IN THIS CHAPTER, I EXPLORE what the past forty years of sentencing reform, crime control, and the correctional boom have accomplished. One of the challenges in analyzing the effects of policy is assuring that the appropriate range of goals, outcomes, and impacts, both intended and unintended, have been identified. Some of the goals will be obvious, such as enhanced public safety, reductions in crime rates, and reductions in recidivism. Others perhaps less so, such as the challenges of prisoners returning to the free world after a period of incarceration, the variety of costs (both financial and collateral) of crime control, and the extent to which sentencing reform actually reduced sentencing discretion. I begin with a discussion of the impact of crime control on public safety.
DID IT WORK?
At first glance, the trends in incarceration and crime rates provide compelling evidence for success. Assuming that a primary goal was to reduce crime, some might look at the trends—increasing incarceration and declining crime—and conclude that punishment reduced crime. Consider the following from the popular media, a May 23, 2005, Forbes article by Dan Seligman entitled “Lock ’Em Up”:
A big story, inadequately memorialized by the media is that crime in America has become a much smaller story. … Counterintuitive as it might seem, this happy result came about via a massive social program. The program did not promote job training or administer therapy to thugs. Instead it consisted of putting them behind bars. … The connection of incarceration to crime is hard to ignore. The number of Americans in prison during 1984–2003 correlates –.71 with the number of violent crimes in in the country. That powerful negative coefficient says that increases in the prison population go hand in hand with declines in crimes committed.
And frankly, this conclusion makes sense. Punishment is intuitive, it is logical, and it is something easily understood. Punishment is something everyone has experienced, and it worked most of the time.
However, the world is just not that simple. In fact, looking at the evidence from a variety of vantage points, the unmistakable conclusion is that any public safety impact of crime control, sentencing reform, and the correctional boom is quite modest. While it is difficult to know what the expectations for public safety have been as a consequence of crime control policies, it seems plausible to assert that the crime reductions that can realistically be attributed to crime control are relatively underwhelming. That is not to say that crime rates have not declined substantially over the course of crime control policies. How much of those declines are due to tough on crime policies is rather limited. I now turn to the evidence.
THE CRIME DECLINE OF THE 1990s
The sustained decline in crime since the early 1990s was relatively consistent throughout the United States. That is, crime declined in all states and in all major metropolitan areas. As Michael Tonry notes (2004: 119) in his excellent analysis of the crime decline:
Crime trends in cities and states that adopted especially celebrated crime control policies can be compared with trends in comparable cities and states that did not. When such comparisons are made, it becomes clear that the comparable trends occurred everywhere. Including in cities that did not adopt aggressive zero-tolerance policing styles, and in states that adopted neither truth-in-sentencing nor three strikes laws.
The U.S. crime decline that began in 1992 was not a uniquely U.S. experience. Far from it. The International Crime Victimization Survey (ICVS) data (van Dijk, van Kesteren and Smit 2007) consist of periodic crime surveys of European and non-European industrialized nations. The report presents analyses of trends between 1988 and 2005 in property and personal crimes. The trends in the fifteen participating industrialized nations (United States, Canada, Australia, New Zealand, Belgium, England and Wales, Estonia, Finland, France, Netherlands, Northern Ireland, Poland, Scotland, Switzerland, and Sweden) are remarkably similar. While these nations have experienced varying levels or rates of offending, they share the rather dramatic and persistent decline in crime that began in the early 1990s. Those declines are comparable in relative magnitude to those experienced in the United States.
What can be learned from this? Analyses of international crime policies and practices (for example, Padfield, van Zyl Smit, and Dunkel 2010; Tonry 2004; Tonry and Frase 2001) show that none of the comparison nations implemented U.S. crime control policies. The simple conclusion is that whatever has driven the crime decline of the 1990s and 2000s, the outcome was not largely dependent on sentencing reform, crime control, and a correctional boom.
The comparative U.S. and Canadian experiences are particularly informative. In his book The Great American Crime Decline, Zimring (2007) devotes an entire chapter to the U.S.–Canadian comparison and notes: 1) the crime declines in both nations occurred at the same time and shared all of the relevant defining characteristics of size, breadth and length; and 2) the punishment policies of these two nations could not be more different—the Canadian incarceration rate varied between 98 and 116/100,000 between 1980 and 2000, while the U.S. rate tripled. The U.S. incarceration rate was 45 percent higher than the Canadian rate in 1980; by 2000, the U.S. rate was four times higher than the Canadian rate.
These comparative analyses of Canada and the European and non-European nations at least bring into question the wholesale assertion that the U.S. crime decline was largely a consequence of crime control policies, massive incarceration, and correctional control.
WHY SHOULD PUNISHMENT REDUCE CRIME?
Crime control is premised on the expectation that punishment will reduce crime and recidivism through the mechanisms of deterrence (specific and general) and incapacitation. I begin with a discussion of deterrence.
Deterrence theory identifies two types of deterrence: specific and general. Specific deterrence suggests that the experience of punishment will deter future offending on the part of the individual being punished. General deterrence posits that the likelihood of offending by others is a function of awareness of the punishment that follows committing a crime and getting arrested. When a court sentences in order to discourage others from engaging in crime, the intent is general deterrence.
