\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 1

24-AUG-09

9:01

Reducing Mass Incarceration: Implications of
the Iron Law of Prison Populations
1

Todd R. Clear*
James Austin**
Beginning in the 1970s, the United States embarked on a three-decadelong shift in its penal policies. In these years, state and federal governments
tripled the percentage of convicted felons sentenced to confinement and
doubled the length of their sentences. As a consequence of these changes,
punishment in the United States has become an outlier, not only among prevailing practices in the Western world, but also in comparison to the United
States’ own long-standing practices.2 United States imprisonment rates are
now almost five times higher than the historical norm prevailing throughout
most of the twentieth century, and they are three to five times higher than in
other Western democracies.3
The amount of writing by scholars and analysts during this thirty-year
period regarding the exceptional nature of U.S. penal policy could fill a library. Early on, many writers suspected that the U.S. prison population was
too small and needed to grow.4 But in recent years, as the growth of the
prison population reached levels that were well beyond those anyone had
anticipated and that few believed were needed, the literature about the U.S.
prison system has shifted to emphasize deep concerns about the wisdom of
our burgeoning prison population. Today, a broad consensus has emerged

1
This paper draws upon three recent volumes in which the authors were principal
writers: TODD R. CLEAR, IMPRISONING COMMUNITIES: HOW MASS INCARCERATION MAKES
DISADVANTAGED PLACES WORSE (2007); JAMES AUSTIN ET AL., UNLOCKING AMERICA: WHY
AND HOW TO REDUCE AMERICA’S PRISON POPULATION (2008), available at http://www.jfaassociates.com/publications/srs/UnlockingAmerica.pdf; and James Austin, Reducing America’s
Correctional Populations (2009) (unpublished working paper for the National Institute of
Corrections, on file with author).
* Todd R. Clear is Distinguished Professor of Criminal Justice at John Jay College of Criminal Justice of the City University of New York. He is president of the American Society of
Criminology.
** James Austin is a criminologist and the President of the JFA Institute. Dr. Austin is
currently working with a number of states and county jails on reducing their correctional populations and evaluating the impact of various legislative and correctional reform policies.
2
For the data in support of this conclusion, see MARIE GOTTSCHALK, THE GALLOWS: THE
POLITICS OF MASS INCARCERATION IN AMERICA 1–6 (2006); INVISIBLE PUNISHMENT: THE COLLATERAL CONSEQUENCES OF MASS IMPRISONMENT 279–92 (Marc Mauer & Meda ChesneyLind eds., 2002); BRUCE WESTERN, PUNISHMENT AND INEQUALITY IN AMERICA 12–15 (2006);
JAMES Q. WHITMAN, HARSH JUSTICE: CRIMINAL PUNISHMENT AND THE WIDENING DIVIDE BETWEEN AMERICA AND EUROPE 3–17 (2003).
3
See AUSTIN ET AL., supra note 1, at 3–4; MARC MAUER, SENTENCING PROJECT, COMPARATIVE INTERNATIONAL RATES OF INCARCERATION, 1–2 (2003), available at http://www.sentencingproject.org/Admin/Documents/publications/inc_comparative_intl.pdf.
4
See, e.g., EDWIN ZEDLEWSKI, U.S. DEP’T OF JUSTICE, MAKING CONFINEMENT DECISIONS
(1987).

\\server05\productn\H\HLP\3-2\HLP209.txt

308

unknown

Seq: 2

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

that the prison population is too large.5 At the same time, pressing fiscal
demands at the state and federal levels have convinced many policymakers
that the current system is no longer affordable.6
The scholarly debate about the prison population today considers three
main topics. First, there is contention about the degree to which the current
drop in crime is a consequence of the larger number of prisoners—a question we consider briefly in the next section of this Essay. Work addressing
this concern tends to show that incarceration rates are not as strongly connected to public safety as one would ordinarily think. Second, there is a
growing literature on the unintended consequences of imprisonment; in particular, new studies show that high rates of incarceration have come at an
indefensible cost in terms of racial and social inequality and have damaged
children, families, and communities.7 Works within this literature tend to
argue for the need to reduce prison populations as a simple matter of social
justice. Third, there is a literature that proposes strategies for reducing
incarceration.8
This Essay contributes to the last topic—what to do about mass incarceration. It is not an argument that “something needs to be done.” We believe that point has been made eloquently and convincingly by a solid body
of work, some of which includes our own previous writing.9 Instead, our
purpose in this Essay is to chart a broad strategic course for anyone seeking
to reduce imprisonment.
This Article is needed because policymakers’ work on incarceration is
hampered today by two important misunderstandings. To begin with, too
many policymakers assume that the connection between incarceration rates
and crime rates is greater than it actually is, and so they are much more
cautious about reducing incarceration rates than they need be. They also fail
to take account of what we refer to as “the iron law of prison populations”—that the total number of prisoners behind bars is purely and simply a
result of two factors: the number of people put there and how long they stay.
As a result, policymakers spend too much time considering policy proposals
that will have little effect on incarceration rates.
In this Essay, we make three points. First, we show why the link between incarceration rates and crime rates is not as great as many policymakers presume. We do so in order to make the case that an aggressive program

5
Compare conservative writers, such as John DiIulio, Two Million Prisoners are Enough,
WALL ST. J., Mar. 12, 1999 at A14, and BERT USEEM & ANNE MORRISON PIEHL, PRISON
STATE: THE CHALLENGE OF MASS INCARCERATION 169–79 (2008), with their more liberal
counterparts, like STEVEN RAPHAEL & MICHAEL A. STOLL, DO PRISONS MAKE US SAFER?
20–22 (2009).
6
RYAN S. KING, SENTENCING PROJECT, THE STATE OF SENTENCING 2008: DEVELOPMENTS
IN POLICY AND PRACTICE 1 (2009), available at http://www.sentencingproject.org/Admin/Documents/publications/sl_statesentencingreport2008.pdf.
7
See CLEAR, supra note 1; WESTERN, supra note 2, at 85–105, 131–162.
8
A recent review of some of these strategies is provided in MICHAEL JACOBSON, DOWNSIZING PRISONS: HOW TO REDUCE CRIME AND END MASS INCARCERATION 173–214 (2005).
9
See generally CLEAR, supra note 1; AUSTIN ET AL., supra note 1; Austin, supra note 1.

