on the criminal histories of felony defendants in a sample of felony filings in the largest 75 countries in the U.S., in conjunction with data from the National Corrections Reporting Program (NCRP), which reports on the number of persons admitted into state prisons, we derived estimates of the criminal histories of person admitted into prison in 1991.

2. Sentences to incarceration include persons sentences to prison plus sentenced prisoners in local jails.


Austin, James, and Robyn L. Cohen. 1996. "Are Crime Rates Declining?" NCCD Focus. The National Council on Crime and Delinquency, November 1996.

Beck, Allan, Darrell Gilliard, Lawrence Greenfeld, Caroline Harlow, Thomas Hester, Louis Jankowski, Tracy Schnell, James Stephan, and Danielle Morton. 1993. Survey of State Prison Inmates. Washington, D.C.: Bureau of Justice Statistics.

Biderman, Albert. 1995. "Statistics of Average Time Served Are Fallacious Indicators of the Severity of Punishment." Paper presented at the Annual Meeting of the American Society of Criminology. Boston, MA, November 16.

Blumstein, Alfred, Jacqueline Cohen, and Daniel Nagin, editors. 1978. Deterrence and Incapacitation: Estimating the Effects of Criminal Sanctions on Crime Rates. Panel on Research on Deterrent and Incapacitative Effects, Committee on Law Enforcement and the Administration of Justice, Assembly of Behavioral and Social Sciences, National Research Council. Washington, D.C.: National Academy of Sciences.

Blumstein, Alfred. 1995a. Youth Violence, Guns, and the Illicit Drug Industry. Working Paper Series. Pittsburgh: H. John Heinz III School of Public Policy and Management.

Blumstein, Alfred. 1995b. "Prisons," In Crime, James Q. Wilson and Joan Petersilia, editors. San Francisco: ICS Press: 387–419. Bonzcar, Thomas P., and Allen J. Beck. Forthcoming. Lifetime Likelihood of Going to State or Federal Prison. Washington, D.C.: Bureau of Justice Statistics.

Brown, Jodi M., Darrell K. Gilliard, Tracy L. Snell, James J. Stephan, and Doris James Wilson. 1996. Correctional Populations in the United States, 1994. Washington, D.C.: Bureau of Justice Statistics.

Brooklyn Treatment Court Program. 1997. Eligibility Report, April 25, 1997. Unified Court District of New York, Kings County, NY. Bureau of Justice Assistance (BJA). 1996. National Assessment of Structured Sentencing. Washington, D.C.: U.S. Department of Justice.

Bureau of Justice Assistance (BJA). 1996. National Assessment of Structured Sentencing. Washington, D.C.: U.S. Department of Justice.

Caulkins, Jonathan P., C. Peter Rydell, William L. Schwabe, and James Chiesa. 1997. Mandatory Minimum Drug Sentences: Throwing Away the Key or the Taxpayers’ Money? Santa Monica, CA: RAND.

Cohen, Jacqueline, and Jose A. Canela-Cacho. 1994. "Incarceration and Violent Crime: 1965–1988," In Understanding and Preventing Violence: Volume 4, Consequences and Control, Albert J. Reiss, Jr. and Jeffrey A. Roth, editors. Washington, D.C.: National Academy Press.

Darity, William A. Jr., and Samuel L. Myers, Jr. 1994. The Black Underclass: Critical Essays on Race and Unwantedness. New York: Garland Press.

Farrington, David J., and Patrick A. Langan. 1992. "Changes in Crime and Punishment in England and America in the 1980s." Justice Quarterly 9:5–46.

Freeman, Richard B. 1996. "Why Do So Many Young American Men Commit Crimes and What Might We Do About It?" Journal of Economic Perspectives 10(1):25–42.

Gilliard, Darrell K., and Allen J. Beck. 1996. Prison and Jail Inmates, 1995. Washington, D.C.: Bureau of Justice Statistics.

Gilliard, Darrell K., and Allen Beck. 1994. Prisoners in 1993. Washington, D.C.: Bureau of Justice Statistics.

Grogger, Jeffrey. 1997. "Market Wages and Youth Crime." NBER Working Paper No. 5983. New York.

Harer, Miles. 1993. An Analysis of Non-Violent Drug Offenders with Minimal Criminal Histories. Washington, D.C.: U.S. Department of Justice.

Kleiman, Mark A. R. 1997. "Drug-Free or Unfree: To Get Heavy Users to Stay Clean, Link Parole and Probation to Abstinence." The Washington Post February 2, 1997: C03.

Levitt, Steven D. 1996. "The Effect of Prison Population Size on Crime Rates: Evidence From Prison Overcrowding Litigation." The Quarterly Journal of Economics May: 319–51.

Lott, John. 1992. "Do We Punish High Income Criminals Too Heavily?" Economic Inquiry 30:1307–16.

Lynch, James P., and William J. Sabol. Forthcoming. "The Use of Coercive Social Control and Changes in the Race and Class Composition of U.S. Prison Populations." The Howard Journal.

Lynch, James P. 1995. "Crime in International Perspective," In Crime, James Q. Wilson and Joan Petersilia, editors. San Francisco: ICS Press.

Lynch, James P. 1993. "A Cross-National Comparison of the Length of Custodial Sentences for Serious Crimes." Justice Quarterly 10(4).