Deterrence theory further stipulates that in order to deter, punishment must be swift, certain, and severe. The punishment must quickly follow the commission of the act, it must have a high probability of occurring, and it must be sufficiently severe.
General Deterrence
What is known about the effectiveness of general deterrence? Perhaps the most familiar version of this debate is the death penalty. As recently as the late 2000s, national media and conservative policy organizations have promoted the idea that the death penalty deters (Associated Press, June 11, 2007; The Wall Street Journal, June 21, 2002;; The Heritage Foundation, August 28, 2007). Despite recent statements by economists asserting the deterrent effect of the death penalty, the scientific consensus (Apel and Nagin, 2011; Bailey and Peterson 1999; Doob and Webster 2003; Radelet and Akers 1996) is clear—there is no valid, reliable scientific evidence supporting the conclusion that the death penalty deters homicide.
Moreover, despite over forty years of federal, state, and local policy based on the premise that punishment severity deters crime—after all, crime control has been about enhancing the severity of punishment—the conclusions from the scientific community are clear. While it may feel appropriate and justified under the circumstances when the District Attorney or judge announce that they are sending a message to the community by imposing a severe sentence, the research indicates either the community is not listening, or if they are, it does not alter their behavior.
Over a decade ago, an exhaustive review of a very impressive amount the scientific research (Doob and Webster 2003) focused on the general deterrent effect of sentence severity. The assessments of the scientific research on general deterrence of specific offenses, offense categories, or crime more broadly lead to the same conclusions, and those conclusions are shared by the vast majority of experts in the field. As Doob and Webster (2003: 187, 189) state:
Can we conclude that variation in the severity of sentences would have differential (general) deterrent effects? The reply is a resounding no. We could find no conclusive evidence that supports the hypothesis that harsher sentences reduce crime through the mechanism of general deterrence. Particularly given the significant body of literature from which this conclusion is based, the consistency of the findings over time and space, and the multiple measures and methods employed in the research conducted, we would suggest that a stronger conclusion is warranted. … The effects are consistent: the severity of sentence does not matter.
Does this mean that sanctioning of criminal offenders has no general deterrent effect? No, it does not. Crime rates are likely to increase if there are no consequences for offending. What it means is that the increasing severity of the sanction appears to be irrelevant as a general deterrent.
A more recent assessment of general deterrence by Apel and Nagin (2011) provides evidence that general deterrence can work, if the focus is on the certainty of punishment rather than severity. Moreover, the research clearly indicates that the deterrent effect of sanctioning is conditional on a number of factors, including the particular sanction used, the particular individual being sanctioned, and the jurisdiction. Apel and Nagin help refine the focus by directing away from questions like: Does greater sanction severity generally deter? (No.) Does general deterrence work? (It can.) Directing instead toward questions like: Under what conditions does a particular sanction deter, if at all? Is the sanction sufficiently cost-effective to warrant implementation?
Specific Deterrence
The other side of the deterrence argument is that harsher penalties, especially incarceration, will deter the future offending of the particular individual being punished. In essence, the negative effect of punishment will cause the individual to reconsider the next time around. In the aggregate, punishment or the threat of punishment raises the costs of engaging in crime and when the costs of crime increase, its frequency declines, thus harsher punishment leads to lower crime rates. At the individual level, the argument is in terms of reducing probabilities of reoffending based on the severity of punishment. As the costs to the individual increase, the likelihood of engaging in crime is diminished, all else equal. So the theory asserts. Let’s look at the evidence.
One incomplete, but telling bit of evidence regarding the ability to deter reoffending is recidivism of offenders released from prison. Recidivism is measured in a variety of ways: rearrest, reconviction, and reincarceration are typical. Reincarceration reflects the stronger measure or more serious version of recidivism because it requires an arrest, a reconviction, and a recommitment to prison or a violation of parole, and a subsequent revocation of parole and return to prison. It is also important to point out that recidivism is a measure of official response, meaning first that law enforcement must be aware of a crime (or a parole officer must be aware of a parole violation if the individual is under parole supervision). For that reason, official recidivism measures are underestimates of overall reoffending because crimes known to the police represent approximately 50 to 60 percent of crimes committed.
Recidivism data (rearrest and reincarceration) are available for offenders released from state prisons in 1983 and 1994 (Bureau of Justice Statistics 2002). These will serve as “baseline” statistics. The Pew Center on the States (2011) recently published a report on state-specific recidivism of inmates released from U.S. prisons in 1999 and 2004. All studies follow releases for three years subsequent to release.
It is important to emphasize that comparing the BJS and Pew recidivism data is just illustrative because they consist of different states reporting rearrest and reincarceration statistics. It is also important to note that these rates are aggregations of state rates and that there is considerable variation across states in recidivism rates, release laws and policies, revocation policies, and so on.
Comparison of the 1983 and 1994 release data with the 1999 and 2004 release data indicate the following four things:
1. Approximately 45 to 50 percent of inmates released in 1983, 1994, 1999, and 2004 were reincarcerated within three years, either for a new offense or for a technical violation of parole.