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 3

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

309

to reduce prison populations can proceed without a substantial negative impact on public safety. We then describe the iron law of prison populations.
Finally, taking advantage of this iron law, we propose a set of penal changes
that, if implemented, would cut the correctional population roughly in half
and return the prison system to an incarceration rate similar to that of almost
thirty years ago—before the trend toward mass incarceration picked up
steam. These changes would reverse the trend with limited impact on public
safety. We begin the Essay by explaining why.
I. THE CONNECTION BETWEEN CRIME RATES

AND

INCARCERATION RATES

As David Garland once noted, it would be silly to argue that there is no
connection between incarceration rates and crime.10 After all, it takes a crime
to put someone behind bars, and surely imprisonment has some suppression
effect on crime. Thus, as crime rates rise, imprisonment rates might also be
expected to rise. When this occurs, crime rates might be expected to fall as a
result. Empirical studies confirm this pattern, but the size of the impact in
both directions is surprisingly small. In fact, in terms of the big picture,
today’s crime rate is roughly what it was in the early 1970s, but the incarceration rate then was one-sixth what it is today. Moreover, state incarceration
rates and state crime rates have a mildly positive correlation, although not a
strong one.11
Several recent studies have attempted to calculate the relationship between the size of the prison population and crime. Their findings depend
upon the time period they study and whether they analyze national patterns
or provide state-specific estimates. Studies investigating national-level patterns from the mid-1980s to the early 1990s estimate that prison population
sizes have a large effect on crime.12 Later studies, especially those that investigate the impact of incarceration growth patterns at state and local levels,13
find much smaller effects, even approaching zero. A consensus has emerged
among criminologists that the impact of imprisonment on crime is modest
compared to other factors.14
10
DAVID GARLAND, PUNISHMENT AND MODERN SOCIETY: A STUDY IN SOCIAL THEORY
18–19 (1993).
11
RYAN S. KING ET AL., SENTENCING PROJECT, INCARCERATION AND CRIME: A COMPLEX
RELATIONSHIP (2005), available at http://www.sentencingproject.org/Admin/Documents/publications/inc_iandc_complex.pdf.
12
See, e.g., Stephen D. Levitt, The Effect of Prison Population on Crime Rates: Evidence
From Prison Crowding Legislation, 111 Q. J. ECON. 319 (1996); Thomas B. Marvell & Carlisle E. Moody, The Impact of Prison Growth on Homicide, 1 HOMICIDE STUD. 205–233
(1997).
13
See, e.g., Tomislav Kovandsic & Lynne M. Vieraitis, The Effect of County-Level Prison
Populations on Growth in Crime Rates, 5 CRIMINOLOGY & PUB. POL. 213, 215 (2006). See
also WESTERN, supra note 2, at 182; Peter Reuter & Shawn Bushway, Revisiting Incapacitation: Can We Generate New Estimates?, 23 J. QUANTITATIVE CRIMINOLOGY 259 (2007).
14
See, e.g., DON STEMEN, VERA INST. OF JUSTICE, RECONSIDERING INCARCERATION: NEW
DIRECTIONS FOR REDUCING CRIME (2007), available at http://www.vera.org/publication_pdf/

