The Math and Science of Decarceration
Summary and Keywords
Despite a growing consensus that “mass incarceration” in the United States has reached unacceptable levels, there has been little movement in its decline. National imprisonment rates seem to have stabilized and will remain so absent a major decarceration effort. To implement such a decarceration effort requires a strategic plan that will lower prison admissions and lengths of stay for all prisoners—especially those convicted of violent crimes. It will also need to reduce the more pervasive nature of other forms of correctional control (jails, probation, and parole). Such a strategy, which relies upon current and past policies, is entirely feasible. But to take hold on a national level, the plan must negate economic and public safety concerns that favor maintaining high imprisonment and correctional control rates.
The Need for a Decarceration Plan
Criminologists and prison reform advocates have done a good job in detailing the historic increase in the U.S. incarceration rates. In the United States, incarceration entails local jails, state prisons, and federal prisons. As shown below, all three populations as well as probation and parole populations have risen dramatically over the past four decades. While lamenting the state of “mass incarceration” there has been, with a few exceptions, less success in reducing the any of these correctional populations. There seems to be a new consensus that crosses traditional political lines that the level of “mass incarceration” is excessive and should be reduced. Former incarceration advocates, such as John DiIulio, former U.S. Attorney General Edward Meese, Congressman Newt Gingrich, Grover Norquist, and other conservatives are joining liberals in challenging the wisdom of imprisonment policies they formerly embraced (Sullum, 1999; Head & Norquist, 2015; Gingrich & Van Jones, 2014). But there is far less consensus on how to significantly lower prison and jail populations.
Criminologists and other advocates have been unable to develop a comprehensive strategic plan that (1) is acceptable to policy makers and the public, and (2) would significantly lower the current rates of incarceration without jeopardizing public safety. The most recent approaches of justice reinvestment and re-entry have produced, at best, modest results in reducing state prison populations (Austin et al., 2013). The reasons for the absence of such a strategy can be traced to a better understanding of the math and politics of decarceration.
The two major sticking points are (1) public safety and (2) economics. The former is the widespread belief that the rise in state prison populations was effective in lowering crime rates. The latter recognizes that shrinking the current criminal justice system could entail the loss of jobs in that industry. The “science” of decarceration would be used to craft policies, laws, and incentives that would lower incarceration rates that are consistent with current crime rates.
Advocates of decarceration have paid too little attention to those key quantifiable factors that produced the sharp rise in U.S. imprisonment rates.1 Consequently, their strategies are uniformly minimal in their projected impact on the state prison population. Prisons and jails (and any confined population of interest) are produced by two metrics—admissions per year and their length of stay (LOS). Significant reductions in both of these drivers are required to significantly reduce a prison or jail population. Clearly, it is not difficult to determine by how much the current rate of prior admissions and/or by how much the current LOS needs to be reduced to significantly lower the U.S. prison population. But the devil is in the details. More specifically, what kinds of people should not be diverted from prison or not have their LOS reduced?
The fear of decarceration is the requirement that any large-scale effort must ensure there will be no increased crime rates and worsened public safety. Often decarceration plans are held to an unreasonable standard that no impact on crime rates can occur which, on its face, is impossible to achieve. Yet for the United States to return to a more normal incarceration rate, the public will need to be assured that public safety will not worsen and may even improve. The inability of criminologists to provide such assurances to the public allows critics of decarceration to repeatedly assert that any effort to reduce the current incarcerated populations will have a direct and significant impact on crime rates. Until this standard is discredited, efforts to decarcerate the U.S. imprisoned populations will lack some level of public support.
Finally, there are the economics of decarceration. There is a considerable amount of money invested in not only the prisons but in all other components of the criminal justice system. The current estimate is that over $265 billion is being spent on police, the courts, jails, probation, parole, and prison systems. It directly employs over 2.5 million people in federal state and local jobs (Kyckelhahn, 2015). This number does not include the number of people employed by private prisons and businesses that provide medical, mental health and other support services. Academic institutions, think tanks, consulting groups, and a wide variety of businesses profit off of the current criminal justice system. Any effort to shrink this level of spending, which must occur under any serious decarceration effort, will be resisted.
Understanding the Sources of Rising Incarceration and Correctional Control Rates
In order to identify effective methods of decarceration, one needs to understand the origins of those factors that paved the way for rising incarcerations rates. There were three “mega quantitative” factors that created the current level of incarceration—(1) rising crime rates, (2) increasing prison admissions, and (3) longer lengths of stay. But the rise also had the strong and persistent support of its two major political parties.2
Beginning in the 1970s, the United States embarked on a three-decade-long shift in its sentencing and penal policies. Prior to 1970, there had been, at best, modest changes in the rate of incarceration. Indeed, there had been some reductions in periods of war, when large numbers of males had been drafted to serve in WWII, the Korean War, and the Vietnam War (see Figure 1). But these fluctuations were relatively mild compared to what has occurred since 1970, after which the prison population grew from less than 200,000 to nearly 1.6 million by 2008.
Figure 1 also plots the U.S. crime rates for the same time period. It is instructive to note that the rapid increase in the crime rates predates the rise in prison incarceration rates. Prior to the mid 1960s, crime rates generally remained below the 2,000 per 100,000 population level. But beginning in 1965, they began a steady increase that did not abate until 1980 when they peaked at about 6,000 per 100,000 population. By contrast, prison (state and federal) incarceration rates remained relatively steady at the 100–110 per capita rate until 1975 some ten years after the crime rate began to climb.
Figure 1 also shows the dramatic decline in crime rates that began in 1995 and have now reached a level that matches the 1967 crime rate. In that year (and for decades prior to 1967) the incarceration rate was at the 100 per capita rate level. But as crime rates have plummeted, incarceration rates have remained high and stabilized at the 470 per 100,000 rate. So while crime rates may have justified the greater use of incarceration, they did not initially cause imprisonment rates to rise, nor have their sharp decline since 1995 been associated with a decline in federal and state imprisonment rates.