2. Reincarcerations for a new offense increased 12 percent between 2002 and 2007, the end years for the 1999 and 2004 Pew release cohorts. This was during a period of declining crime rates and increasing incarceration rates.
3. These recidivism statistics have remained essentially unchanged for the past thirty years, the period of crime control and the corrections explosion.
4. The most important predictors of recidivism are age (younger), number of prior arrests, and offense type (primarily property offenders).
While this is not a strong test of the impact of crime control, the fact that recidivism rates remained stable over time is suggestive. After decades of sentencing reform, increased incarceration, longer sentences, truth in sentencing resulting in greater time served, and dramatic increases in correctional control—all of which presumably were designed to reduce recidivism—it does not appear that the investment has produced noticeable dividends. Again these are aggregate statistics and thus do not reflect interstate variation, as well as varying recidivism probabilities associated with demographics, prior criminal history, and so on. However, the point is telling—it does not appear that the harshest form of punishment the United States metes out (short of the death penalty) deters all that much. It is not known what policymakers in years past expected with regard to the level of deterrence when they implemented crime control. There were no bold statements on the record that X percent drop in crime or Y percent drop in recidivism should be expected. However, it seems likely that the results seen today in terms of recidivism would be viewed as underwhelming.
One ironic discovery in reviewing the research on the specific deterrent effects of incarceration is how relatively little informative research has been conducted, especially farther back in time (Nagin, Cullen, and Jonson 2009). What is ironic about this is the fact that crime control was premised on the deterrent effect of harsher punishment. As discussed in chapter 1, the massive expansion of prison capacity, unprecedented increases in incarceration rates and correctional control, and dramatic changes in sentencing statutes all characterize and define U.S. criminal justice policy. All of this, with very little scientific evidence to support the purported and often articulated merits of these policies. We are not naive to the point of expecting policymakers to consult the scientific community regarding each policy decision made. Policy formation is a product of many influences and inputs. Scientific evidence is only one of those. However, for policies of the scale, expense, and longevity of those focused on here, it is a bit disconcerting that decision makers did not occasionally seek such guidance. Apparently, the intuitive, common sense appeal of crime control was sufficient to launch and sustain it for decades, despite mounting, contrary evidence.
Relatively exhaustive reviews of the scientific research literature on the specific deterrent effects of imprisonment have been completed over the past fifteen years or so (Gendreau, Goggin, and Cullen 1999; Nagin, Cullen, and Jonson 2009; Villettaz, Killias and Zoder 2006). These reviews and assessments of the state of the scientific knowledge impose screening standards for studies in their assessments (to assure their scientific validity and rigor) and carefully consider the scientific merit of the research. The most informative research is comparative or controlled, contrasting recidivism outcomes for individuals receiving a sentence of imprisonment versus a noncustodial sentence such as probation. On balance, three things are known. First, there is a remarkable lack of scientific support for the assertion that harsher punishment deters those who experience it. Second, compared to noncustodial sentences, incarceration is either no different in terms of future offending (that is, recidivism), or has a crime-promoting or criminogenic effect (Bales and Piquero 2012; Nagin, Cullen, and Jonson 2009). Finally, there is sufficient evidence to give scientific credibility to the common assertion that criminals who go to prison typically come out even worse.
In terms of the policy relevance of these findings on deterrence, a discussion I will return to in detail in the subsequent chapters, Nagin, Cullen, and Jonson (2009: 183) echo the concerns of the scientific community in their concluding remarks:
There is little convincing evidence on the dose-response [severity of punishment] relationship between time spent in confinement and reoffending rate. Because imprisonment is so costly, $30,000 per person-year or more, knowledge of this relationship has important implications not only for public safety but also for the state, local and federal budgets.
So how does one make sense of the failure of punishment to deter criminal offending and recidivism? How does one reconcile the fact that science smacks right up against common sense and individual experiences with punishment?
There are a few ways of understanding the lack of a general deterrent effect of punishment severity. Most of us are law abiding (exceptions for the occasional traffic ticket, ordinance violation, or minor misdemeanor). Thus it is reasonable to suggest that for most Americans, the severity of the punishment is irrelevant because they are disinclined to engage in crime in the first place. Many Americans have grown up and live as adults in a world in which crime is neither a desirable nor necessary option.
For others, who are more crime prone, the null effect of sentence severity might be understood in terms of the series of predicate events that are required before an offender is punished: law enforcement must become aware of the offense, then must apprehend/arrest the offender, then the prosecutor must indict, adjudicate, convict, and then the court must impose a severe sentence. The odds are generally in the offender’s favor. First, only 50 percent of violent crimes and 40 percent of property crimes are actually reported to law enforcement. Of those reported, 47 percent of violent crimes are cleared (that is, lead to an arrest) and 18 percent of property crimes are cleared. Obviously, there is considerable variation in arrest rates by type of felony, a reflection largely of seriousness. The likelihood of reporting a crime to the police is largely a function of its seriousness, as is the likelihood of an arrest. Moreover, the vast majority of indictments are plea negotiated and, most typically, it is a form of charge bargaining in which the offender pleads guilty to a lesser charge in exchange for less punishment. Offenders know how this system works. They may not have precise probabilities at hand, but in all likelihood, they have some general impressions about their odds of getting caught and punished. Given these probabilities, it is generally true that crime pays.