\\server05\productn\H\HLP\3-2\HLP209.txt

310

unknown

Seq: 4

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

There are several reasons why the size of the prison population has so
little to do with crime, including one crucial limit on the penal system’s
ability to affect crime. That is, when one person is locked up, another person frequently comes along to replace him, thereby maintaining the rate of
crime. This is particularly true for drug-related crime.15 But it is also true for
much of the crime committed by young men in groups, because loosely
formed and intermittently criminally active groups quickly find new members when old ones go to prison.16 Thus, the generally increased likelihood
that a felony conviction will lead to imprisonment has resulted in an increased number of people going to prison, but it has not produced a decrease
in the criminal activity of those who remain behind.
What would happen if we increased the prison-release rate?17 First, the
size of the prison population would drop. But what would happen to crime
rates? Increasing the rate of release would increase the number of people
reentering society from prison, certainly a risk factor for crime rates. Yet
recent studies show that the length of stay in prison is not associated with a
change in the risk of recidivism.18 So sending people to prison for shorter
periods would not make them more likely to commit crimes upon release.
Undeniably, a larger release cohort will have some effect on crime
rates. But there are good reasons to think that the overall risk would be
small.19 One study of the nation’s 1994 prison release cohort found that it
accounted for only five percent of all arrests for felonies in the three-year
time period that followed its release, and it produced only one percent of
arrests for violent crimes.20 Even though people released from prison are
themselves at a high risk of committing new crimes, they commit only a
small fraction of all crimes in any given year. Given that prisoners do not
become less likely to commit crimes upon release as they spend more years
in prison, and given that the contribution of former inmates to the overall
379_727.pdf; PEW CTR. ON THE STATES, THE IMPACT OF INCARCERATION ON CRIME: TWO
NATIONAL EXPERTS WEIGH IN (2008), available at http://www.pewcenteronthestates.org/
uploadedFiles/Crime%20Incarceration%20QA.pdf.
15
See KING ET AL., supra note 11, at 6. For a journalistic account of this phenomenon, see
DAVID SIMON & EDWARD BURNS, THE CORNER: A YEAR IN THE LIFE OF AN INNER-CITY
NEIGHBORHOOD (1997).
16
Marcus Felson, The Process of Co-offending, in THEORY FOR PRACTICE IN SITUATIONAL
CRIME PREVENTION 149 (Martha J. Smith & Derek B. Cornish eds., 2003).
17
We return to this question later in the paper. See infra Part V.B.2.
18
Daniel S. Nagin et al., Imprisonment and Reoffending, in 38 CRIME AND JUSTICE: AN
ANNUAL REVIEW OF RESEARCH (forthcoming 2009).
19
Of course, no matter how small the risk, the fact that more than 700,000 prisoners are
released each year already has the capacity to alarm people, so we can imagine how policymakers would feel about an increase in that number. See HEATHER C. WEST & WILLIAM J.
SABOL, U.S. DEP’T OF JUSTICE, PRISONERS IN 2007, at 3 (2008), available at http://
www.ojp.usdoj.gov/bjs/pub/pdf/p07.pdf.
20
PATRICK A. LANGAN & DAVID J. LEVIN, U.S. DEP’T OF JUSTICE, RECIDIVISM OF PRISONERS RELEASED IN 1994, at 5 (2002), available at http://www.ojp.usdoj.gov/bjs/pub/pdf/
rpr94.pdf. Higher estimates have been made by Richard Rosenfeld, et al., The Contribution of
Ex-Prisoenrs to Crime Rates, in PRISONER REENTRY AND CRIME IN AMERICA 80 ( Jeremy
Travis & Christy Visher eds, 2005).

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 5

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

311

crime rate is relatively small, increasing the prison release rate would seem
to have little disadvantage. Certainly, some prisoners will commit crimes
upon release. However, we can assume these individuals would commit
these crimes upon their release at a later date anyway, and it is wildly impractical (and entirely inconsistent with current practice) to propose keeping
all potential recidivists incarcerated permanently.
There is further good news: as the prison population drops, the number
of people available to be released from prison would also drop. So in the
later years of a plan to accelerate prison releases, the size of the reentry
cohort—the one we are most worried about for new crimes—would also
drop. In other words, an accelerated release program moves the risk of new
offenses to an earlier period of time, purely as a consequence of an increase
in the cohort size. But it does so in a trade for an equivalent reduction in the
risk of new offenses at the later time, as the release cohort gets smaller.
There is even better news for those reformers who would reduce the cohort
who enter prison in the first place. The same studies that now show length of
stay to have no relationship to rate of recidivism also show that going to
prison in the first place does not reduce the likelihood that a criminal offender will be a repeat offender and may make it marginally higher.21
The conclusion we can draw from this analysis is that the size of the
prison population and the amount of crime are related, but not strongly. In
particular, the speed at which people are released from prison is not related
to their likelihood of staying crime-free. This suggests that prisoners can
serve shorter sentences without triggering an increase in the crime rate. Furthermore, maintaining a large prison population does not necessarily significantly decrease the number of crimes committed. So a fairly substantial
amount of leeway exists to change the rate at which people are released from
prison without much long-term net impact on public safety.
In other words, this is an area where policymakers can innovate without
imperiling the public, certainly in the long term. The remainder of this Article addresses the question: What kind of innovations ought they consider?
II. THE IRON LAW

OF

PRISON POPULATIONS

In addition to assuming too great a relationship between high imprisonment rates and low crime rates, policymakers too often ignore the way
prison populations grow. They commonly engage in debates over “rehabilitation versus punishment,” and they are deeply invested in advocacy for
drug treatment programs and reentry planning. The amount of attention absorbed by the Second Chance Act of 2007 is a great example.22 This legisla21
See, e.g., Paul Nieuwbeerta, et al., The Relationship Between First Imprisonment and
Criminal Career Development: A Matched Samples Comparison, J. QUANTITATIVE CRIMINOLOGY (forthcoming).
22
Second Chance Act of 2007, Pub. L. No. 110-199, 122 Stat. 657 (codified as amended
in scattered sections of 42 U.S.C.).

\\server05\productn\H\HLP\3-2\HLP209.txt

312

unknown

Seq: 6

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

tion provides extensive federal monetary support for states’ efforts to
improve services and supervision for people who are released from prison.
One might argue that this Act is a good thing, but it will clearly not do much
about prison populations. Why? Because all of its effects are side-issues that
ignore the way prison populations are created.
The Iron Law of Prison Populations states that the size of a prison
population is completely determined by two factors: how many people go to
prison and how long they stay. If either of these factors changes, the size of
the prison population will also change. The corollary to this iron law is
equally important: There is no way to change the prison population without
changing either the number of people who go to prison or how long they
stay there.
Viewed in light of the iron law, the growth of the prison population in
the United States can be divided into three rough periods. In the 1970s, the
prison population grew because the crime rate grew, resulting in greater
numbers of people going to prison. In the 1980s, and stretching into the early
1990s, a host of sentencing policies restricted the use of probation as a sentence for felons, causing a substantial increase in the number of people entering prison during a period when crime rates were semi-stable.23 After that,
legislation that enhanced penalties for felonies greatly increased the average
length of prison terms, which led to growing prison populations even as
crime rates dropped and the number of people entering prison began to stabilize. The result was a growing backlog of people serving long sentences,
who made up a permanent population base upon which the flow into and out
of prison was grafted.
There is substantial debate about whether these changes in penal policy
were wise. But there can be no doubt that the recent growth in prison populations in the United States can be viewed almost entirely as a matter of
changes in penal policy taking place since 1980. This history supports our
argument that America will have whatever prison population its penal policies create. It is even conceivable that, with a new set of policies, the prison
population could be cut in half—and this could be done without greatly affecting crime rates in the long run.24 But the other extreme is also true. With
over five million people under correctional control who are not behind bars,
the prison population could, through legal and administrative policy, be
doubled, again without much impact on crime. The point is that the size of
the prison population is a matter of penal policy, and over the last thirty-six
years, the United States has built a policy designed to grow prisons.