Lynch, James P., and William J. Sabol. 1992. "Punishment Policy and Social Control: Incarceration and the Growth of the Underclass 1970–1990." Paper presented at the meetings of the Academy of Criminal Justice Sciences, Pittsburgh, PA, March 14.

Lynch, James P. 1988. "A Comparison of Imprisonment in the United States, Canada, England, and West Germany: A Limited Test of the Punitiveness Hypothesis." Journal of Criminal Law and Criminology, Spring.

Lynch, James P., Steve Smith, Helen Gradziadei, and Tanutda Pitthaythukin. 1994. Profile of Inmates in the United States and in England and Wales, 1991. Washington, D.C.: Bureau of Justice Statistics Special Report, U.S. Department of Justice.

Mauer, Michael. 1991. Americans Behind Bars: A Comparison of International Rates of Incarceration. Washington, D.C.: The Sentencing Project.

McGahey, Richard M. 1986. "Economic Conditions, Neighborhood Organization, and Urban Crime," In Communities and Crime, Albert J. Reiss, Jr., and Michael Tonry, editors. Chicago: University of Chicago Press: 231–70.

Perkins, Craig. 1994. National Corrections Reporting Program, 1991. Washington, D.C.: Bureau of Justice Statistics.

Petersilia, Joan, and Susan Turner. 1993. "Intensive Probation and Parole," In Crime and Justice: A Review of Research, Michael Tonry, editor. Chicago: University of Chicago Press: 17:281–335.

Roberts, Julian V. 1992. "Public Opinion, Crime, and Criminal Justice," In Crime and Justice: A Review of Research, Michael Tonry, editor. Chicago: University of Chicago Press: 99–180.

Rossi, Peter H., and Richard A. Berk. 1997. Just Punishments: Federal Guidelines and Public Views Compared. New York: Aldine de Gruter.

Sampson, Robert J. 1987. "Urban Black Violence: The Effect of Male Joblessness on Family Disruption." American Journal of Sociology 93:348–82.

Sampson, Robert J. 1995. "Unemployment and Imbalanced Sex Ratios: Race-Specific Consequences for Family Structure and Crime," In The Decline in Marriage Among African Americans: Causes, Consequences, and Policy Implications, M. Belinda Tucker and Claudia Mitchell-Kernan, editors. New York: Russell Sage Foundation.

Sullivan, Mercer L. 1989. Getting Paid: Economy, Culture, and Youth Crime in the Inner City. Ithaca, NY: Cornell University Press.

Shryock, J., and L. Seigel. 1973. Methods of Demography. Washington, D.C.: U.S. Bureau of the Census.

Tonry, Michael. 1992. "Mandatory Penalties," In Crime and Justice: A Review of Research, Michael Tonry, editor. Chicago: University of Chicago Press: 243–74.

Tonry, Michael, and Mary Lynch. 1996. "Intermediate Sanctions," In Crime and Justice: A Review of Research, Michael Tonry, editor. Chicago: University of Chicago Press: 20:99–144.

Waldfogel, Joel. 1993. "Criminal Sentences as Endogenous Taxes: Are They ‘Just’ or ‘Efficient’?" Journal of Law and Economics 36: 139–51.

Wilson, James Q., and George Kelling. 1989. "Making Neighborhoods Safe." The Atlantic 263:46-52.

Wilson, William J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press.

Zawitz, Marianne W., Patsy A. Klaus, Ronet Bachman, Lisa D. Bastian, Marshall M. DeBerry, Jr., Michael Rand, and Bruce M. Taylor. 1993. Highlights from 20 Years of Surveying Crime Victims: The National Criminal Victimization Survey, 1973–92. Washington, D.C.: Bureau of Justice Statistics.

About the Authors

James P. Lynch is an Associate Professor in the Department of Justice, Law, and Society at the American University. His fields of specialization include social organizations, communities and crime, and international comparisons of crime and justice-system responses. He has served on several advisory panels for the National Institute of Justice, the Bureau of Justice Statistics, and the American Statistical Association, and he was an editor of the Public Policy Series published by Georgetown University Press. Dr. Lynch's many publications cover a range of criminal justice issues and include "The Decomposition and Graphical Analysis of Crime and Sanctions Data," The Journal of Quantitative Criminology (forthcoming), and "The Use of Coercive Social Control and Changes in the Race and Class Composition of U.S. Prison Populations" (with William J. Sabol), The Howard Journal (forthcoming).

William J. Sabol is a senior research associate at The Urban Institute, where he directs several projects on criminal justice systems, including the Federal Justice Statistics Program (which produces the Compendium of Federal Justice Statistics for the Bureau of Justice Statistics, and a project on the impacts of law enforcement on community organization). Before coming to The Urban Institute, Dr. Sabol served as a senior social scientist at the U.S. General Accounting Office, where he conducted studies of the Federal justice system. His publications include "The Use of Coercive Social Control and Changes in the Race and Class Composition of U.S. Prison Populations" (with James P. Lynch), The Howard Journal (forthcoming), and "Racially Disproportionate Prison Populations: Historical Patterns and Contemporary Issues," Contemporary Crises.

Advisory Board

Adele Harrell
Blaine Liner
Shelli Rossman
Jeffrey A. Roth
William J. Sabol

Topics/Tags: | Crime/Justice

Usage and reprints: Most publications may be downloaded free of charge from the web site and may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact publicaffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.

Disclaimer: The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute.

Email this Page