Moreover, there is the assumption (at least implied in deterrence theory) that individuals would: 1) have access to information regarding outcomes like conviction and sentences; and 2) make decisions in a rational manner. It is probably reasonable to suspect that neither of these conditions is uniformly present among the offender population.
What about the apparent lack of a specific deterrent effect? Why is it that personal experiences and logical reasoning do not seem to apply to offenders considered to have been harshly punished by the justice system? Once again, there are several explanations that are relevant. One is that the punishment is not sufficiently harsh. This line of reasoning leads to the conclusion that it is not the theory that is faulty, but the implementation of it. There are a number of problems with this argument. First, there is no scientific evidence to support it. In today’s political and fiscal climate, it would be extraordinarily difficult to muster much support for going down that road. Moreover, there is always the concern with constitutional challenges (Eighth Amendment cruel and unusual punishment issues, equal protection issues, due process issues).
A more reasonable explanation for the disconnect between the common sense of deterrence and the reality is that while nonoffenders can apply the threat or imposition of punishment in their own lives and rightfully conclude that it would work for them, this logic misses the fact that typical offenders are not like nonoffenders. Research has consistently and decisively shown that there is a very strong link between poverty and disadvantage and criminal involvement. Criminal offenders typically do not have the resources, the opportunities, and the alternatives that nonoffenders have. Nor do they typically have access to the education and employment resources that many Americans have. Drug and alcohol abuse are prevalent among the offender population, as is mental illness. Most have prior criminal justice involvement, which creates additional barriers to desistence from crime. The point is that with the typical barriers and challenges that offenders face, it is not surprising that they view crime as a viable, often necessary option.
In most states and for some federal programs, there is usually an ongoing effort to punish offenders well after they have been discharged from the justice system by, for example, preventing them from being employed in certain types of jobs and obtaining certain types of occupational licenses, denying access to things like public assistance and public housing, Pell grants for educational advancement, and health care through Medicaid. Research reported by the Pew Center on the States (2010) indicates that before incarceration, two-thirds of male inmates were employed and one-half were the primary source of support for their families. After release from incarceration, inmates work nine fewer weeks annually and earn 40 percent less than prior to incarceration. Moreover, society continues to punish by restricting where ex-inmates can live and work. Landlords deny housing to individuals who do not pass the criminal background criteria set up by management, and employers deny employment to individuals with criminal backgrounds (despite the fact that there are laws in place preventing such actions). The point is simply that once someone enters the justice system, his or her options, opportunities, and alternatives diminish rapidly, simply as a consequence of their justice system involvement. These restrictions and ongoing barriers tend to further perpetuate a life of crime.
This is not an apology for criminal offenders. Instead, the reality of the criminogenic circumstances and deficits that offenders present with, combined with the additional limits and restrictions the justice system and society place on them, begs the question: What do we expect? What is expected of a parolee who has few resources, limited family ties, employment and housing challenges, and substance abuse problems, for whom there are few resources or forms of assistance available either from the parole system or the public sector? Is it really surprising that many offenders return to crime after they have been punished? Is it really surprising that harsh punishment does not work? Again, this is not an exercise in feeling sorry for criminal offenders. Instead, it is a realistic look at a system that is premised on several very basic principles, a system that just seems to miss a very fundamental point. When someone has limited options, it is rational that they take the one that seems most available and viable. If anything, it is surprising that official recidivism rates are not higher than they are.
A similar reasoning applies to specific deterrence as to general deterrence regarding the probability of experiencing subsequent punishment. This involves the reoffender’s perception of the likelihood of getting caught. Deterrence theory holds that in order for punishment to deter, it must be swift, certain, and severe. The certainty of punishment is contingent on the perceived probability of apprehension, which is somewhere well south of 100 percent. Moreover, there is not much about processing criminal defendants that can be characterized as swift. Absent the ability to address the swift and certain attributes of deterrence, it probably makes sense that policymakers chose severity (if they are or were aware of deterrence theory at all).
Finally, punishment may not specifically deter for the simple fact that punishment does not target the reasons individuals engage in crime, and therefore punishment does not alter behavior. As Cullen and Gendreau (2000: 146, 154, 155) note early on in the scientific scrutiny of crime control policies:
A central policy issue is whether the movement to “get tough” on crime has enhanced public safety. In particular, the massive rise in the prison population … has created an intense interest in whether the extensive use of imprisonment has a meaningful incapacitation effect, or, at the aggregate level, deterrent effect. Punishment does not target criminogenic needs and thus is one of the most ineffective, if not counter-productive, strategies … there is no evidence that punishment-oriented “treatment” programs specifically deter or otherwise reform offenders.