23
Alfred Blumstein & Allen J. Beck, Reentry as a Transient State Between Liberty and
Recommitment, in PRISONER REENTRY AND CRIME IN AMERICA 50–51 (Jeremy Travis &
Christy Visher eds., 2005).
24
A short-run effect might occur because a cut in the prison population would result in a
larger release cohort. But the impact of releases on crime would soon approach zero, as the
prison population reached a new homeostasis.

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 7

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law
III. THE FAILURE OF THE CONTEMPORARY DEBATE
MASS INCARCERATION

9:01

313

ON

Despite this history, policymakers focused on reducing mass incarceration have failed to heed the iron law’s two conditions. Following three decades of policies that increased the prison population through changes in
sentencing and (to a lesser extent) post-release supervision, policymakers
continue to act as though they think they can reduce the size of the prison
population without directly taking on the rate and length of sentences. Table
1 lists recent sentencing reforms passed by seventeen states that were designed to reduce incarceration and were widely publicized in a report by The
Sentencing Project this year.25 What can be seen from this list is how little
these reforms have to do with either the number of people being sentenced to
prison or their lengths of stay there. Some alter the nature of community
supervision, and so arguably address the rate at which felons are recycled
into prison by being revoked from community supervision. And some aim to
make non-incarceration alternatives more attractive by setting up new or
more incentivized systems of community supervision. Some set up advisory
commissions to provide strategies for reducing prison costs. But there is not
a single new law or initiative designed to change the length of sentences or
to reduce restrictions on the use of probation as a sentence.
Sentencing reforms can succeed in reducing mass incarceration only if
these elements are at their core. It has already been remarked that the
agenda of sentencing reform is easily side-tracked. In a recent article in this
Journal’s online edition, Douglas A. Berman argued that a focus on innocence and debates about the death penalty have made it harder to have a
serious, engaging discussion about mass incarceration.26 But these are not
the most significant distractions from our necessary national debate about
mass incarceration. Far more important are the emphases on reentry, alternatives to incarceration programs, and the philosophy of rehabilitation, which
often serve as crucial distractions from the agenda of mass incarceration, a
problem that can be addressed only with a focus on the iron law’s two
elements.
IV. REHABILITATION PROGRAMS
When people hear complaints about mass imprisonment they often automatically assume that opposition to big prisons means favoring rehabilitation. But the debate between punishment and treatment is unwarranted, as
they are not opposed. They are merely different coercive penal strategies.
25

See KING, supra note 6, at 2.
Douglas A. Berman, Reorienting Progressive Perspectives for Twenty-First Century
Punishment Realities, HARV. L. & POL’Y REV. ONLINE 1, 8 (2008), http://hlpronline.com (follow “Reorienting Profressive Perspectives for Twenty-First Century Punishment Realities”
hyperlink).
26

\\server05\productn\H\HLP\3-2\HLP209.txt

314

unknown

Seq: 8

24-AUG-09

Harvard Law & Policy Review

9:01

[Vol. 3

Indeed, some judges will say that they send people to prison for rehabilitation purposes.
TABLE 1. SEVENTEEN STATE REFORMS

TO

REDUCE INCARCERATION

Arizona

Established probation revocation and crime reduction performance
incentive system

Arkansas

Declared marijuana enforcement lowest law enforcement priority
(Fayetteville)

Colorado

Amended criminal code to permit certain juveniles charged with
murder to have their cases adjudicated in the Youthful Offender
System

Connecticut

Authorized racial and ethnic impact statement to be prepared in
conjunction with certain criminal justice legislation

Hawaii

Declared marijuana enforcement lowest law enforcement priority
(Hawaii County)

Illinois

Created Commission to Study Disproportionate Justice Impact

Iowa

Authorized racial and ethnic impact statement to be prepared in
conjunction with certain criminal justice legislation

Kentucky

Amended parole release policies and expanded home incarceration
for persons convicted of certain offenses; created committee to
study Kentucky Penal Code and make recommendations for
reform; rescinded certain requirements for persons seeking to have
voting rights restored after the completion of sentence

Louisiana

Expanded dismissal of prosecution to persons who have completed
a drug court diversion program

Massachusetts

Declared marijuana enforcement lowest law enforcement priority

Mississippi

Amended parole release policies; expanded eligibility for
compassionate release

New Jersey

Expanded drug court eligibility and permitted early termination of
probation supervision for persons making exemplary progress

Pennsylvania

Created Recidivism Risk Reduction Incentive sentence to provide
for accelerated release for eligible individuals upon completion of
certain programs

South Carolina Established the South Carolina Sentencing Reform Commission
Utah

Appropriated state funds for the provision of postsecondary
education for persons in prison

Vermont

Expanded substance abuse programming for persons in prison and
under community supervision and permitted courts to reduce
probation sentence for persons making progress under supervision