It is not just the state and federal prison systems that have experienced such massive growth. The three other major components of the correctional system have grown at a similarly rapid pace. Table 1 shows the relative changes in all four correctional populations dating back to 1980. Overall the entire correctional system has increased from 1.8 million to 7.0 million people on any given day.
The overall increase in the entire correctional system is instructive on a number of levels. First, the large increases in parole, jail, and probation demonstrate that the rise in the prison population was not the result of people being shifted from other systems of control to state and federal prisons. Rather, imprisonment was only one small part (22%) of the entire correctional system, which also grew dramatically.
Table 1. Changes in the U.S. Correctional Populations, Population, Crime, and Arrests, 1980–2014
% Change 1980–2014
Total Correctional System
Reported Index Crimes
Sources: Bureau of Justice Statistics, U.S. Census, and Federal Bureau of Investigation.
Second, the fact that other forms of correctional control are more pervasive and also grew as dramatically as state and federal prison populations raises questions about efficacy of previous studies that claimed incarceration rates were responsible for a significant portion of the dramatic declines in crime rates (Levitt, 1996; Nagin, 2013; Spelman, 2006; Tonry, 2015). These studies only included changes in state incarceration rates in their regression models and not the other forms of correctional supervision. Had they done so, the relative contribution of higher incarceration rates in reducing crime rates since 1995 would no doubt have been suppressed as the absolute numbers of people that have been added to probation, parole, and local jails clearly exceeded the increases in state and federal prisons.
Third, the size of what is essentially a “snapshot” of the current correctional system as represented in Table 1 severely underreports the number of people who experience some form of imprisonment and/or invasive correctional control each year. To better appreciate the extent to which these four correctional systems impact the U.S. adult population one needs to measure the flow of people through these control systems. There are nearly 15 million admissions into the four major correctional systems each year. The largest contributor to the correctional system admissions stream is the jail, which reports over 11 million people per year (Table 2).
Undoubtedly, some of these admissions are the same people who in a given year were arrested, booked into jail, and later admitted to probation or state and federal prison. But even assuming a 25% double counting of individuals in these numbers would exceed 10 million people per year or nearly five per every 100 adults in the United States.
The fact that the entire correctional system has grown to historic levels suggests that our concern with mass incarceration should not be limited to state or federal prison populations, which make up a small share for the national correctional pie. Rather, reforms are also needed to lower the probation, parole, and jail populations, which often serve to “feed” state prison populations. For example, a large share of prison admissions are people who have failed the terms of probation and parole supervision. Similarly, lengthy periods of pretrial detention often serve to facilitate plea bargains for probation and prison sentences (Feeley, 1979). If one wishes to lower prison populations, probation, jail, and parole, populations must also be reduced.
Table 2. Correctional Populations and Admissions, 2014
Assume 25% Duplication
Rate per 100 Adults
Source: Bureau of Justice Statistics website data collections available at http://bjs.ojp.usdoj.gov and U.S. Census.
A Closer Look at Changes in State Prisoner Population Trends
While it’s clear that all forms of correctional control have dramatically increased, less is understood about the so-called components of this trend. Was it because judges were more likely to send a convicted felon to prison, more people were sent to prison for violent crimes, prison sentences became longer, or rising recidivism rates? It turns out that none of these factors fully explain the rise in incarceration rates. Rather, it was other more basic factors that fueled and continue to fuel the “imprisonment binge” (Austin & Irwin, 2012).
Prisons and all other correctional populations are the products of two basic factors—admissions and their length of stay (or LOS) (Clear & Austin, 2009). As shown below, both have dramatically increased over the past few decades, which is the math of higher incarceration rates. For decarceration to occur, the math that produced mass incarceration will have to be reversed. For a decarceration plan to be successful it must target those two fundamental drivers of correctional populations.
In the analysis that follows, there are two limitations that should be noted. First, much of the statistical data and analysis focuses only on state and federal prison populations. This is not to ignore the importance of the other three major components of the correctional system (local jails, probation and parole) which serve as “feeders” of people into state and federal prisons. For example, virtually all people admitted to state and federal prison have spent weeks and months in local jails awaiting the disposition of their criminal charges. Similarly, large percentages of prison admissions are people who have violated the terms of probation and parole supervision. The lack of attention to these other correctional systems is largely an artifact of the lack of national data available to conduct detailed analysis.
Second, throughout this paper, one needs to be aware of the undue influence of California’s recent reforms on recent correctional population trends. With the intervention of the federal courts and the passage of several ballot initiatives and associated legislative reforms, there have been significant reductions in the state’s prison, parole, jail, and probation populations, which can skew the national trends (Austin, 2016).
A comparison between the state prison population in 1978 (the earliest year that such data are available) and 2014 shows that the distribution of inmates by their primary offense has changed (Table 3). Notwithstanding the obvious jump in the state prison population by over one million people, the percentage of the population incarcerated for violent crimes has declined while the percent for drug offenses has increased. The rise in drug offenses is even more dramatic if one adds the nearly 100,000 people in federal prisons, which grew by over 160,000 people.
The rise in the federal prison population occurred after the Democratic-controlled Congress passed the Sentencing Reform Act of 1984 with bipartisan support, which mandated long prison terms for possession of crack cocaine, increased the likelihood of a prison sentence and the LOS for most crimes. At that time the Reagan administration initiated a “War on Drugs.” George Herbert Walker Bush later appointed a “Drug Czar.” The public was bombarded with news about drugs, like the drug death of sports figure Len Bias and the confessions of celebrities about personal struggles with substance abuse. More punitive forms of parole and probation supervision emerged with mandatory drug testing becoming the norm. For the first time, large numbers of people on probation and parole were incarcerated for testing positive for non-prescribed drugs and alcohol.