Perhaps thirty years ago it was reasonable for state and federal officials to make promises that enhanced punishment will pay crime reduction dividends because it will deter reoffending. So too the tough on crime prosecutor who recommends a maximum sentence in order to send a message to the community and to a specific offender that this kind of behavior will not be tolerated. Those assertions and the subsequent implementation of tough on crime punishments had little basis in terms of scientific evidence pointing to their likely success, but at the same time, there was little scientific evidence indicating that they would not work. Certainly, there are examples of individuals who were deterred from reoffending because of the punishment they received. There are individuals who have decided not to engage in crime because of the threat of punishment. Nevertheless, in the aggregate, these deterrence strategies are ineffective.
Even if the evidence pointed toward a crime-deterring effect of severity, there is little basis for knowing or even determining how severe the punishment must be in order to produce a marginal deterrent effect. The amount of punishment has been determined for decades in terms of harm perpetrated by the offender, not generally in terms of any particular understanding of the individual’s circumstances that might inform the proper dosage of punishment. The severity of punishment is in the eyes of the giver (the sentencer) and the receiver (the convicted offender), and it is presumptuous to think that the appropriate level of severity could be determined on a case-by-case basis.
So is there any reason not to just bury the deterrence argument and move on? Yes. Research indicates that punishment may deter if it is swift and certain (Farabee 2005; Grasmick and Bryjak 1980; Nichols and Ross 1990; Paternoster 1989; Rhine 1992; Taxman, Soule, and Gelb 1999). In practice, the swift and certain attributes of sanctioning or punishment appear to be producing enhanced compliance and accountability among probationers. The HOPE Court, developed in Hawaii in 2004, is based on the premise that deferred and low-probability threats of severe punishment are less effective than immediate and more certain threats of mild punishment. Probationers participating in the HOPE Court program receive an initial warning by the court that spells out what will happen when anyone violates any conditions of probation. When a violation occurs, the offender is immediately brought before the judge and typically a short sentence in jail is imposed and the individual is immediately taken to jail. The length of jail time is increased for successive violations. Evaluation research indicates dramatic reductions in rearrest, drug use, missed appointments with the probation department, and revocations. I will discuss the HOPE Court model in more detail in later chapters.
In 2009, Mark Kleiman published a book entitled When Brute Force Fails. The title is a bit misleading, as some readers probably figure this is a book about alternatives to deterrence-based approaches. Instead, the premise of Kleiman’s book and the premise of subsequent discussions along the same lines (for example, Durlauf and Nagin 2011) is that a greater police presence can deter crime by increasing the perceived certainty and swiftness of punishment. The logic is based in part on the well-established lack of a general or specific deterrent effect of the severity of punishment and the criminogenic effect of incarceration (Apel and Nagin 2011; Nagin, Cullen, and Jonson 2009). The recommendation is to shift resources away from severe incarceration sentences to a greater police presence. Reducing the severity of punishment is evidence based. Increasing the presence of police is less so. As Tonry (2011) notes, the evidence is problematic regarding the crime-reducing impact of police crackdowns, zero-tolerance policing, and increased police presence. The coincidence of police initiatives and crime rate declines is at least partly spurious, as crime began declining before the police initiatives were introduced. Tonry states (2011: 148): “Fifteen years later, the clear weight of opinion concerning these [certainty enhancing] policies [the three-strikes law in California and zero-tolerance policing in New York City] is that they do not deserve major credit for the crime declines in those places.” Moreover, there is significant concern about whether the findings of drug market crackdowns and hotspot policing are generalizable across time and space, and whether there are sizable crime-reduction effects net of displacement. As Tonry (2011: 148) concludes “If crackdowns have no long-term effects, then they are little different from castles of sand built on ocean beaches: the waves will wash them away as if they never existed.” Tonry expresses substantial reservations about the potential civil rights and racial profiling concerns with this strategy. He cites law enforcement’s track record in this regard for a lack of confidence in mitigating these concerns. Finally, there is little in Kleiman’s strategy that results in changes to major criminogenic circumstances, so it is hard to know how longer-term reductions in crime are achieved.
In short, there is considerable appeal to the concept of implementing strategies that increase the certainty and celerity of punishment. It is intuitive and there is research to support a deterrent effect. There are some programs that are based on this premise that appear to be effective and are gaining widespread attention in research and policy circles (Hawken and Kleiman 2009). There is mounting evidence that the certain and swift sanction model, in conjunction with therapeutic interventions, have very promising potential. However, absent the therapeutic intervention, swift and certain punishment is unlikely to change behavior in the longer term.
Incapacitation
Crime control also promotes the incapacitation function of correctional control, especially incarceration. Incapacitation is the removal of criminal opportunity, the effect of preventing criminal offending in the free world while an offender is incarcerated. The intuitive appeal of incapacitation is compelling. Lock up the criminal, lock up the crime.