Wisconsin

Established a coordinated strategy for the collection and analysis
of criminal justice data for the purposes of identifying unwarranted
racial disparities and created a Racial Disparities Oversight
Commission

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 9

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

315

The more important point is that rehabilitation programs, no matter how
good, cannot by themselves truly substantially reduce incarceration rates by
reducing recidivism. It is true that a growing body of work shows, quite
persuasively, that certain kinds of programs reduce recidivism rates more
than trivially.27 Especially if our aim is to help those who have broken the
law become restored to their communities as pro-social citizens, we would
be unwise to ignore this work. But even in a best-case scenario—meaning
that rehabilitation programs now offered to a fraction of the population
would be brought to scale, so that every person behind bars would be exposed to an effective program—rehabilitation programs would have relatively modest effects. Recidivism rates might fall from forty percent to
thirty-two percent. By extension, instead of 300,000 parole violators a year,
there would be 240,000. All other things remaining equal, instead of 700,000
prison admissions a year, there would be 640,000.28 This is a non-negligible
improvement but hardly a program that would bring mass incarceration
down to a new homeostasis. To achieve even these reductions would require
bringing rehabilitation programs to scale for everyone at an unrealistic rate,
given current knowledge and capacity. In other words, rehabilitation may be
the right thing to do but not because it overcomes mass incarceration.
V. ALTERNATIVES

TO

INCARCERATION

A panoply of strategies, generally thought of as “alternatives to incarceration,” have been offered to entice judges to place offenders into community programs rather than incarceration, including, for example, intensive
probation programs and drug treatment diversion programs. However, these
programs rarely substantially replace incarceration and drive down incarceration rates. There are two reasons. First, to be politically feasible, most “alternatives” have to promise to be tough and uncompromising. As a result,
they end up having high rates of “technical” failures, which occur when
program participants are unable to live under the programs’ strict rules.29
Second, these strategies typically promise not to put the public at risk, so
they forego dealing with serious law violators and instead deal with lawbreakers who would not be sent to prison anyway. In the first case, these
tougher “alternatives” increase incarceration by sending people back to

27
LAWRENCE SHERMAN ET AL., U.S. DEP’T OF JUSTICE, NAT’L INST. JUSTICE, PREVENTING
CRIME: WHAT WORKS, WHAT DOESN’T, WHAT’S PROMISING 6–12 (1998), available at http://
www.ncjrs.gov/pdffiles/171676.pdf.
28
For estimates of the size of the rehabilitation effect, see Don A. Andrews & Craig
Dowden, Managing Correctional Treatment for Reduced Recidivism: A Meta-analytic Review
of Programme Integrity. 10 LEGAL & CRIMINOLOGICAL PSYCHOL. 173 (2005) and Mark Lipsey
& Francis T. Cullen, The Effectiveness of Correctional Rehabilitation: A Review of Systematic
Reviews, 3 ANN. REV. L. & SOC. SCI. 297 (2007).
29
Susan Turner et al., Evaluating Intensive Supervision Probation/Parole (ISP) for Drug
Offenders, in 38 CRIME & DELINQUENCY 539, 553–554 (1992).

\\server05\productn\H\HLP\3-2\HLP209.txt

316

unknown

Seq: 10

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

prison at higher rates; in the second, they are irrelevant to incarceration
rates.
A. Reentry Programs
Interest in reentry programs has been intense ever since the National
Institute of Justice announced about a decade ago that nearly 700,000 people
reenter communities from prison each year.30 It was as though nobody had
realized before how many people were returning to communities from
prison. Researchers and policymakers have become focused on this issue.
The potential for improving the prospects of those reentering society from
prison is significant now that the problem has surfaced and received so much
public attention, but there is at least an equal chance that this new attention
will backfire by promoting the usual kinds of changes: closer surveillance,
more restrictions, and greater emphasis on being “tough.”
But the central problem is even more basic. Even if reentry programs
are wildly successful—and there is no reason to think they will be, given
current studies—they cannot solve the problem of mass incarceration.31 The
inescapable fact is that reentry comes after the person has already gone to
prison. We have 700,000 people in reentry each year because we have removed each of them in prior years. Our concern with reentry cannot solve
the problem of mass imprisonment because it is, itself, a consequence of
mass imprisonment.
B. The Obvious Need: Sentencing Reform
Any solution to the problem of mass incarceration must begin with two
points. First, programmatic tinkering has not reduced the prison population
to date, and it will never have much effect, even under the most optimistic
assumptions. Second, to overcome mass incarceration requires that we incarcerate fewer people. There is no getting around it. If the problem is mass
imprisonment, then the solution is to change the laws that send people to
prison and sometimes keep them there for lengthy terms. That means reducing the number going in, their length of stay, or both.
1. The Number Going In
There are two main ways to reduce the number of people entering
prison. They are, first, to eliminate mandatory sentencing and, second, to
eliminate technical revocations of probation and parole.

30
JEREMY TRAVIS, BUT THEY ALL COME BACK: FACING THE CHALLENGES OF PRISONER
REENTRY 35 (2005). The number of people released from prison has been growing and now
exceeds 720,000.
31
James A. Wilson & Robert C. Davis, Good Intentions Meet Hard Realities: An Evaluation of the Project Greenlight Reentry Program, 5 CRIMINOLOGY & PUB. POL’Y 303–38 (2006).