Table 3. Primary Offense for Sentenced State Prisoners 1978 versus 2014
Total State Prison
Total Federal Prison
Source: Bureau of Justice Statistics, Prisoners in 2014, and Survey of Inmates in State Correctional Facilities 1978 (SISCF), Bureau of Prisons Statistical Report, 1978.
By 1990 the prison population had risen to 740,000 with no signs of abating. Under both Democrat and Republican administrations and Congresses under the control of both parties, the prison populations continued to rise. It was President Clinton’s administration that endorsed the Violent Crime bill in 1994. Written in part by then-Senator Joe Biden, the bill funded 200,000 police officers and boot camps, along with $10 billion to help pay for prison construction for states willing to pass laws that would require longer prison terms for people convicted of violent crimes (truth in sentencing). By then the Democratic Party had achieved a major political goal—the public would view them as conservative as Republicans on the crime issue.
Trends in Court Sentencing
National criminal court sentencing data are more problematic to assess. Such data are actually based on two reporting series from the Bureau of Justice Statistics (BJS). One is an annual report that summarizes the results of “large urban courts,” while the other covers all of the state criminal courts.
Table 4 shows detailed state court sentencing data for the time frame between 1992 and 2006, and adds the number of index crimes reported, arrests, and the size of the prison population. The overall major trends are as follows:
• While the number of convictions made by the state courts grew by 21% the state prison population grew by 55%;
• The number of felony convictions has increased at a rate slightly higher than the increase in arrests;
• The number of drug convictions grew at a faster pace than most other crimes sentenced by the courts;
• There has been no change in the conviction rate;
• There has been little, if any, change in court prison disposition rates.
Table 4. State Felony Court Sentencing Trends 1992 versus 2006
State Prison Population
Source: Bureau of Justice Statistics website data collections available at http://bjs.ojp.usdoj.gov.
Trends in Prison Admissions
There is no doubt that prison admissions grew at a very fast rate, which had a direct impact on prison population growth. Up until the mid 1990s, part of that growth is explained by the associated increase in arrests, which for many crimes leads to a jail booking and some form of correctional supervision (Figure 2). But as crime and arrests began to decline, prison admissions continued to rise, reaching a peak in 2008 and then declining. Much of the decline is associated with major reforms in only California as part of its historic re-alignment legislation.
One reason that arrests and prison admissions do not always correlate is because a sizeable portion of the prison admission stream is made up of people who have violated the terms of probation and parole. National data show that parole violators constitute about one third of the prison admission stream. There are no national data on probation violators who are admitted to prison, but estimates from other major states suggest that another third of prison admissions are probation violators (Austin & Irwin, 2012, p. 30).
Specifically, behavior that one could not be arrested for let alone sent to prison or jail. Such reasons include use of recreational drugs or alcohol, failing to report for treatment, and/or failing to report to a parole or a probation officer. A recent study of Kentucky recidivism rates for state prisoners released in 2012 found that most of the reasons why a person was returned to prison were for non-criminal behavior (Table 5). Similar phenomena occur within probation departments.
In Kentucky over half of the prison admissions are either parole or probation violators. Until recently, over 60% of California’s prison admissions were parole violators (Austin & Irwin, 2012, p. 29). This is not to say that there is no criminal activity associated with these violations. Often there is. But they are at a level of severity that could result in a prison term unless they are on probation and parole supervision.
Table 5. Most Frequent Reasons for Parole Violation Warrants Kentucky Prison Releases 2012
Parole Violation Warrant Reason
Absconding parole supervision
Use of controlled substance—Marijuana
New misdemeanor arrest
New felony arrest
Failure to report to Parole Officer as directed
Failure to complete treatment for substance abuse
Failure to complete Halfway House program
Use of controlled substance
Failure to make restitution as directed
Use of controlled substance—Methamphetamines
Use of alcohol
Use of controlled substance—Cocaine
Absconding probation supervision
Failure to pay supervision fee as directed
Use of controlled substance—Opiates
Failure to pay for drug testing as directed
Leaving area of supervision without permission of Parole Off
Failure to report to Probation Officer
Failure to attend treatment for substance abuse
Failure to report arrest within 72 hours to Parole Officer
Failure to report change in home address to Parole Officer
Trends in State Prison Length of Stay
The significance of how long a person is under the custody and control of the correctional system is largely misunderstood by the public and policy officials. Too much attention has been directed at longer prison sentences when in fact the key driver has been increases in duration or LOS. Longer prison sentences have not been the cause of higher incarceration rates. Rather, it has been steady increase in LOS since the mid 1990s. Between 1993 and 2009, average overall sentence length actually declined as well for most offenses (Table 6).
At the same time, the average LOS for released prisoners has increased from 21 months to 29 months (Table 7). While eight months may not seem too significant, it has a huge impact on the size of the prison population. Using the basic formula of admissions x LOS = prison population, if the United States were to return to a LOS of 21 months as existed prior to 1994, the nation’s state prison population would decline by about 28% or about 430,000 inmates.
Table 6. Average State Prison Maximum Sentence Length (in months) for New Court Admissions by Crime Category 1993 versus 2009
Motor vehicle theft
Driving while intoxicated
Source: Bureau of Justice Statistics, National Corrections Reporting Program, 1993 and 2009 at http://bjs.ojp.usdoj.gov.
The increase in LOS is directly the result of tougher sentencing legislation that required prisoners to serve higher proportions of their sentences (truth in sentencing laws) and more restrictive parole board release practices. Similarly, parole supervision lengths have also increased by nine months. The total amount of time under state supervision has increased by 15 months to 55 months. And this statistic does not include the average of four to six months inmates are spending in the local jail under pretrial detention. In all, the total time of pretrial detention, incarceration, and parole supervision is about five years.