The scientific evidence on incapacitation is consistent in one respect: the incapacitation effect is relatively small. Where the evidence is inconsistent is in terms of how small (somewhere in the 10 to 25 percent range, meaning incarceration accounts for somewhere between 10 and 25 percent of crime declines over the past twenty years). In a thoughtful review of the research, and assessment of a variety of statistical and methodological concerns and challenges, Spelman (2000: 485) concludes:
In a nutshell then, what the studies of the past ten years tell us is that crime responds to prison capacity and that continued expansion of prisons nationwide will reduce the crime rate. What the studies don’t tell us is whether the reduction is large enough to warrant continued expansion.
A recent analysis by Western (2006) concludes that incapacitation accounts for 2 to 5 percent of the reduction in serious crime during the 1990s and 10 percent overall. Three studies have addressed some of the more compelling methodological difficulties of prior research (Spelman 2000, 2005; Levitt 1996) and their estimates of the incapacitation effect converge (Steman 2007). The range of estimates produced by these studies is 2 to 4 percent, meaning that a 10 percent increase in incarceration will produce a 2 to 4 percent reduction in crime. Zimring (2007) argues that the incapacitation effect is problematic in light of the comparison of the U.S. and Canadian crime declines during the 1990s. As discussed earlier, both nations experienced similar crime reductions, but Canada did not dramatically expand incarceration.
Some research further indicates that the scale of incarceration matters in that as scale increases, the marginal return of reduced crime erodes dramatically (Liedka, Piehl, and Useem 2006). Put differently, incarceration does appear to have a modest crime-reducing effect at low levels, but as incarceration increases, the negative effect on crime is diminished.
Another consideration regarding the presumed crime-reduction effect of incapacitation has to do with the extent to which criminal offending is eliminated when an offender is removed from the street and incarcerated. It is the idea that the incapacitation effect, to the extent that it exists, is dependent at a minimum on the type or nature of crime that was being committed. For example, in the case of gang-related or organized crime, simply incarcerating a gang member may not eliminate the crime. It is reasonable to expect that the crime for which the incarcerated gang member was responsible will continue, simply being committed by another gang member.
A similar argument applies to the drug trade. Taking a street-level dealer off of the street will probably not disrupt the distribution of drugs. That incarcerated dealer will in all likelihood be replaced by one of any number of dealer wannabes who have been waiting for the right opening or opportunity. There is clear ethnographic evidence that this is in fact what happens (Johnson et al. 1990; Padilla 1992; see also Blumstein 1994). The overall caution here regarding incapacitation effects is that they are conditional, in part depending on the nature of the crime. Some criminal activity is like a labor market. If a worker is removed, another takes his or her place. The irony is obvious. Prison sentences for some mean opportunities for others.
Strictly speaking, incarceration does not always prevent or incapacitate those in prison from being involved in crime on the outside. Prison gangs are extraordinarily prevalent in the nation’s prisons, and many have street equivalents. These gangs are organized by race or ethnicity. White gangs include the Aryan Brotherhood, the Nazi Lowriders, the Aryan Circle, Dead Man Incorporated, and the Confederate Knights of America. Typical Hispanic gangs are the Mexican Mafia (or La Eme), Nuestra Familia, Barrio Azteca, and the Texas Syndicate. Black gangs include the Crips, the Bloods, the Black Guerilla Family, United Blood Nations, and Folk Nation. While these gangs provide protection for their members while incarcerated, they are primarily criminal enterprises that have extensive connections in the free world and are often heavily involved in drug trafficking and money laundering, among others. While they may not be physically committing free-world crime themselves, members of prison gangs can influence and/or direct crime on the outside.
Enthusiasm for the incapacitation effect of incarceration is reflected in mandatory sentences, especially habitual offender laws (commonly known as three strikes laws). Such repeat offender laws are, in part, a statutory attempt to implement selective incapacitation—a strategy based on the fact that the majority of crime is committed by a relatively small number of repeat offenders. Selective incapacitation is the attempt to incarcerate those offenders responsible for the bulk of crime. One substantial challenge to selective incapacitation is that it is difficult to accurately identify these offenders. The success of selective incapacitation depends on the presumption that an offender will continue offending and therefore incapacitating him or her will reduce crime. Not only is there an error rate in selective incapacitation (false positives deprive individuals of liberty for lengthy periods of time, when in fact they are not the repeat offenders who were targeted), there are errors involved in the implementation of the repeat offender or three strikes laws. The likelihood of those false positives combined with cases in which the mandatory sentence was radically inconsistent with the conviction offense has led to a reduction in the use of such laws in most states.
Another consideration associated with mandatory sentences and the incapacitation effect is as the prison population ages as a function of length of incarceration, we run the risk of incarcerating larger and larger numbers of inmates who have aged out of offending. To the extent this is the case, an incapacitation effect should not be expected from those individuals who are taking up space that might be more productively used for offenders still in the crime-prone age span. Recent data indicate that between 2007 and 2011, the number of U.S. prison inmates ages sixty-four and over increased at a rate ninety-four times the rate for those ages in the general, free-world population. The prison population is graying at a very rapid rate, a trend that further limits the ability of incarceration to effectively and cost-effectively incapacitate. As increasing numbers of older inmates are incarcerated, the marginal ability to incapacitate declines.