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 11

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

317

a. Mandatory Sentencing
The main reason for prison growth in the United States in the 1980s
was a reduction in the use of probation as a sentence for people convicted of
felonies, especially drug crimes. Before laws mandating prison sentences in
place of probation came into effect, drug offenders were a small fraction of
the U.S. prison population, about six percent; they are now about twenty
percent.32
Eliminating mandatory prison terms across the board would have a substantial impact on the size of the prison population. Much of this impact
would come from having fewer people serve time for drug-related crimes.
Enabling judges to choose non-prison penalties for other kinds of felonies
would also have an effect on the size of the prison population, but because
many of these felonies are serious enough to warrant some loss of freedom—and often prison sentences result from defendants’ prior felony
records—the overall impact of eliminated non-drug mandatory penalties
would likely be small.
b. Technical Revocation of Probation and Parole
When a person who is under community supervision fails to comply
with the requirements set by probation or parole, such as “reporting as directed,” the privilege of community supervision can be revoked, and the
person can be sent back to prison or jail, even if no new crime is alleged.
This is referred to as a “technical revocation” of probation or parole, and at
least one-third of prison admissions come by this route.33 It is notable that
these prisoners are people who have not been convicted of new crimes but
are returned to prison as a consequence of rules violations. Prohibiting reincarceration for technical violations of probation or parole could cut the number of prisoners substantially.
Undoubtedly, in some of these cases, the probation or parole violation
involves the commission of a new crime; in such cases, violators will likely
end up in prison if the option of technical revocation is eliminated. In California, for example, it has been estimated that as many as eighty percent of
“technical” violations involve allegations of criminal misconduct.34 Yet one
study of parole revocations in that state found that only sixteen percent of
those criminal allegations involved serious and/or violent felonies. The other
alleged crimes are less serious; they are the kinds of crime that quite often
result in sentences of probation when fully prosecuted.35 And California, by
all accounts, presents the most extreme version of this overuse of technical
32

See CLEAR, supra note 1, at 54–56.
JACOBSON, supra note 8, at 146.
34
See Joan Petersilia, California’s Correctional Paradox of Excess and Deprivation, 37
CRIME & JUST.: REV. RES, 206 (2008)
35
JEREMY TRAVIS, URBAN INST., PAROLE IN CALIFORNIA, 1980–2000: IMPLICATIONS FOR
REFORM (2003), available at http://www,urban.org/UploadedPDF/900598_Parole_in_Calif.
pdf.
33

\\server05\productn\H\HLP\3-2\HLP209.txt

318

unknown

Seq: 12

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

violations for alleged misconduct; in other states the problem is less dramatic, and technical revocations are less likely to reflect uncharged criminal
activity.36 As many as one-fourth of technical violations nationwide involve
no allegation of new criminal conduct.37
The problem of misconduct on supervision has been the subject of extensive discussion. Most experts recognize that cycling minor violations,
even criminal ones, through prison is expensive and counterproductive. They
propose developing a range of sanctions that replace return to prison and yet
reinforce the importance of complying with supervision requirements and
avoiding infractions. Referred to as “graduated problem-solving responses,”
these strategies deal with misconduct through a variety of in-community
controls in place of return to prison. It has been estimated that most lowlevel violations, even those that involve minor crimes, can be dealt with
through such mechanisms.38
If technical revocations are eliminated and graduated strategies put into
place, the rate of parolees returning to prison will be cut substantially, perhaps as much as two-thirds. And because there is no evidence that technical
revocations prevent crime, a policy that eliminates them might be pursued
with minimal public safety implications.
2. The Length of Incarceration
In the last thirty years, the average time served by people going to
prison has almost doubled, and the amount of time felons spend under parole
supervision has also increased.39 Lengthening prison terms and high rates of
technical revocation have resulted in a consistently growing prison population, even in the face of declining crime rates. Considerably rolling back
the length of prison terms would leave the United States with a smaller
prison population and a punitive policy more in line with those of other
Western democracies. Because decreasing a prisoner’s length of stay does
not lead to an increased chance of recidivism (if anything, the relationship is
in the other direction), and because almost everyone going to prison gets out
eventually, we can reduce sentence lengths substantially without affecting
crime rates or prison reentry rates.40
A crucial policy target is extremely long sentences. They used to be rare
in the United States but are becoming more common. Currently, 140,000
prisoners are serving life terms (twenty-eight percent without possibility of
parole), and one-fifth of prisoners serve sentences with a minimum term of
twenty-five years or longer. People who receive sentences of this magnitude
36

See JACOBSON, supra note 8.
JOAN PETERSILIA, WHEN PRISONERS COME HOME: PAROLE AND PRISONER REENTRY
(2003). See also Petersilia, supra note 35.
38
See, e.g., AMY SOLOMON, ET AL., URBAN INST., PUTTING PUBLIC SAFETY FIRST: 13
PAROLE SUPERVISION STRATEGIES TO ENHANCE REENTRY OUTCOMES (2008), available at
http://www.urban.org/UploadedPDF/411791_public_safety_first.pdf.
39
See Austin, supra note 1, at 12, 20.
40
See Nagin et al., supra note 18, at 51–56.
37

R

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 13

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

319

nearly always have committed atrocious acts that shock the conscience,41
and their lengthy sentences often result from decisions imposed at a time of
heightened public outrage in the face of the crime. Yet with little exception,
the outer years of these terms have no public safety value—most people who
serve long sentences and reach their forties or fifties pose little threat to the
public.42 Placing an upper limit on sentences and making release more readily available to people in their fifties would help reduce incarceration with
little effect on public safety.
These are the implications of the iron law. To affect the size of prison
populations, we have to change the dynamics that produce prison populations by altering how many people go to prison and how long they stay.
This means reform of sentencing and parole practices. We have already recommended placing upper limits on sentences and making release more readily available for older prisoners. What other specific reforms might be
implemented?
a. The Big Three: Length of Stay, Mandatory Penalties for Drugs,
and Technical Revocation
In a recent report, we showed the likely impact on prison populations of
adjusting three major pressure points on the prison population.43 We chose
those pressure points because they, more than any others, were the places of
change in the 1980s and 1990s that produced the prison population we have
today. Here, we summarize these recommendations to illustrate the power of
the iron law as an organizing principle for prison reform.
Table 2 shows the cumulative effects on state prison populations of implementing the following reforms over a five- to eight-year period of time:
1. Reduce length of stay (LOS) for sentenced prisoners to 1988
levels;
2. Divert technical parole and parole violators from prison and reduce
their LOS;
3. Eliminate mandatory penalties for drug crimes.
These reforms would cut the incarceration rate by about half. They would
also reduce the state and federal prison populations by over 400,000 inmates.