Table 7. Changes in Sentence Lengths, Prison and Parole LOS 1993–2009
Length of Supervision
Average Prison Time Served
Average Parole Supervision
Average Total State Time
With Pretrial Detention
Source: Bureau of Justice Statistics, National Corrections Reporting Program, 1993 and 2009 at http://bjs.ojp.usdoj.gov.
Three States that have Decarcerated Prison Populations
There are three major states (California, New York, and New Jersey) that have significantly reduced (by over 20%) their prison populations (Figure 2) and, at same time, significantly lowered their crime rates (Figure 3). It is interesting that relative to crime rates, the three states had higher or equivalent crimes rates as compared to the U.S. rate through the 1990s and now have rates below the U.S. crime rate. Just how these three states were able to do what other states have not done serves a useful guide for future reforms.
California has dropped its state prison population from a high of 174,282 in 2008 to 127,816 at the end of 2015. This reduction was the combination of persistent litigation by private lawyers over many years that resulted in a federal court order by a Three-Judge Panel (Plata v. Brown) that required California to find ways other than building more prisons to depopulate its horribly crowded prisons.
The major legislation that was passed in reaction to the Three-Judge Panel was AB 109, or more commonly known as “Realignment.” That legislation, which took effect on October 1, 2011, forbade counties to send people to state prison if they were convicted of certain non-violent crimes or had no convictions for a violent or “serious” crime(s). Prior to AB 109, a large portion of the prison admissions were people convicted of such crimes who were receiving relatively short sentences and spending a short period of time (three to nine months) in state prison before being released to one to three years of parole. But under AB 109, local jurisdictions can sentence the AB 109 inmates to a “full” prison term or a “split” or “blended” term. The latter means that only some proportion of the sentence must be served in the jail and the remainder under some form of community (typically probation) supervision.
A key justification for passage of AB 109 was California’s declining jail population that had created excess jail bed capacity. In 2007, the jail population reached a peak of 83,184 but had declined to 69,515 by the summer of 2011. While the California Department of Corrections and Rehabilitation (CDCR) had estimated that AB 109 would reduce the state prison population by as much as 40,000 inmates, it was clear the local jail systems could not accommodate that number of diverted state prisoners. In order to address the potential jail crowding issue and provide an economic incentive for the counties to use re-alignment as designed by the legislation, a “split-sentencing” provision was imbedded in the legislation.
There were other aspects of AB 109 that impacted the number of people on parole supervision. Prisoners currently incarcerated in the CDCR as of October 1, 2011 who were convicted of non-serious, non-violent, or non-high-risk sex offense (regardless of prior convictions) would, upon their release from state prison, be supervised by county probation departments. This population, known as Post-Release Community Supervision (PRCS), was formerly supervised by state parole and will decline over time.
Tied to the AB 109 legislation was the economic incentive of substantial funds being provided by the state to the counties to be used to manage the formerly state-imprisoned people at the local level. Specifically, the state offered a steady and permanent stream of block funding that matched the costs of housing each county’s prisoner in state prison.
In 2014, a ballot-sponsored initiative (Prop 47) redefined seven felony crimes, which further lowered the prison population but had the added benefits of lowering the state jail and probation populations. All total there are over 150,000 fewer adults in prison, jail, or on probation and parole, or a 23% decline in the total correctional system (Austin, 2016). The largest decline was in the parole population where lengthy parole terms were eliminated for many released prisoners Table 8). While these reforms were occurring the crime rate was continuing to decline and has reached its lowest rate since the 1950s.
Table 8. Changes in California Correctional Populations 2007–2015
Source: Austin (2016).
New Jersey has dropped its state prison population from 31,493 in 1999 to 21,590 by 2014 (31% decline). Somewhat related to California, the decline in New Jersey’s prison population was first triggered by ligation (Hawker v. Consovoy) that was directed at the state’s parole board. New Jersey, unlike California and New York, has indeterminate sentencing, so parole board decision-making has a significant impact on how long a person is incarcerated. The litigation was brought by three inmates in 2000, which showed that the parole board was not hearing cases in a timely manner, which resulted in extended periods of incarceration.
Shortly after the litigation was settled, the parole grant rate increased from 30% to 50% and has remained at that level. Other reforms that followed were reductions in parole violations and the reclassification of drug offenses as not eligible for state imprisonment (Maurer & Ghandnoosh, 2014). The drug law reforms were heavily influenced by the advocacy of drug reform organizations. As with California, crime rates continued to decline as the prison population dropped.
New York and New York City
New York’s state prison population has declined from 72,899 in 1999 to 52,518 by 2014. Unlike California and New Jersey, which were state-wide declines, New York’s decline can be linked to reforms that only occurred in New York City. The rest of the state has seen no reduction in its prison population. While there have been several reforms implemented within the NYC criminal courts, the major factor that explains the decline in state prisoners was the shift in arrests by the New York Police Department (NYPD). Beginning in 1994, the number of felony arrests began to drop while misdemeanor arrests increased. This reduction in felony arrests meant that prison admissions also began to decline.
Other important trends were sharp declines by 2009 in the NYC jail population (from 22,000 to 13,200), NYC probation population (77,000 to 44,000), and NYC parole populations (36,000 to 23,000). In total, there was an approximate 75,000 fewer people in state prison or jail, or on probation or parole by 2009. This dramatic overall decline was largely triggered by the change in arrest practices by the NYPD and other locally initiated reforms (Austin & Jacobson, 2013).