So part of the disconnect between the intuitive, common sense appeal and the scientific reality of the incapacitation effect is that removing an offender from the street does not always remove the crime in which he was engaged. The disconnect is also a function of how well the justice system accurately identifies who needs to be incapacitated. This raises what will become a common concern as I move forward in my discussion. Effective sentencing (in a utilitarian sense) is really a matter of accurately sorting offenders into one of a limited number of categories: diversion, probation, prison, and so on. Part of the failure of incapacitation to live up to expectations, whatever they may have been, may reflect a historic inability to make optimal, utilitarian decisions about who should go into which category. Clearly, although there may be merit to the incapacitation function of incarceration, the implementation has not been terribly effective.
THE CONSEQUENCES OF SENTENCING REFORM
In this section, I focus on the consequences of two pivotal aspects of sentencing reform: the implementation of mandatory sentences and the shift to determinate sentencing. These attributes of sentencing are important for reform because absent such changes, crime control and the corrections explosion would have been mere shadows of themselves. I begin with mandatory sentencing.
Mandatory Sentencing
In 1970, Congress repealed virtually all of the mandatory minimum drug sentences that had been implemented some twenty years earlier under the Boggs Act of 1952. This repeal of mandatory minimums was based on the observation that “lengthening prison sentences had not shown the expected overall reduction in drug law violations” (quoted in Mascharka 2001). It was a short fifteen years later that Congress reversed course in the Sentencing Reform Act, which paved the way for a variety of federal mandatory minimum sentences, many of which are aimed at drug law violations. By 1994, all fifty states had at least one mandatory sentencing law in place. The Anti-Drug Abuse Act of 1986 instituted new mandatory sentences for a variety of drug law violations. The Rockefeller Drug Laws, the quite severe mandatory sentences for drug offenses in New York, were put in place in 1973. Michigan followed suit in 1978. Over the next twenty years, thirty-seven states implemented harsher drug law provisions. At the same time, opposition to mandatory minimum sentences was broad, including the Federal Sentencing Commission, the Judicial Conference of the United States, the Federal Courts Study Commission, the Federal Judicial Center, the American Bar Association, and the vast majority of federal judges, as well as Supreme Court justices Rehnquist, Kennedy, and Breyer (Mascharka 2001).
Mandatory minimum sentences and other mandatory sentences, such a habitual offender laws, have been in place for decades and constitute a characteristic element of the U.S. criminal justice landscape. Mandatory sentences have been an essential mechanism for accomplishing the goals of crime control because they enhance the punitiveness of criminal sentences by increasing the number of offenses for which incarceration is the sentencing outcome (therefore increasing admissions to prison), as well as length of incarceration, increasing both the up-front sentence as well as time served. The evidence supports the conclusion that one of the important consequences of the imposition of mandatory sentences is growth in incarceration rates. Stemen, Rengifo, and Wilson (2006) conclude that states with more mandatory sentences had higher incarceration rates. Truth in sentencing reforms also appear to be related to increases in inmate populations in the states. Research by Sabol et al. (2002) shows that states that implemented truth in sentencing laws experienced increases in prison population, however they caution that this effect of truth in sentencing laws is tied to broader sentencing reforms that states implemented, such as mandatory sentences.
But what about public safety? What is the impact of mandatory sentencing on crime and recidivism? Once again, Michael Tonry (2009) has provided us with an excellent review and critique of the existing research on the deterrent effects of mandatory sentences. Referencing three sources of evaluation research—National Advisory bodies, surveys of the research literature, and evaluations and impact assessments—Tonry underscores the convergence of the conclusions that there are no detectable deterrent effects of mandatory sentences. Because mandatory sentences focus on severity of punishment exclusively, one would not expect a deterrent effect based on the research on punishment severity and deterrence. Several National Academy of Sciences panels, research by the British Home Office and the Canadian Sentencing Commission, to name but a few, all concur. In fact, it is ironic that the Home Office and Canadian Sentencing Commission recognized in the late 1980s and early 1990s that “deterrence cannot be used with empirical justification, to guide the imposition of sentences” (Canadian Sentencing Commission 1987: xxvii). The research communities’ assessment of the marginal deterrent effect of increasing punishment severity through mandatory sentence enhancements is consistent with that of national advisory bodies: “the effects of severity [of punishment] estimates and deterrence/sanctions composites, even when statistically significant are too weak to be of substantive significance” (Pratt et al. 2006: 379). Finally, with regard to individual state-level evaluations, Tonry (2009: 95) concludes: “No individual evaluation has demonstrated crime reduction effects attributable to enactment or implementation of a mandatory minimum sentence.”
What about an incapacitation effect of mandatory sentencing? In theory, this is a plausible expectation. However, many of the mandatory sentences are for drug law violations, and removing drug dealers from the street typically does not remove the offending. Moreover, mandatory sentences, especially habitual offender laws, are an attempt at selective incapacitation—lock up the career offenders for a lengthy incarceration and crime reduction benefits should be realized. The problem is that a statutory mechanism for implementing selective incapacitation is fraught with error. Just because the three strikes laws work well as a baseball metaphor does not mean they are effective at capturing true habitual offenders. What is the logic behind using a law rather than clinical assessments to determine who is more or less likely to reoffend? What is magical or predictive about three? And then there is the inefficiency at the other end of the sentence, where there is an increasingly aging inmate population. Twenty-one percent of prison inmates in the United States are serving a sentence of at least twenty years; 10 percent are serving a life sentence. With mandatory sentences, offenders are kept well past the age of desistence from offending.