41
The exception is prisoners convicted under extreme versions of three-strikes legislation,
such as that of California. Cal. Penal Code §§ 667(e)(2)(A), 1170.12(c)(2)(A)(ii) (1994).
42
See Alfred Blumstein & Kiminori Nakamura, Redemption in the Presence of Widespread Criminal Background Checks, CRIMINOLOGY (forthcoming Aug. 2009).
43
See AUSTIN ET AL., supra note 1.

\\server05\productn\H\HLP\3-2\HLP209.txt

320

unknown

Seq: 14

24-AUG-09

Harvard Law & Policy Review
TABLE 2. PROJECTED EFFECTS

OF

State Prison Admissions

Admits
684,656

New Court Admissions

351,510

[Vol. 3

THREE RECOMMENDED REFORMS

Current Length of Stay and
Admissions
Total

9:01

LOS

Prisoners

All Reforms Plus
Diversion of Victimless
Crimes
Admits

1,401,605

418,109

LOS

Prisoners
731,077

860,321

311,965

21

551,168

Violent

91,393

49

373,186

91,393

34

261,231

Property

101,938

23

195,381

101,938

16

136,767

Drug

112,483

23

215,593

84,362

16

113,186

45,696

20

76,161

34,272

14

39,984

351,931

94,310

21

176,950

Public Disorder
Technical Probation Violators

143,792

Violent

37,386

49

152,659

33,647

34

96,175

Property

41,700

23

79,924

31,275

16

41,960

Drug

46,013

23

88,192

25,883

16

34,726

Public Disorder

18,693

20

31,155

3,505

14

4,089

189,354

12

189,354

11,835

3

2,959

Technical Parole Violators
U.S. Population

303,000,000

303,000,000

463

241

Incarceration Rate per 100,000
citizens

b. Companion Efforts: Community Penalties
Two additional reforms might be enacted as complements to the big
three strategies described above. They are complementary in the sense that
they are not directly related to the iron law; they do not alter the rate of
prison commitments or the length of stay in prison. But they do relieve pressure on the non-prison alternatives to incarceration, thus enhancing the capacity of these alternatives to absorb the increased demand they will
experience when the “big three” are put into effect. The complementary
strategies are:
1. Reduce length of stay for persons placed on probation and parole;
2. Make greater use of fines, restitution, and community service in
lieu of probation.
Presently, there are over five million people on probation and parole.
This number is about three times higher than it was in the 1970s, when the
prison boom began.44 Thus, the same kind of expansion that took place with
regard to prison populations has also taken place in the use of these nonimprisonment criminal penalties. These penalties are subject to their own
version of the “iron law”: the size of probation and parole populations de44

See CLEAR, supra note 1, at 15–48

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 15

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law

9:01

321

pend on the number of people placed under community supervision and their
length of stay under that sanction. A companion effort to the one designed to
reduce prison populations could target reductions in the size of community
supervision populations.
Studies show that the effects of parole supervision on recidivism fade
after about a year, and longer supervision periods are not associated with
higher success rates.45 This means that community supervision caseloads
can be reduced substantially, merely by reducing length of stay on supervision.46 Finally, use of fines, restitution, and community service as the sole
sanctions would make our penalties look more like those of other Western
countries.47
The impact of these additional changes on parole and probation populations is shown in Table 3.
TABLE 3. PAROLE

Population

PROBATION NUMBERS
ADMISSIONS AND LOS

AND

Current LOS and
Admissions

New LOS – Only
Population

CHANGES

IN

Plus Diverted Probationer/
Parolees

Admits

LOS

Admits

LOS

Admits

LOS

2,362,100

21

4,190,896

2,362,100

14

2,755,783

1,771,575

14

2,066,838

Parole

512,800

17

710,065

512,800

10

427,333

384,600

10

320,500

Totals

2,874,900

4,900,961

2,874,900

3,183,117

2,156,175

Probation

Population

AFTER

Population

2,387,338

c. Projected Impact on Mass Incarceration
Table 4 shows the grand totals for all correctional populations, including the federal prison system and the jails, that would follow from adoption
of our strategies.48 We have built in the reductions indicated in the two preceding tables. We have also projected modest reductions in the nation’s jail
population based on what appears to be a growing trend of reduced arrest
rates, which fuel jail bookings.49 If our proposals are implemented, there
will also be fewer jail bookings, as fewer probationers and parolees will
become violators and be required to remain in jail until they receive their
revocation hearings. The federal prison population is also projected to de-

45
See, e.g., Avinash Singh Bhati, Investigating the Dynamics of the Criminal Recidivism
Process (Urban Inst. Working Paper, 2004), cited in Joan Petersilia, Employ Behavioral Contracting for “Earned Discharge” Parole, 6 CRIMINOLOGY & PUB. POL’Y. 807, 810 (2007);
Patrick C. Jackson, Some Effects of Parole Supervision on Recidivism, BRIT. J. CRIMINOLOGY,
Jan., 1983, at 17.
46
AUSTEN ET AL., supra note 1, at 24.
47
See Sally T. Hillsman, Fines and Day Fines, 12 CRIME & JUST. 49, 50 (1990).
48
See Austin, supra note 1, at 35.
49
See id.