The Science of Decarceration and Crime Rates
The basis for a decarceration effort must be grounded in a science that ensures the public and public officials, responsible for passage of the needed decarceration laws, that crime rates will not rise as prison populations fall. The connection between crime rates and incarceration rates is clearly the biggest political obstacle to be acceptable of a decarceration strategy. Whenever a decarceration policy is proposed most law enforcement officials are quick to predict that crime rates will increase. And when decarceration occurs, the first test made by criminologists is whether crime rates went up or not. There are several studies that have addressed the relationship between crime and incarceration rates and whether abrupt changes in imprisonment may impact crime or the recidivism rates of early released prisoners.
Relationship Between Aggregate Imprisonment and Crime Rates
Is there a correlation between incarceration rates and crime rates? We already noted that U.S. crime (and incarceration) rates were relatively stable for several decades and did not begin to increase in the 1960s in response to some major decline in incarceration rates. It is also true that there is a strong, positive correlation between crime rates and incarceration rates by states. In general, states with higher rates of imprisonment have higher crime rates while states with lower incarceration rates tend to have lower crime rates (Figure 5). If imprisonment by itself was having powerful deterrent and incapacitation effects, the relationship would be in the opposite direction. Obviously other non-criminal justice factors are having a much stronger influence on crime rates. Linsky and Strauss (1986) found that macro level measures of social stress (e.g., business failures, unemployment, etc.) were associated with higher rates of violent crime, mental illness, and suicides in various states.
One could interpret Figure 5 as showing imprisonment acts as having a reactive or passive effect. Recent studies based on individual states and counties all have estimated the crime-reduction impact of prison growth to be small or nonexistent (Western, 2006; Lynch & Sabol, 1997; Stemen, 2007).
Reductions in Both State Imprisonment and Crime Rates
If there is not an inverse relationship between higher incarceration rates and crime rates, what do we know about the impact of abrupt reductions changes in imprisonment rates on crime rates? We have already seen in the experiences of California, New York, and New Jersey that crime rates did decline as correctional populations were lowered. But this may have been an example of good timing, since crime rates were dropping everywhere.
Durlauf and Nagin (2011) argued through their simulation models that both imprisonment and crime rates can be simultaneously reduced but by only increasing the presence of police in targeted high crime areas. Bushway (2016), in a review of research completed on California’s re-alignment decarceration efforts, argues that crime rates did increase as a result decarceration but, with the exception of motor vehicle theft, the increases were small. It’s interesting that Bushway reaches this conclusion while the authors of the study do not. Given that crime rates in California have continued to decline since re-alignment occurred speaks volumes to the tentative nature of such a conclusion. Unless the author is suggesting that crime rates would have declined even more without re-alignment.
The late James Q. Wilson, noted criminologist and early advocate of higher incarceration rates, noted that the United States had reached a tipping point of “diminishing returns” on the investment in prisons (Wilson, 2011). Professor Wilson reported that judges have always been tough on violent offenders and have incarcerated them for relatively long sentences. However, as states expanded incarceration, they dipped “deeper into the bucket of persons eligible for prison, dredging up offenders with shorter and shorter criminal records” (Wilson, 2011, p. 501).
The bulk of the evidence points to three conclusions: (1) The effect of imprisonment on crime rates, if there is one, is small; (2) If there is an effect, it diminishes as prison populations expand; and (3) the negative side effects of incarceration and over-crowding far outweigh the potential, unproven benefits of incarceration.
Prison Length of Stay and Prisoner Recidivism Rates
The evidence is more powerful on the issue of increasing or lowering LOS. Here the research shows that modest reductions in the current LOS for released prisoners do not increase recidivism rates. Perhaps one of the strongest arguments for lowering the length of imprisonment, and, thus, the prison population as shown in the U.S. Department of Justice recidivism studies for 1994 releases (California dominates the results by virtue of its large numbers) is the lack of a relationship between how long persons are incarcerated and recidivism rates (Table 9).
Although the recidivism rates appear to decline slightly, those declines are not statistically significant, except for people who serve more than 24 months, and that reduction in recidivism may be due to the maturation of the prisoner—older prisoners have lower recidivism rates. When this study is replicated, controlling for age and other related factors, there is no statistically significant difference in recidivism rates by LOS.
Table 9. Three-Year Follow-Up Rate of Rearrest of State Prisoners Released in 1994 by Time Served in Prison
3-Year Rearrest Rates
6 Months or Less
61 months or more
Source: Bureau of Justice Statistics, Prison Statistics. Retrieved from https://www.bjs.gov/.
A meta analysis of 12 studies of early release programs on recidivism rates by the National Council on Crime and Delinquency found early released inmates had the same or lower recidivism rates compared to those who were not early releases. Further, the state crime rates either declined or remained the same during the period that the early release program or policy was in effect (Guzman, Krisberg, & Tsukida, 2008).
Finally, there is no relationship between each state’s LOS, which varies significantly by state, and their reported crime rate (Figure 6). In other words, increasing or lowering a state’s length of incarceration seems to have no impact on crime rates. But LOS has a very large impact on incarceration rates. The PEW Center on the States estimated that in 2009, the costs of adding an additional nine months to the 1990 national LOS was costing states over $11 billion per year (PEW, 2012).
Rehabilitation and Treatment
Some have argued that prison populations and its associated costs can be reduced by enhancing rehabilitative treatment opportunities for those imprisoned (Aos et al., 2006). The problem with this approach is that despite increased progress in the science risk assessment and modest increases in program funding recidivism rates as measured by the BJS have remained unchanged over the past three decades (Table 10).