A RAND study published in the late 1990s (Caulkins et al. 1997) focused on assessing the relative cost-effectiveness of mandatory minimum drug sentences. The researchers found that mandatory minimum drug sentences were not justified on cost-effectiveness grounds. Their research indicated that incarceration, whether through mandatory minimum sentences or indeterminate sentences, is less cost-effective than drug treatment for problem users. More on cost-effectiveness later.
Determinate Sentencing
It appears that the conclusions are the same when the movement to determinate sentencing is considered. There were two primary drivers of determinate sentencing: to reduce leniency in sentencing due to judicial discretion and to reduce sentencing disparity/unfairness due to judicial discretion. Regarding the former, the jury has rendered that verdict. There is little evidence to support the assertion that judges tend to be too lenient. Secondly, in the aggregate, enhanced sentence severity does not deter either generally or specifically and if there is an incapacitation effect, it is minimal.
Where are we with regard to sentence disparity? Many observers (including Alschuler 1979; Engen 2008; U.S. Sentencing Commission 1991; Forst and Bushway 2010; Mascharka 2001; Miethe 1987; Piehl and Bushway 2007; Tonry 2009; Vincent and Hofer 1994) have pointed to the obvious and troubling consequence that sentencing reform did not eliminate sentencing discretion, it simply displaced it from the bench to the prosecutor. Determinate sentencing and mandatory sentences remove much of the latitude from the judge at sentencing, but the sentencing-relevant characteristics of the case are largely established by the prosecutor through decisions such as what charges to indict, what evidence to use, what to negotiate in the inevitable plea bargain, and what to recommend at sentencing. Jurisdictions that restrict discretion by judges simply enhance the discretion that prosecutors exercise in influencing sentencing outcomes. One of the key concerns expressed by the shifting of discretion to prosecutors is that decision making moves from the “neutral, fair adjudicator” to the adversarial lawyer who represents the state. I shall return to this issue of prosecutorial discretion in chapter 4.
On the one hand, crime control has a very logical, familiar, intuitive basis. It just feels right. It seems just, deserving, appropriate. If someone errs, we do what we do with children: we punish. Punishment may not work the first time, but eventually this approach will probably produce the intended results. It worked for us, it works in our everyday lives, thus it has real-world validation.
And therein lies the problem. Criminal offenders are not “us.” The wisdom of crime control fails when similar results are expected from individuals whose lives are significantly different. Most criminal offenders do not enjoy the opportunities and options that most of us have and that policymakers and legislators who write the laws have.
There was very little scientific basis for the efficacy of crime control. In fact, there is evidence that sentencing policy developed, at least in part, based less on scientific evidence and more on logic and personal preference. Hoelter (2009: 54) relays the experiences of one Federal Sentencing Commissioner who lamented the fact that sentencing policy was based on personal preference and opinion, rather than scientific evidence.
The personal preferences of sentencing commissioners as to what is “good” or “right” or “just” should not be the basis for the Commission’s policy decisions. The basis for those decisions must be information, and information of the costs and benefits of various policy options. What concerns me about these unsupported amendments is not only that the substantive changes may not be warranted but also that the Commission’s process for generating guideline amendments is developing in such a way as to hinder rational policy-making.
One can appreciate the difficult situation that gave traction to crime control in the late 1960s and 1970s and even the 1980s. But at some point, the wisdom of crime control came into question, as the simple argument of “incarceration goes up and crime goes down means it works” was no longer credible to those who looked carefully at the issues.
Thirty years ago phrases like “lock ’em up and throw away the key” and “do the crime, do the time” were common. The states and the federal government seemed as if they could not build enough prisons. Rural communities competed for prisons because they were viewed as an economic boon, and in many respects they were. However, of late, it seems some Americans have awakened after a thirty-five year sleep with a really bad hangover and the realization that not only has this experiment failed, it has wasted enormous amounts of public resources, jeopardized public safety, and resulted in hundreds of thousands of avoidable victimizations.
A 2004 study of felony defendants from the seventy-five largest metro areas demonstrated what many observers already knew, that 75 percent of the felony defendants had a prior arrest history, and over 50 percent had at least five prior arrests (Kyckelhahn and Cohen 2008). It is truly a revolving door. What policymakers appear to ignore, or at least not appreciate sufficiently, is that each time someone reoffends and law enforcement arrests them, the public cash register rings once again, as costs are incurred by law enforcement, the jail, the prosecutor, the judge at a preliminary hearing, pretrial services, the grand jury, the arraignment judge, defense counsel (more than likely assigned counsel or a public defender), the trial court, and corrections. Moreover, ineffective policies result in future victims being irresponsibly placed in jeopardy. It is as inappropriate as it is unnecessary.