\\server05\productn\H\HLP\3-2\HLP209.txt

322

unknown

Seq: 16

24-AUG-09

Harvard Law & Policy Review

9:01

[Vol. 3

crease as Truth in Sentencing laws are relaxed,50 and judges make greater use
of probation. With these two additions, as shown by the chart, an overall
fifty percent reduction in correctional population will be achieved by 2016.
Of course, these reductions would occur gradually, and the amount of time
they would require would depend on the precise nature of the policy reforms
chosen to achieve them. Policy reforms requiring the enactment of new
legislation or modification of existing laws would take longer to implement,
and their effects would therefore be more delayed.
TABLE 4. SUMMARY OF THE RECOMMENDATIONS’ IMPACT
CORRECTIONAL POPULATIONS

Correctional
Populations

ON

New LOS
and
Diversions

Plus Diverted
“Victimless”
Crimes

% Change

4,071,866

3,842,879

-49%

Current

New
LOS

Grand Totals

7,496,425

5,126,628

State Prisons

1,398,698

1,037,930

803,948

731,077

-48%

199,618

125,000

100,000

100,000

-50%

Federal Prison
Probation

4,293,163

2,755,783

2,066,838

2,066,838

-52%

Parole

824,365

427,333

320,500

320,500

-61%

Local Jails

780,581

780,581

780,581

624,465

-20%

It is worth noting that there is nothing truly revolutionary about these
proposals. They are “evidence-based,”51 having been shown to be effective
in reducing correctional populations without adversely impacting on crime
or recidivism rates.52 The point follows: these proposals are essentially public safety neutral. To the degree that an increase in the release cohort will
carry with it a slight increase in felony commissions from that cohort, at
least in the short term, there is an effect. But, as discussed in Part I, it is
entirely an effect of moving crimes to an earlier point in time, as releases
occur earlier. Once the release cohort bubble has worked its way through the
system, people in reentry will begin to account for fewer crimes, because
there will be fewer people in reentry. That is, over a short period of time, any
increases in crime due to increases in prison releases will be counteracted by
equivalent decreases in crime as the number of releases shrinks to a smaller
level than before.

50

See JACCOBSON, supra note 8.
For a review of evidence-based policies in corrections, see Center for Evidence-Based
Corrections, http://ucicorrections.seweb.uci.edu/pubs (last visited Jul. 10, 2009)
52
Indeed, a strong case can be made that public safety could be greatly enhanced using
the billions of dollars that would be saved each year by the states. They could invest those
funds in preventive programs in the communities that produce the vast majority of admissions
to the correctional system.
51

\\server05\productn\H\HLP\3-2\HLP209.txt

unknown

Seq: 17

24-AUG-09

2009] Reducing Mass Incarceration: Prison Populations’ Iron Law
PAST

AND

9:01

323

PROJECTED DECLINES IN CORRECTIONAL POPULATIONS
1988-2016
1,600,000

7,000,000

1,400,000

6,000,000

1,200,000

5,000,000

1,000,000

4,000,000

800,000

2,000,000

600,000

2,000,000

400,000

1,000,000

200,000

Prison, BOP, Jail, Parole

8,000,000

0

0
1988 1990
Total

1995
Probation

2000

2005
State

Federal

2010

2015 2017
Parole

Jail

VI. DOING SOMETHING ABOUT MASS INCARCERATION
It is clear that the criminal justice system today feeds upon itself. There
are many ways this is true. Increased police presence in a “troubled” neighborhood increases the probability of detecting violations of the law and subjecting people in those places to criminal justice. When concentrations of
arrest get to a certain point, crime begins to rise rather than fall, creating
more eligibility for detection and arrest. This cycle is fueled as more adults
go to prison in places with more police presence and thus more arrests. In
turn, the chances of a child becoming involved in delinquency are increased
by having a parent go to prison, in part because socialization and adjustment
are affected by loss of the parent. Being exposed to neighborhood violence
and being victimized as a child increase the chances of adult criminality.
Similarly, a child who becomes involved in the formal juvenile justice system as an adolescent is at greater risk of criminal activity in early adulthood.
People who are convicted of crimes as juveniles are targeted for prison terms
as recidivists; people who go to prison are more likely to recidivate once
again. As crime grows, pressure for more aggressive policing concentrated
in problem communities increases, and pressure for more stringent penalties
does as well. And so the cycle begins again.53
The cycle described above is now widely viewed as a serious problem,
and a situation from which we must move away. But how do we do that?
The answer is simpler in principle than we might think, but it will be more
53

For research in support of these conclusions, see generally CLEAR, supra note 1.

\\server05\productn\H\HLP\3-2\HLP209.txt

324

unknown

Seq: 18

Harvard Law & Policy Review

24-AUG-09

9:01

[Vol. 3

challenging to implement than we might expect. We got to this place as a
consequence of a generation of policymaking. It will take a sustained effort
of policymaking to get us out.
There is a temptation to look toward the penal system for the policies
we need: better rehabilitation programs, high quality reentry practices, and
strong and attractive alternatives to incarceration. These are worthy ideas,
but they will not cure the ills that undergird mass incarceration. If we are to
take meaningful steps toward decreasing the size of the U.S. prison population, we must change the incarceration policies that produce it.
We must reduce the number of people who go to prison and the length
of time they stay there.