The reasons why recidivism rates remain stubbornly high vary. One major reason is that the treatment effects, even under the best conditions, are low. Experimental groups at best show a 10% reduction as compared to equivalent control populations. This means that the number needed to treat (NNT) in order to get one successful case is ten. Further, the fidelity of such programs is suspect. Even if a successful program is located that does not mean that programs like it will have the same success (the issue of external validity). Finally, the most successful programs are operated in the community versus prisons and jails, which are neither designed nor funded as treatment centers. While such programs are desirable in terms of humanizing prisons and allowing prisoners a means to lower their lengths of imprisonment, they, by themselves have not had an impact on rates of imprisonment or the associated costs.
Table 10. Three-Year Follow-up National Recidivism Rates Released State Prisoners
Source: Bureau of Justice Statistics website recidivism studies 1983, 1994, and 2005, http://www.bjs.gov/index.cfm?ty=tp&tid=17.
Decarceration and the Proportionality of Punishment
Reversing these trends will require a redefinition and acceptance, by the public, of the primary objectives of the correctional system. In particular, it will be necessary to reduce the various dosages of punishment now being handed out by the criminal justice system.
A key argument that transcends the question of whether imprisonment impacts crime rates is that prison sentences being handed out by the courts are not proportionate to the harm being suffered by victims. The disproportional level in punishment fuels the prison population and can produce widespread perceptions of injustice and inequities. A growing body of research shows that the burgeoning use of incarceration generates racial and social inequality. Moreover, it may damage children, families, and communities, with limited impact on the original objective: public safety (Clear, 2007).
One measure of the level of disproportionality is the cost of imprisonment versus the economic loss to victims of such crimes. As noted earlier, the United States is spending approximately $265 billion a year in federal, state, and local government criminal justice services. Over $80 billion is spent on corrections (jails, prisons, probation, and parole). Conversely, the most recent estimate of total economic loss to victims as reported by the National Crime Victimization Survey in 2008, was an estimated $17.4 billion (BJS, 2011). In that year, the total number of victimizations was 21.3 million, of which the vast majority (16.3 million) were property crimes. Within the property crime category, theft was the most common reported crime (12.3 million).
While the number of crimes reported by victims is much higher than the number reported to police in that year, the average victim losses are quite low. The overall average economic loss (which includes medical expenses and loss of wages) was $816 with a median loss of just $125. For the most frequent crime of theft, the average loss was $524 and a median figure of $100. The much lower median figures show that about one third of the reported victim losses are under $50.
It’s instructive to compare these costs with the five common crimes that people are being incarcerated for (Table 11). For example, the median loss associated with an assault is $80. The typical prison sentence for assault is 62 months, or about five years, of which the typical time served is 33 months. Together with the time spent in jail pretrial (estimated at five months), the average assault offender is incarcerated for 38 months at a cost of approximately $98,167. Even if a person is a repeat offender, of which may are, total losses for multiple crimes will not approach the total punishment costs for assault. The same can be said about the other crimes. And these costs do not include the cost of post-release supervision. From an economic perspective, neither victims nor taxpayers are getting a good return on their criminal justice system investments.
Table 11. Economic Loss to Victims and Costs of Incarceration, 2008
Type of Crime
Median Victim Loss
Prison Sentence (in mos)
Pretrial Time (in mos)
Prison Time Served (in mos)
Total Time (in mos)
Source: Criminal Victimization, 2008. Washington, DC: Bureau of Justice Statistics National Corrections Reporting System. Retrieved from https://www.bjs.gov/.
Note: (*) Incarceration costs based on an average annual incarceration cost of $31,000 (Henrichson & Delaney, 2012).
What a Decarceration Plan Would Look Like
How does one begin lowering not only the prison population, but also other forms of correctional control? The good news is that the necessary reforms have been widely adopted in many states either currently or in the past. For many decades the United States has much lower rates of correctional control that correlate with today’s crime rates. So, the means required to decarcerate are neither radical or need to take a long time to implement. Rather, one only needs to return to those same practices that were in existence in the past or exist today in some jurisdictions.
Significant variations in state and county incarceration and correctional control rates already exist where many states and counties have very low rates of incarceration. For these jurisdictions, returning to previous incarceration rates that match today’s crime rates will require minimal reforms. But for those where incarceration rates are high, a long-term process involving systemic policy and legislative changes in law enforcement, prosecutorial practices, sentencing, and correctional policy will be required.
In the analysis that follows, a series of crucial pressure points that have substantial impact on the size of the prison, jail, probation, and parole populations are identified. Using the most current national data, simulations of how adjustments in these pressure points will prompt change in the targeted population will be presented. These simulations are based on current national data, which have been disaggregated by admission type and offense group. The status quo simulation is based on the correctional system that existed as of 2007, which has remained largely unchanged. The key reforms that will have to be implemented are as follows:
1. Reduce length of stay for sentenced prisoners including people convicted of violent crimes;
2. Divert technical parole violators from prison and reduce their length of stay;
3. Divert technical probation violators from prison and reduce their length of stay; and
4. Divert persons convicted of victimless crimes from prison.
The reductions in LOS are key to achieving sizable population reductions. The justifications for such a change are (1) a state’s prison LOS is not related to recidivism or crime rates and (2) the current LOS is well above prior levels. Reducing LOS will require repeal of mandatory minimums and truth in sentencing laws, increasing parole grant rates (for states with indeterminate sentencing), and increasing good time credits for work and program participation. Prisoners convicted of robbery and assault must be included in these reforms. Similarly, those with life sentences with the possibility of parole who are low risk to recidivate must start to be paroled at their parole eligibility dates.
The recommendation to divert technical parole and probation violators is similarly rooted in the lack of proportionality of punishment for such behavior. Technical violations and re-arrests for misdemeanor crimes for which a person cannot be sentenced to prison can no longer be used to re-incarcerate people simply because of their parole or probation status.
Finally, diversion of people convicted of drug possession and other victimless or public disorder crimes would occur. Here again the basis for this recommendation is that state imprisonment is not proportionate for such deviant behavior.
Relative to reducing the size of the parole and probation populations, the key reform would be to reduce the period of supervision. There is ample evidence that traditional probation and parole supervision can be both criminogenic and excessive. As shown earlier, people who discharge from prison have similar or lower recidivism rates as compared to those who are placed on parole. Further, there is no relationship between length of community supervision and recidivism rates (Austin & Irwin, 2012, p. 257).
The above reforms would also lower local jail populations, which have already been declining as crime and arrests have been reduced. Reductions in the bookings of probation and parole violators who are detained for several weeks until their violations are disposed of by the courts or parole boards would moderately lower jail populations.
If the above reforms were implemented, which largely reflects the sentencing practices that existed several decades ago, what would the cumulative effects be?
Table 12 mimics the decarceration policies on the state prison population that were in play in 2007 (prior to California’s reforms). For violent crimes, admissions remain the same but LOS is reduced to the levels that existed in the 1980s. Both admissions and LOS are reduced for property and drug crimes. Technical parole violators and some victimless crimes are virtually eliminated.
Table 13 shows the effects of both lowering probation and parole populations and relying less on these forms as supervision as a sentence or for post release purposes. All of these suggestions (reduced time in prison, parole, and probation) have been successfully implemented in a state or county.
Table 12. Estimated Effects of Decarceration Reforms on State Prison Population
New Court Admissions
Technical Parole Violators
State Incarceration Rate
Table 13. Estimated Effects of Decarceration Reforms on Probation and Parole Populations
Current LOS and Admissions
New LOS and Diverted Probationer/Parolees
The cumulative effects result in the entire correctional system declining by approximately 50%. The prison incarceration rate would fall to 241 per 100,000 but would still be well above the historic rates of 100. Prison admissions for violent crimes would remain the same but their LOS would be lowered to pre-1990 levels. Technical probation violator admissions that are imprisoned for pure technical violations would be diverted from prison while those arrested and convicted of new crimes would serve reduced prison terms comparable to other new court commitments. Parole technical violators would largely be prohibited from being re-incarcerated. And people placed on probation and parole would serve shorter periods of community supervision. The federal prison system would be significantly lowered by repealing its 85% truth in sentencing requirement and returning to a more normal standard of 50%.
Table 14. Estimated Impact of Decarceration Reforms on Current Correctional System
Plus Court Diversions
Plus Diverted “Victimless” Crimes
Remaining Obstacles to Decarceration
Correctional populations are the product of a very simple equation—admissions x length of stay. The only way to significantly lower correctional populations is to reduce the current number of admissions and the lengths of stay for people imprisoned or placed on probation or parole. People convicted of violent crimes, repeat property offenders, and drug offenders cannot be excluded from such reforms. The obvious solution lies in simply returning to the same sentencing and correctional policies that existed a few decades ago when the U.S. crime rate was what it is today and the incarceration rate was much lower.
But this is obviously a simplistic solution. While the “science of decarceration” and the experiences of some jurisdictions indicate that correctional systems can be significantly lowered without increasing crime rates, there will continue to be considerable resistance to such reforms. Elected officials continue to benefit supporting muscular criminal justice policies and laws that serve to define crime prevention as the product of high incarceration rates. The most recent experience in California illustrates that reform is unlikely to occur within the criminal justice system. Rather, external forces like litigation and public ballot initiatives seem to be more powerful agents of change.
It is absolutely clear that the United States had much lower incarceration rates for many decades that match contemporary crime rates. But there remain lingering doubts in the minds of criminologists, public officials, and the public on whether decarceration can be implemented without negatively impacting crime rates and at what costs. Admittedly this is a very high and arguably an unfair standard. But unless there are clear economic and public safely benefits to decarceration, current levels of imprisonment and correctional control are likely to be remain where they are for many decades to come.
Andrews, D., Zinger, I., Hoge, R. D., Bonta, J., Gendreau, P., & Cullen, F. T. (1987). Does correctional treatment work? A clinically relevant and psychologically informed meta-analysis. Criminology, 28(3), 369–404.Find this resource:
Currie, E. (2013). Crime and punishment in America. New York: Picador.Find this resource:
Gendreau, P., & Ross, P. (1987). Revivification of rehabilitation: Evidence from the 1980s. Justice Quarterly, 4(3), 349–407.Find this resource:
Liedka, R. V., Piehl, A. M., & Useem, B. (2006). The crime control effect of incarceration: Does scale matter? Criminology and Public Policy, 5, 245–276.Find this resource:
Nagin, D. (2010). Imprisonment and crime control: Building evidence-based policy. In R. Rosenfeld, K. Quiner, & C. Garcia (Eds.), Contemporary issues in criminological theory and research: The role of social institutions (pp. 309–316). Belmont, CA: Wadsworth.Find this resource:
Petersilia, J., & Cullen, F. T. (2014). Liberal is not stupid: Meeting the promise of downsizing prisons. Stanford Journal of Criminal Law and Policy, 2, 1–43.Find this resource:
Sundt, J., Salisbury, E. J., & Harmon, M. G. (2016). Is downsizing prisons dangerous? The effect of California’s Realignment Act on public safety. Criminology and Public Safety, 15, 315–341.Find this resource:
Tonry, M. (2010). Less imprisonment less crime: A reply to Nagin. In R. Rosenfeld, K. Quiner, & C. Garcia (Eds.), Contemporary issues in criminological theory and research: The role of social institutions (pp. 317–330). Belmont, CA: Wadsworth.Find this resource:
Travis, J., Western, B., & Redburn, S. (2014). The growth of incarceration in the United States: Exploring the causes and consequences. Washington, DC: The National Academies Press.Find this resource:
Western, B. (2006). Punishment and inequality in America. New York: Russell Sage.Find this resource:
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based public policy options to reduce future prison construction, criminal justice costs, and crime rates. Federal Sentencing Reporter, 19, 275–290.Find this resource:
Austin, J. (2010). Reducing America’s correctional populations: a strategic plan. Justice Research and Policy, 12(1), 1–32.Find this resource:
Austin, J. (2016). Regulating California’s prison population: The use of sticks and carrots. In C. Kubrin & C. Seron (Eds.), The ANNALS of the American academy of political and social science: The great experiment: Realigning criminal justice in California and beyond (Vol. 664, pp. 84–107).Find this resource:
Austin, J., Cadora, E., Clear, T., Dansky, K., Greene, J., Maurer, M., et al. (2013). Ending mass incarceration: Charting a new justice reinvestment. Washington, DC: The Sentencing Project.Find this resource:
Austin, J., & Irwin, J. (2012). It’s about time: America’s imprisonment binge. Belmont, CA: Wadsworth Publishing.Find this resource:
Austin, J., & Jacobson, M. (2013). How New York City reduced mass incarceration: A model for change? New York: Brennan Center for Justice.Find this resource:
Bureau Justice of Statistics. (2011). Criminal victimization in the United States, 2008. Statistical Tables. Washington, DC: United States Department of Justice, Bureau of Justice Statistics.Find this resource:
Bushway, S. D. (2016). Evaluating realignment : What do we learn about the impact of incarceration on crime?Criminology and Public Policy, 15, 309–314.Find this resource:
Clear, T., & Austin, J. (2009). Reducing mass incarceration: Implications of the iron law of prison populations. Harvard Law & Policy Review, 3, 307–324.Find this resource:
Clear, T. R. (2007). Imprisoning communities: How mass incarceration makes disadvantaged neighborhoods worse. Oxford: Oxford University Press.Find this resource:
Cullen, F. T., Johnson, C. L., & Nagin, D. (2011). Prisons do not reduce recidivism: The high cost of ignoring science. The Prison Journal, 91, 48–65.Find this resource:
Durlauf, S. N., & Nagin, D. S. (2011). Overview of “Imprisonment and crime: Can both be reduced?” Criminology & Public Policy, 10, 9–12.Find this resource:
Feeley, M. M. (1979). The process is the punishment: Handling cases in a lower criminal court. New York: Russell Sage.Find this resource:
Gingrich, N., & Jones, V. (2014). Prison system is failing America. CNN Opinion. Retrieved from http://www.cnn.com/2014/05/21/opinion/gingrich-jones-prison-system-fails-america/.Find this resource:
Guzman, C., Krisberg, B., & Tsukida, C. (2008). Accelerated release: A literature review. Oakland, CA: National Council on Crime and Delinquency.Find this resource:
Head, T., & Norquist, G. (2015, August 13). The high costs of over-incarceration. The National Review. Retrieved from http://nationalreview.com.Find this resource:
Henrichson, C., & Delaney, R. (2012). The price of prisons: What incarceration costs taxpayers. New York: The Vera Institute.Find this resource:
Kyckelhahn, T. (2015). Justice expenditure and employment extracts, 2012—preliminary (NCJ 248628). Washington, DC: Bureau of Justice Statistics.Find this resource:
Levitt, S. (1996). The effect of prison population size on crime rates: Evidence from prison overcrowding litigation. Quarterly Journal of Economics, 111, 319–351.Find this resource:
Linsky, A. S. & Strauss, M. (1986). Social Stress in the United States. Dove, MA: Auburn House.Find this resource:
Lynch, J. P., & Sabol, W. J. (1997). Did getting tough on crime pay? Crime Policy Report No. 1. Washington, DC: Urban Institute.Find this resource:
Maurer, M., & Ghandnoosh, N. (2014). Fewer prisoners, less crime: A tale of three states. Washington, DC: The Sentencing Project.Find this resource:
Minton, T., & Zeng, Z. (2015). Jail inmates at midyear 2014 (NCJ 248629). Washington, DC: Bureau of Justice Assistance.Find this resource:
Nagin, D. S. (2013). Deterrence in the twenty-first century. In M. Tonry (Ed.), Crime and justice (Vol. 42). Chicago: University of Chicago Press.Find this resource:
PEW. (2012). Time served. The high cost, low return of longer prison terms. Washington, DC: The Pew Center on the States.Find this resource:
Schiraldi, V., & Jacobson, M. (2014, June 4). Could less be more when it comes to probation supervision? [Blog post]. Retrieved from http://americancityandcounty.com/blog/could-less-be-more-when-it-comes-probation-supervision.Find this resource:
Spelman, W. (2006). The limited importance of prison expansion. In A. Blumstein (Ed.), The crime drop in America (Rev. ed., pp. 97–129). New York: Cambridge University Press.Find this resource:
Stemen, D. (2007). Reconsidering incarceration: New directions for reducing crime. New York: Vera Institute of Justice.Find this resource:
Sullum, J. (August/September 1999). Prison conversion: After studying non-violent drug offenders, a criminologist who once said “let ‘em rot” now says “let ‘em go.” Reasonline. Retrieved from http://www.reason.com.Find this resource:
Tonry, M. (2015). The fog around cost-of-crime studies may finally be clearing: Prisoners and their kids suffer too. Criminology and Public Policy, 14, 663–671.Find this resource:
Wilson, J., & Davis, R. C. (2006). Good intentions meet hard realities: An evaluation of the Project Greenlight reentry program. Criminology and Public Policy, 5, 303–338.Find this resource:
Wilson, J. Q. (2011). Crime and public policy. In J. Q. Wilson & J. Petersilia (Eds.), Crime and public policy (pp. 489–507). Oakland, CA: ICS Press.Find this resource:
Western, B. (2006). Punishment and inequality in America. New York: Russell SageFind this resource: