12 Prison to Work Western

8/3/2019 12 Prison to Work Western

1/38

HAMILTONTHE

PROJECT

Advancing Opportunity,

Prosperity and Growth

The Brookings Institution

From Prison to Work:A Proposal for a NationalPrisoner Reentry Program

D I S C U S S I O N P A P E R 2 0 0 8 - 1 6 D E C E M B E R 2 0 0 8

Bruce Western


2/38

The Hamilton Project seeks to advance Americas promise o

opportunity, prosperity, and growth. The Projects economic

strategy reects a judgment that long-term prosperity is best

achieved by making economic growth broad-based, by

enhancing individual economic security, and by embracing a

role or eective government in making needed public

investments. Our strategystrikingly dierent rom the

theories driving economic policy in recent yearscalls or fscal

discipline and or increased public investment in key growth-

enhancing areas. The Project will put orward innovative

policy ideas rom leading economic thinkers throughout the

United Statesideas based on experience and evidence, not

ideology and doctrineto introduce new, sometimes

controversial, policy options into the national debate with

the goal o improving our countrys economic policy.

The Project is named ater Alexander Hamilton, the

nations frst treasury secretary, who laid the oundation

or the modern American economy. Consistent with the

guiding principles o the Project, Hamilton stood or sound

fscal policy, believed that broad-based opportunity or

advancement would drive American economic growth, and

recognized that prudent aids and encouragements on the

part o government are necessary to enhance and guide

market orces.

HAMILTONTH E

PROJECT


3/38

DECEMBER 2008

HAMILTONTHE

PROJECT

Advancing Opportunity,

Prosperity and Growth

NOTE: Thi dicion paper i a propoal rom the athor. A emphaized in The Hamilton

Project oriinal tratey paper, the Project wa deined in part to provide a orm or leadin

thiner acro the nation to pt orward innovative and potentially important economic policy

idea that hare the Project broad oal o promotin economic rowth, broad-baed partici-

pation in rowth, and economic ecrity. The athor are invited to expre their own idea in

dicion paper, whether or not the Project ta or adviory concil aree with the pecifc

propoal. Thi dicion paper i oered in that pirit.

From Prison to Work:A Proposal for a NationalPrisoner Reentry Program

Brce Wetern


4/38

From Prison to Work: A ProPosAl For A nAtionAl Prisoner reentry ProgrAm

2 THE HAMILTON PROJECT|

THE BROOkINgs INsTITuTION

Copyriht 2008 The Brooin Intittion

Aac

Around seven hundred thousand mostly low-income and minority men and women are released rom

prison each year. Returning to lives o low wages and high rates o unemployment, about two thirds

will be rearrested within three years. I propose a national prisoner reentry program whose core ele-

ment is up to a year o transitional employment available to all parolees in need o work. Transitional

jobs are supplemented by substance-abuse treatment and housing ater release, expanded work and ed-

ucational programs in prison, and the restoration o eligibility or ederal benets or those with elony

records. The program costs are oset by increased employment and reduced crime and correctional

costs or program participants. By shiting supervision rom custody in prison to intensive programs

in the community, the national reentry program improves economic opportunity and reduces prison

populations.


5/38


WWW.HAMILTONPROJECT.ORg|

DECEMBER 2008

C

Introdction 5

1. The Problem o Ma Imprionment and Pot-Prion Employment 6

2. Evidence on Prioner Reentry Proram 10

. A Propoal or a National Prioner Reentry Proram 14

4. Cot and Beneft 2

5. Objection and Alternative 27

6. Conclion 28

Appendix: Cae stdie 29

Reerence 2


6/38





7/38



DECEMBER 2008 5

In the current era o mass incarceration, low-in-

come young men with little schooling are per-vasively involved in the criminal justice system.

Those returning rom state or ederal prison ace

high rates o unemployment and recidivism. Both

these measuresunemployment and recidivism

refect the acute challenge o reentering society and

assuming mainstream social roles.

I propose a national prisoner reentry program

whose main element is a year o community service

employment buttressed by transitional services and

in-prison education. The national prisoner reentryprogram aims to increase employment among re-

leased prisoners while reducing prison populations.

Achieving these objectives will yield a sustainable

public saety that overcomes the long-term nega-

tive consequences o criminal punishment and pro-

motes the economic improvement o poor commu-

nities.

idc


8/38




The growth o the penal system over the past

thirty years has redrawn the landscape o ur-ban poverty in America. Prison and jails now

hold 2.25 million inmatesmostly minority and

poorly educated young men. Swelled largely by

drug oenders and parole violators, state and ed-

eral prisons return more than seven hundred thou-

sand prisoners each year to inner-city communities

across the country. Although growth in the prison

population has helped reduce crime rates over the

past decade, todays penal system presents two re-

lated challenges or public policy.

First is the problem o prisoner reentry. In the late

1970s around one hundred and ty thousand in-

mates were released rom state or ederal prison

each year. Today, that number is about ve times

as large. These enlarged cohorts o released pris-

oners return overwhelmingly to inner-city neigh-

borhoods o concentrated poverty where jobs are

scarce, crime rates are high, and social disorganiza-

tion is itsel deepened by the population turnover

associated with mass incarceration. Under these

conditions, the benets to impoverished amiliesand communities o post-prison employment are

potentially large.

Frequently returning to social and economic ad-

versity, ormer prisoners themselves are poorly

equipped to lead productive lives. Mostly minorites

and aged in their thirties or older, prisoners aver-

age about a tenth-grade education (Table 1). Sur-

vey data show that about one-third o state prison

inmates were jobless and two-thirds had a history

o heavy drug or alcohol use at the time o theirincarceration (U.S. Department o Justice 2004).

The disadvantage o prisoners is also indicated by

chronic health problems, high rates o mental ill-

ness, and cognitive scores well below grade level.

Further, prisoners have very little work experience,

even compared to others with similar schooling and

demographic characteristics.

1. th P ma ip ad P-Pep

Ater returning home, ex-prisoners are out o work

about hal the time, earn on average around $9,000

a year, and experience virtually no growth in earn-

ings (Western 2006, ch. 5). Prison time itsel may

impede successul reintegration into society; studies

show that incarceration is associated with reduced

earnings and employment rates, and increased rates

o divorce and separation (Western 2006). Perhaps

ailure ater release rom prison is indicated most

vividly by recidivism rates: the 1994 national re-cidivism study showed that more than two thirds o

ormer state prisoners were rearrested within three

years o release, and hal o those rearrested were

back in prison within that time (Langan and Levin

2002).

Whereas the problem o prisoner reentry has grown

tAble 1.

Daphc, idca s ad ep-a, ad Pa Pacpa saP

1991 2004

Demographics

Blac (percent) 47.5 4.2Latino (percent) 15.5 19.4Median ae (year) 0.0 4.0

Skills and employability

Averae choolin (year) 10.4 10.4Employed beore

imprionment (percent) 67.2 72.4Reportin heavy dr e

(percent) 62.2 69.1Program participation

sbtance-abe treatment

(percent) 6.4 71.2Wor or edcation proram(percent) 44.2 25.

sorce: srvey o Inmate o state Correctional Facilitie, 1991 and 2004

(Brea o Jtice statitic 199, 2007).


9/38



DECEMBER 2008 7

with the incarceration rate, assistance or prisoners

and their amilies has contracted. Resources or ed-

ucational and other rehabilitative programming in

prison have shrunk, and social services ater release

vary substantially across jurisdictions. Despite the

acute human capital decits o prisoners, participa-tion in work and education programs has declined

rom 44 percent in 1991 to 25 percent in 2004. Be-

cause o these changes, released prisoners may be

less prepared or the labor market than they were

in the past. Adding to the challenge o prisoner re-

entry, the labor market or low-skilled men has de-

teriorated. Earnings among men with only a high

school education have stagnated and joblessness

among young non-college-educated blacks remains

persistently high.

The second public policy challenge o todays penal

system is presented by the scale o correctional ex-

penditures in state budgets. The growth o the pris-

on population has changed the unctions o state

governments. For most o the twentieth century,

the imprisonment rate in the United States hovered

around one hundred per one hundred thousand

(Figure 1a). From 1975 to 2005, the raction o the

population in prison grew ve-old, and the costso corrections ballooned. In 2005, total correctional

spending was $70 billion, up rom $19 billion (in

2007 dollars) in 1982. This represents an average

annual cost o about $27,000 per prison inmate.

Increased spending on prisons means ewer re-

sources or other budget priorities. Spending on

corrections as a share o states general unds in-

creased about 40 percent rom 1987 to 2007 (Fig-

ure 1b). Over this same period, spending on higher

education as a share o state spending declined byabout 30 percent. These gures indicate a shit in

priorities away rom human capital investment to-

ward criminal punishment.

Figure 1(A)

ip ra, 19252005 (p 100,000 u.s. ppa)

Fire 1(a): Imprionment per one hndred thoand o the u.s. poplation, 1925 to 2005

1925 1935 1945 1955 1965 1975 1985 1995 2005

0

100

200

300

400

500

600

ImprisonmentRate(per100,0

00)


10/38




The policy problems o reentry and rising cor-rectional budgets are the most visible signs o the

challenge to social justice created by extraordinary

rates o incarceration among young black men.

Black men are seven times more likely to be incar-

cerated than white men, and large racial disparities

can be seen or all age groups and at dierent levels

o education. The large black-white disparity in in-

carceration is unmatched by most other social indi-

cators. Racial disparities in unemployment (two to

one), nonmarital childbearing (three to one), inant

mortality (two to one), and wealth (one to ve) areall signicantly lower than the seven-to-one black-

white ratio in incarceration rates. Among black men

under age orty, around one in nine is currently be-

hind bars in prison or jail. Among black male high

school dropouts under orty, one in three is incar-

cerated. Over a lietime, about one in ve black men

born since 1965 will serve time in prison. Indeed,

black men are now more likely to go to prison thanto graduate rom college with a our-year degree

(Western 2006, p. 29). At the very bottom o the edu-

cation distribution, a third o non-college-educated

black men and two-thirds o black male high school

dropouts born since 1965 will go to prison at some

point in their lives.

The historically novel normality o imprisonment

or young black men with little schooling was pro-

duced by a newly punitive criminal justice policy

applied most zealously in poor urban neighbor-hoods that oered ew legitimate economic oppor-

tunities. Mandatory minimum sentencing, truth in

sentencing, and habitual oender enhancements

or those on their second and third strikes in-

creased prison commitments among those arrested,

and increased time served among those in prison.

Through the 1990s the growth in the incarceration

Figure 1(b)

sa Cca ad Hh edcaa spd, 19872007 (pc)

Fire 1(b): state pendin on correction and hiher edcation a a percentae o tate eneral nd, 1987 to 2007.

sorce: National Aociation o state Bdet Ofcer (1996, 2007); Patore and Maire (2008).

1987 1992 1997 2002 2007

0

2

4

6

8

10

12

14

16

18

GeneralFundSpending(%)

Higher EducationCorrection


11/38



DECEMBER 2008 9

rate was swelled by increasing rates o parole revo-

cation (Blumstein and Beck 2005). As the criminal

justice system became more punitive, high levels o

joblessness exposed young low-skill men in inner

cities to the scrutiny o the police, the lure o il-

legal income, and the disorder o chronic idleness(Western 2006, ch. 3). Harsh punishment and the

jobless ghetto combined to produce the mass im-

prisonment o young black men with no more than

a high school education.

The penal system now refects the contours o se-

vere disadvantage among young men, and deepens

inequality by diminishing the lie chances o those

with prison records. In the long run, public saety

itsel is threatened by mass incarceration because

those released rom prison have trouble joining themainstream o social lie. Increasing employment

and reducing crime among those released rom

prison has become central to improving economic

opportunity among todays urban poor, and central

to reducing the scale o a penal system that now

shapes the lie path or a generation o young black

men.


12/38




The problem o prisoner reentry is an active

area o policy interest; many programs orimproving employment and reducing recidi-

vism have been proposed. For example, the Reen-

try Policy Council (2005) provides an encyclopedic

discussion and makes dozens o policy recommen-

dations. A wide variety o work, training, and educa-

tion programs, in prison and ater release, orms an

uneven patchwork o services that requently oper-

ate with only limited success.

Recent reviews oer a mixed assessment o the

eects o reentry programs on employment. DanBloom (2006) observes that there have been ew

randomized evaluations, although ex-prisoners

were sometimes included in studies o programs

that were more broadly designed to assist disadvan-

taged workers. Visher, Wintereld, and Coggershall

(2005) review eight random-assignment studies o

employment-based programs and nd that the av-

erage eect on recidivism is small and insignicant.

Conversely, a broad survey by the British Depart-

ment o Education and Skills concludes that well-

designed programs successully promoted employ-ment among ex-prisoners, although evidence or

the eects on recidivism is weaker (Hurry, Brazier,

Parker, and Wilson 2006).

Despite conficting reviews, policy lessons can be

drawn rom a small number o well-designed stud-

ies (see Table 2). Many evaluations o programs or

prisoners report large reductions in recidivism, but

these results are oten artiacts o weak research de-

signs. In particular, selection into programs is oten

poorly controlled and program dropouts are otenignored. The evaluations on which I ocus are all

based on experimental or strong matching or re-

gression designs and report eects or all program

participants and not just program graduates.

Four kinds o programs have tried to increase

employment and reduce crime among those with

2. evdc P r Pa

criminal records: (1) transitional employment pro-

grams, (2) residential and training programs ordisadvantaged youth, (3) prison work and educa-

tion programs, and (4) income supplements or the

unemployed.

Transitional employment programs provide sub-

sidized work to parolees who work in small crews

under close supervision. An early randomized ex-

periment, the National Supported Work (NSW)

Demonstration (197578), placed parolees and

probationers in construction industry jobs. Three

years ater entry to the program, about 42 percento NSW clients over the age o twenty-six had

been rearrested, compared with 54 percent in the

control group (Uggen 2000). NSW participants

over age twenty-six were also less likely to report

illegal earnings. There were no signicant dier-

ences between program and control groups among

those aged twenty-six and younger. The value o

transitional employment or ex-prisoners is also

indicated by recent evaluations o the two transi-

tional jobs programs rom New York. An evaluation

o the CEO program (200405) ound that parol-ees entering transitional jobs experienced increased

employment and were 19 percent less likely to be

rearrested ater a year. However, this eect was only

ound or those entering the program within three

months o release rom prison (Bloom, Redcross,

Zweig, and Azurdia 2007). Because o small sample

sizes, these program eects were not signicant.

The ComALERT program (200406) in Brook-

lyn, New York, provides up to a year o subsidized

employment in combination with housing and sub-

stance-abuse treatment. Program participation wasassociated with signicant improvements in em-

ployment and a 18 percent reduction in arrest rates

compared to a matched control group with simi-

lar demographics and criminal history (Jacobs and

Western 2007). In sum, transitional employment

or up to six to twelve months immediately ater

prison release is associated with reduced recidivism


13/38



DECEMBER 2008 11

Pa sap z Dcp mhd Pa c

Transitional employment

NsW (197578) 1,497 Minimm wae Random 22% on arret i

contrction job ainment over ae twenty-ix;in mall pervied +6% (n..) on arretcrew or ex-prioner i ae twenty-ix or

releaed le than yonerix month

CEO (200405) 977 Minimm wae Random 19% (n..) on

manal job, job ainment arret or thoereadine trainin, enterin proram

and placement within three monthor parolee o prion releae,

+6% (n..) on arretor thoe enterinproram ater three

month o prion releae;+144% on employment

over a year

ComALERT (200406) 996 Minimm wae Matchin 18% on arret;manal job, dr +45% on uI

treatment, and hoin employment (N = 128)mandated to drtreatment

Other residential and training programs

OPTs (199497) 98 Family conelin, Random 16% (n..) on dr e,hoin aitance, ainment 7% (n..) on arret, +9%

job readine, and (n..) on ll-timeplacement or employment

probationer and parolee

Job Corp (199496) 2,450 Reidential edcation Random -% (n..) amonand trainin or hih ainment nonerio arretee,chool dropot aed 8% (n..) amon erio

ixteen to twenty-or, arretee on arret;bample with prior +10% or nonerio

arret arretee, -2%

(n..) or erioarretee onorth-yearemployment

JTPA (198789) 416 Claroom trainin, Random +6% (n..) on arret;on-the-job trainin or ainment no eect on uI earnin

job earch aitance,

incldin a bampleo male yoth arretee

tAble 2.

r ep ad ta Pa ex-P


14/38


12 THEHAMILTONPROJECT | THEbROOkINgs INsTITuTION

Pa sap z Dcp mhd Pa ffc

Prison Work and Education

PREP(198385) 750 Vocationaltraininor Matchin 24%onreincarceration atleat6month ater8to12yearor inmotlyclericalor inmateinprion

manalaricatin indtrie,33%or andrepairjo inmateinvocational

traininorapprentice- hip,23%(n..)or

inmateinprion indtrieandtrainin

FloridagED(199499) 12,956 In-priongEDclae Rereion +6%(n..)onqarterly andexam;gED uIearninaterone

radatecompared year;eectlihtly tohihchooldropot larer,thoh

withgED temporary,ornonwhite.

3-statestdy(199798) 3,170 In-prionproramin Matchin 16%onarret;+5% aicedcation,gED (n..)onemployment

preparation,lieill ateroneyearand andconitiveill, 30%onearnin; econdaryandpot- 3-yearprorameect

econdaryedcation orearninand employment

notinifcantIncome supplement programs

LIFE(197274) 432 unemployment Random 13%onarret;

eneft($252weely) ainment +12%onll-time orparolee employment

TARP(197577) 3,982 unemployment Random +3%(n..)onarret, eneft($250weely) ainment 25%onearnin andplacementor parolee

sorce:EvalationarereportedyManpowerDevelopmentReearchCorporation(1980)orNsW;uen(2000)ordierenteectyae;bloom,Redcro,Zwei,

andAzrdia(2007)orCEO;JacoandWetern(2007)orComALERT;Roman,sridharan,govi,bc,andMorley(1999)orOPTs(drtreatmentaloormedpart

oOPTs,tcontrolaloreceivedtreatment);schochet,brhardt,andglazerman(2001)orJoCorp;bloom,Orr,bell,Cave,Doolittle,Lin,andbo(1997)orJTPA;

saylorandgae(1997)orPREP;klinandTyler(2007)orFloridagED;andsterer,smith,andTracy(2001)orThree-statestdy;MallarandThornton(1978)orLIFE

(proramincldedjoplacementwhichwaineective);Roi,ber,andLenihan(1980)orTARP.

n..=Nottatiticallyinifcant.


15/38


WWW.HAMILTONPROJECT.ORg | DECEMbER 20 08 13

and increased employment, at least or the frst year

or two ater release. (CEO and ComALERT are

described in greater detail in the appendix.)

Whereas transitional employment or ex-prisoners

yields positive results, public service employmentprograms have improved employment and earnings

or other populations with only mixed success (Ell-

wood and Welty 2000, pp. 322331). Many large

employment and training programs involving crim-

inal oenders are ocused on youth. Most youth

involved in serious crime, however, are unlikely to

desist while they are still in late adolescence. The

eectiveness o transitional employment or those

in their late twenties and older is encouraging or a

reentry program or released prisoners, 80 percent

o whom are at least twenty-fve years old.

Other programs have combined several supportive

services like housing and drug treatment, though

not transitional employment, to move ex-prisoners

into the labor market. The Opportunity to Succeed

(OPTS) program (199497) provided mandatory

substance-abuse treatment in intensive residential

placements, as well as job readiness training. A year

ater random assignment, the treatment group had

accumulated an extra month o ull-time employ-

ment and were 9 percent more likely to have helda ull-time job. Recidivism was also modestly lower

in the treatment group, although the program eect

was not signifcant (Rossman, Sridharan, Gouvis,

Buck, and Morley 1999). Job Corps, targeting high

school dropouts under age twenty-fve, also provided

housing in combination with education and training

programs. Perhaps because participants were rela-

tively young, Job Corps ailed to produce signifcant

reductions in one-year arrest rates or to produce

signifcant increases in employment among those

with prior serious arrests (Schochet, Burghardt, andGlazerman 2001). The Job Training Partnership Act

(JTPA, 198789) provided training and job search as-

sistance similar to Job Corps, but in a nonresidential

setting. This less-intensive intervention had no eect

on the earnings and rearrest rates o male youth with

arrest records (Bloom, Orr, Bell, Cave, Doolittle,

Lin, and Bos 1997).

Three large-scale studies suggest the importance o

prison education. The PREP study (198385) ound

that participation in vocational training and work

programs was associated with reduced rates o rein-

carceration in ederal prison as long as twelve years

ater release (Saylor and Gaes 1997). The Three-State Recidivism Study (199798), named or study

groups in Maryland, Minnesota, and Ohio, exam-

ined a variety o educational programs, including ba-

sic education, GED preparation, and secondary and

postsecondary schooling. Although the study did not

distinguish the eects o dierent types o educa-

tional programs, those who participated in classes

in prison had only a 48 percent rearrest rate ater a

year, compared with a 57 percent rearrest rate or the

comparison group (Steurer, Smith, and Tracy 2001).

Program participants had higher earnings in the frstyear ater release, but this earnings advantage disap-

peared ater three years. Similar to the Three-State

Recidivism Study, the Florida GED study (199499)

ound no enduring gains to earnings or employment

or those who obtained a GED in prison. Still, some

immediate improvements in earnings were ound,

particularly or nonwhite GED holders (Kling and

Tyler 2007).

The main alternative to improving economic op-

portunities through work, housing, and educationhas involved paying unemployment benefts to re-

leased prisoners. Beginning in 1971, the Baltimore

LIFE (Living Insurance or Ex-Prisoners) experi-

ment (197274) randomly allocated released state

prisoners to a thirteen-week treatment consisting o

weekly $252 payments and job placement in some

cases, while a control group received no treatment.

Ater twelve months, 49.5 percent o the treatment

group had been rearrested, compared with 56.9 per-

cent o the controls (Mallar and Thornton 1978).

The LIFE program was replicated on a larger scalein Texas and Georgia in the TARP (Transitional Aid

or Released Prisoners) experiment (197577). The

TARP participants had higher rates o unemploy-

ment than the control group, however, and were no

less likely to recidivate (Rossi, Berk, and Lenihan

1980).


16/38




The striking result rom this survey o correc-

tional programming is the substantial uneven-ness o the programs eects. The programs

vary greatly in their content and in their clients.

Less-intensive interventions such as the income

supplements o TARP or the training o JTPA and

interventions directed at male youth have been un-

successul. More-intensive interventions tend to be

more successul, particularly i they target adult o-

enders who may be more motivated than younger

oenders to desist rom crime. The results rom

CEO and ComALERT also suggest that timely in-

terventions ocused on the period immediately aterprison release have a greater chance o success.

Timely and relatively long-term transitional em-

ployment appears promising because it addresses

perhaps the key barrier to steady post-prison em-

ployment: the very low level o work experience

among released prisoners. In many cases, men and

women coming out o prison have never held a reg-

ular, legitimate job. As a result, the rudimentary lie

skills o reliability, motivation, and sociability with

supervisors and coworkers are undeveloped. Otenwe think o these characteristics as noncognitive

skills that are ormed in childhood (Carneiro and

Heckman 2004). These noncognitive skills are as

important or success in the labor market as are the

more amiliar cognitive skills o math and verbal

ability (Heckman, Stixrud, and Urzua 2006). The

evaluation results or transitional jobs suggest that

the habits o everyday work and the noncognitive

skills on which they are based can be developed in

adulthood by the daily rehearsal o the routines o

working lie. Encouraged by the successul resultso timely and large-dose transitional jobs programs,

I propose up to a year o subsidized community ser-

vice employment or all parolees in need o work

as the centerpiece o a national prisoner reentry

program.

3. A Ppa a naa P r Pa

To oster work habits and tackle the problem behav-

iors o ormerly incarcerated men, several addition-al supports are needed. First, transitional housing

and substance-abuse treatment may enhance the

eectiveness o transitional jobs or the homeless

and drug addicted. Second, parole reorms that cur-

tail the reimprisonment o technical parole viola-

tors will acilitate the learning process in which new

noncognitive skills o reliability and persistence are

being developed. Third, prison education programs

should be expanded to improve readiness or tran-

sitional employment and work in the open labor

market. Finally, eligibility or ederal welare andeducation programs should be extended to those

with elony convictions.

In contrast to the prolieration o numerous small-

scale measures to assist ex-prisoners, the national

prisoner reentry program consists o a small num-

ber o large-scale measures that are intended to

work together as a system, moving prisoners out o

custody into the community. Unlike many reentry

proposals, my proposed program has the reduction

in prison populations as an explicit policy objective.The proposal also takes a realistic view o program

eectiveness. Transitional employment by itsel will

only modestly reduce recidivism and improve em-

ployment, and these eects may be short lived. Still,

the impact o transitional jobs can be enhanced by

supplementary services and supportive parole su-

pervision.

We can think o the national reentry program as a

sequence o stages that prepares people or work

in the open labor market. In this sequence, prisoneducation and discharge planning is preparatory or

transitional jobs and other services in the commu-

nity. The eectiveness o transitional jobs is sup-

ported by parole reorm, and the expansion o eli-

gibility or ederal programs.


17/38



DECEMBER 2008 15

i P: edca, W, adDcha Pa

To be prepared or transitional employment, pris-

oners must be equipped with basic literacy, job skills,

and rudimentary job readiness. Prisons have been,historically, a graveyard or rehabilitative criminal

justice. As correctional administrators know well,

the prisons main job is the sae and secure custody

o its inmates. As a result, even orderly and well-run

prisons can be unriendly contexts or teaching pro-

social behaviors. I we must choose between in-pris-

on and community programs, we should probably

spend our money in the community where program

eects or the ormerly incarcerated are larger and

the social benets distributed more widely.

Although the imperatives o custody may compro-

mise rehabilitation, the modest goals o literacy and

basic job skills may be achievable. State prisoners

average a tenth-grade education and score below

their grade level on cognitive tests. Improving the

cognitive skills o prisoners is thus an important

part o a post-prison employment program. In ad-

dition to providing work and education programs,

prisons can also play an important role at the time o

release by connecting inmates to their post-release

social supports.

By setting universal standards or adult education,

the Federal Bureau o Prisons oers a good model

or schooling in custody. Federal prisoners who are

unctionally illiterate or who lack a high school di-

ploma or GED are required to enroll in 240 hours

o educational programs. In 2004 about 40 percent

o new ederal prisoners were enrolled in education

programs, compared to 20 percent o new state pris-

oners. I propose a national minimum standard or

correctional education based on the ederal stan-dard. As in the ederal system, the national mini-

mum standard would aim to achieve a twelth-grade

level o unctional literacy or state prisoners. Such a

standard would help prepare prisoners or the labor

market, GED exams, and postsecondary schooling.

Compulsory correctional education throughout a

state prison system would require the availability

o basic education in virtually all acilities. Widely

oered standard programs would help inmates re-

main in class as they moved rom prison to prison.

Meeting the national standard would require a sig-

nicant expansion o state correctional education.

In most states, schooling is mandated only or pris-

oners under age twenty. States signing on to the na-

tional education standard would receive ederal aid

to make up the shortall between current spending

on correctional education and the level required to

meet the goal o 240 hours o basic education or

unctionally illiterate prisoners.

We can estimate the cost o this eort by using the

current levels o ederal spending on correctional

education and survey data on program participa-

tion by low-education state prisoners. Precise g-

ures are hard to determine, but it appears that the

annual per pupil spending o the Federal Bureau

o Prisons on educational programs roughly equals

the national per pupil average or secondary school,

or about $9,000 in 2007 dollars. A 240-hour literacy

or secondary school program would thus cost about

$2,000 per prisoner.1 About one ourth o state pris-oners lacks a high school diploma or GED and is

not currently involved in any school program. In

line with a national standard or state prisoners, 240

hours o basic education or these three hundred

orty thousand inmates would cost $680 million.

For inmates not enrolled in education programs,

correctional acilities could oer work in prison

industries making products used or state and lo-

cal governments (e.g., oce urniture). Again, the

ederal prison system provides a model. Federalprison industries employ prisoners in clerical, sales,

and manual semiskilled occupations. The program

is entirely sel-unding and employs about 17 per-

cent o ederal inmates, about twice the percent-

1. An average secondary school year is about 180 days o about 6.5 hours each, so 240 hours o correctional education is equal to about 20percent o a school year.


18/38




age o state prisoners in work programs. Earnings

vary rom $0.25 to $1.15 an hour, and are applied

to unpaid nes, victim restitution, and child sup-

port. The PREP evaluation, described above, ound

that six months work in ederal prison industries

was associated with a 24 percent reduction in re-incarceration rates at least eight years ater release

(Saylor and Gaes 1997). Deciencies in research

design probably account or some o this estimated

eect. Operating prison industries on a large scale

will likely reduce the benets we see or ederal

prisoners. Still, the PREP evaluation suggests that

modest reductions in recidivism can be obtained at

modest cost.

Although I propose mandatory participation in

education programs and an expansion o prisonindustries, participation in work and school pro-

grams could be handled in several ways. First, work

or school could be required o all prisoners in need,

just as the ederal system requires education o all

prisoners, and just as some policy analysts have ar-

gued or work programs or all prisoners. In addi-

tion to building the skills o prisoners, mandatory

work and education programs could be seen as a

way to demand more accountability rom prisoners

(Travis 2005). Second, programs could be volun-

tary and linked to reductions in time served. Goodtime is already provided or successul program

participation in many jurisdictions. Adopting this

approach nationally would allow expanded educa-

tion and work programs to reduce prison popula-

tions through early release.

In addition to work and education programs in pris-

on, I also propose national standards or discharge

planning that readies prisoners or release rom in-

carceration. Released prisoners are at highest risk

o recidivism immediately ater release rom prison(Langan and Levin 2002; Visher and Kachnowski

2007). In many jurisdictions, prisoners are released

with a little gate money and no real direction until

their rst parole appointment, which is sometimes

weeks ater reentry into society. During this period,

prisoners are at high risk o rearrest or drug relapse,

particularly i they have uncleared legal obligations

and lack work or housing. Discharge planning helps

released prisoners move quickly into employment,

housing, and substance-abuse treatment. National

guidelines or discharge planning would recom-

mend that departments o correction prepare or a

prisoners release by resolving uncleared warrants,nes, and child support obligations, and providing

a state-issued identication card.

Prisoners would also receive a risk and needs assess-

ment to provide reerrals or employment, hous-

ing, and treatment to ensure the transition to a sup-

portive social context in the rst days out o prison.

The needs assessment would take account o skills,

schooling, employment history, and employment

opportunities o parolees, as well as the many risk

actors associated with recidivism. Discharge plan-ning would regulate entry into transitional jobs and

treatment, ensuring that only those with real needs

would receive reerrals.

We currently have little systematic inormation

about the eectiveness o discharge planning, al-

though it is widely recommended by policy analysts

(Petersilia 2003; Travis 2005). A recent randomized

experiment evaluating the New York prerelease

program, Project Greenlight, ound no reductions

in recidivism or prisoners receiving discharge plan-ning (Wilson and Davis 2006). More evaluations o

discharge planning are needed. National standards

or discharge planning cost relatively little and rep-

resent a modest but realistic step toward enlisting

prisons in a more active role in ex-prisoners reen-

try into ree society.

taa ep, H, adsac-A ta

Sobriety and the habits o regular work oer thebest chance o improving employment among re-

leased prisoners. The path to a steady job will be

prepared by a bundle o intensive transitional ser-

vices (employment, housing, and substance-abuse

treatment), weighted to support the rst months

back in ree society.


19/38



DECEMBER 2008 17

The national program or transitional employment

assigns prisoners to a post-release job as part o their

discharge plan i they have no guaranteed employ-

ment prior to release. Those assigned to the pro-

gram would be required to report or work within

a week o prison release, as a condition o parole.Transitional employment would last up to twelve

months, although job placement services would

aim to quickly move ex-prisoners into the open

labor market. Employment would consist o ull-

time minimum-wage work in a small crew under

the direction o a supervisor. Program participants

would work in community service jobs maintaining

parks, roads, or public buildings and grounds. The

states would develop these programs to best t local

conditions. Transitional jobs might be directly or-

ganized by public agencies, or put under contract tononprot organizations. Administered in this way,

the transitional employment program resembles

the National Supported Work (NSW) Demonstra-

tion o the 1970s, or contemporary welare-to-work

initiatives.

Figures on the parole population and employment

rates among prisoners help us estimate the scale

o the transitional jobs program. At current levels,

about 70 percent o annually released prisoners

our hundred ninety thousand ex-inmatesareon some kind o supervised release and would be

eligible or transitional employment. (Those re-

leased without supervision cannot be mandated to

programs, and would all outside the scope o the

initiative.) A third o all prisoners were unemployed

at prison admission; data on released prisoners in-

dicate one hal to three quarters are out o work in

their rst months ater release (Sabol 2007; Visher

and Kachnowski 2007). I one hal o all parolees

need work immediately ater release rom custody,

the transitional employment program would needto supply two hundred orty-ve thousand jobs an-

nually.

Supplementing the employment program, states

would also provide transitional housing or home-

less ex-prisoners. There are no national statistics on

homelessness among ex-prisoners, but gures rom

major jurisdictions suggest 10 to 20 percent o pa-

rolees are homeless or some period in the two years

ater release (Metraux and Culhane 2004). Around

one hundred thousand additional beds would be

needed to guarantee housing to homeless parolees.

To promote sobriety and the habits o regular work,homeless parolees would be assigned to support-

ive housing that combines accommodations with

substance-abuse treatment and other counseling

services. Such supportive housing, like transitional

employment, would be provided or up to a year.

Because homeless ex-prisoners are likely to have

the most acute needs, supportive housing oers a

promising path to stable and independent housing.

Supportive housing or homeless parolees will also

provide the social benet o reducing the numbers

o ex-prisoners in city shelters or illegally residingin public housing.

Finally, to support employment I propose expand-

ing resources or substance-abuse treatment or

parolees. Around two-thirds o state and ederal

prisoners reported a history o heavy drug or al-

cohol use prior to incarceration. In many jurisdic-

tions, parolees with substance-abuse problems are

mandated to attend treatment programs. We lack

national gures or parolees, although 40 percent

o probationers attend drug or alcohol treatmentas a condition o their supervision. These gures

suggest that around 30 percent o paroleesabout

two hundred thousandmay need treatment and

are currently without a treatment mandate. I hal

o these prisoners were already in some kind o

substance-abuse program, then about one hundred

thousand parolees would require additional treat-

ment.

A national program or transitional employment,

housing, and drug or alcohol treatment would rep-resent a signicant commitment to the economic

and social reintegration o ex-prisoners. The gross

cost o transitional employment depends on how

many parolees will move to unsubsidized work

within a year. Post-release surveys suggest hal o

all program participants obtain employment a-

ter six months. Thereore, two hundred orty-ve


20/38




thousand parolees would enter the program but the

cost would be based on one hundred eighty-our

thousand annual equivalent participants. Using the

NSW Demonstration as a guide to gross costs, each

participant would be paid $14,300 in annual wages

(in 2007 dollars), and the service provider wouldreceive $15,400 in overhead to cover the costs o

supervision and administration (Bartik 2001, p.

194). These costs are similar to those o the New

York CEO and ComALERT programs reviewed

above. The relatively high New York minimum

wage on which CEO and ComALERT costs are

based suggests average national costs may be lower

in practice. Supportive housing would annually cost

about $10,000 or each o one hundred thousand

beds. One year o substance-abuse treatment in

a nonresidential program costs about $4,900 oreach o one hundred thousand parolees (Substance

Abuse and Mental Health Services Administration

[SAMHSA] 2003). In total, the gross annual cost o

transitional programs would be about $7 billion.

Outside a correctional setting, this type o program

might subsidize those who could be sel-sucient

in the open labor and housing markets. As part o

a reentry program, however, released prisoners

would not voluntarily enroll. Instead, transitional

jobs, housing, and treatment would be assigned byprison and parole authorities using a needs assess-

ment at discharge. Participation in the programs

would not be voluntary at the discretion o ex-pris-

oners, but mandatory as a condition o supervised

release. There is oten a tension in the allocation o

program services between helping those who are

likely to do well and helping those who are most in

need. Assigning transitional services on the basis o

a ormal risk and needs assessment at discharge will

tend to channel services to parolees who are more

needy and at higher risk.

Pa r

The eectiveness o services or released prisoners

is reduced by the harshness o criminal punishment.

Imprisonment reduces employment and disrupts

the amily relationships that might otherwise sup-

port desistance rom crime. Although mass incar-

ceration prevents crime in the short run by incapac-

itating criminals, it undermines public saety in the

long run by expanding the population o ex-pris-

oners with ew economic prospects or amily sup-

ports. Correctional budgets also divert resourcesrom public saety investments in police or social

services. An eective plan or prisoner reentry that

builds a sustainable public saety must also reduce

the heavy reliance on imprisonment as the main in-

strument o criminal punishment. To develop this

sustainable public saety, the conditions o parole

supervision and revocation must become less puni-

tive. In particular, or transitional services to sup-

port reentry parole agencies must signicantly limit

reimprisonment or technical violations.

What are technical violations? Parolees are general-

ly required to remain drug ree, gainully employed,

and diligent in reporting to treatment and their pa-

role ocers. In addition to imprisonment or new

crimes, parolees can also be incarcerated or violat-

ing these so-called technical conditions o release.

Failing a drug test, losing a job, or missing appoint-

ments can all trigger reimprisonment or technical

violations. Recommitment o parole violators has

been a signicant driver o state imprisonment rates

through the 1990s (Blumstein and Beck 2005). Bythe early 2000s, parole violators accounted or a

third o state prison admissions. About hal o pa-

role violators were drug oenders (Blumstein and

Beck, p. 63). Unlike a conviction or a new crime,

parole revocation or technical violations is oten

more an administrative than a law enorcement de-

cision. Revocation decisions are oten guided by the

exigencies o parole caseloads and prison capacity

(Jacobson 2005). The role o managerial actors in

revocation decisions is refected in large dierences

in revocation rates across states. Some states, likeFlorida and Illinois, reimprison relatively ew tech-

nical violators, whereas Caliornia revokes nearly 60

percent o parolees or technical violations. Revok-

ing technical violators in response to administra-

tive pressures can result in overincarceration, where

the prison detains those who pose little danger to

the community. This recommendation or parole


21/38



DECEMBER 2008 19

reorm ollows a number o similar proposals by Ja-

cobson (2005), Travis (2005), and Petersilia (2003)

to reduce the recommitment o parolees to prison.

In these proposals, and mine, parolees committing

new crimes should o course be prosecuted in the

courts and sentenced to prison i necessary.

While policy experts proposed limits on parole re-

vocation mostly to control prison populations, cur-

tailing reimprisonment for technical violators promotes

public safety by enhancing the effectiveness of transitional

services. I we view transitional programs as build-

ing the lie skills or successul reintegration, we

should expect ailurerelapse into drug use, job

loss, missed parole appointmentsto be a common

part o the process o reentry. Relapse is part o a

learning process in which new noncognitive skills oreliability and persistence are acquired. I ailure is a

likely stop on the path to steady work, parole super-

vision must also tolerate drug relapse or unemploy-

ment without automatic return to prison. Sending

parolees back to prison or ailing drug tests or

other technical violations truncates the acquisition

o prosocial behaviors that transitional services are

designed to oster. I drug relapse and other kinds

o ailure are common but are ultimately ollowed

by steady employment and other positive behavior,

a reintegrative system o parole release should al-low or ailure within a context o community-based

sanctions. Keeping parolees in the community will

allow them more access to transitional services and

greater chances or success.

Restricting parole revocation will increase the dose

o transitional services, but unchecked technical

violations oten indicate problem behaviors that

lead to crime and other serious ailures. To avoid

this path, technical violators should ace a range o

graduated sanctions designed to control problembehaviors and maintain participation in transitional

services. Instead o reimprisoning technical viola-

tors, graduated sanctions apply more-intensive pa-

role supervision or more-intensive programming

or those who ail to comply with a treatment plan.

Day reporting centers, or example, can require

technical violators to sign in or substance abuse

and other treatment, and or community service.

Attendance at day reporting centers or up to seven

days a week intensies supervision in a way that also

intensies programming. At a higher level o super-

vision, residential acilities strictly monitor parolees

while allowing their participation in community-based programs. Stricter supervision in these ways

oers greater access to services, not just greater ex-

posure to the detection o violations. Persistent vi-

olators would ace disciplinary hearings combined

with short jail stays, up to several weeks. A system

o graduated sanctions oers line ocers a wider

array o responses to parole violations than revoca-

tion and reimprisonment alone. Incarceration re-

mains available, but more in the orm o short jail

stays than extended periods o imprisonment. This

approach reduces the overincarceration o thoseposing little risk to public saety while increasing

parolees use o community programs.

To promote the integrated development o parole

and transitional services, states access to transi-

tional employment unds will be conditional on re-

stricting reimprisonment or technical violations.

Replacing reimprisonment or technical violators

with a fexible range o graduated sanctions rede-

signs parole supervision to work smoothly with

a large-scale transitional employment program.Linking transitional services to restrictions on re-

commitment or technical violators helps shit pub-

lic costs rom custody to services and supervision

in the community. The transitional service package

provides a unied approach to promoting the dual

goals o post-prison employment and a reduction

in prison populations. Though punishment is re-

duced, the prospects or a sustainable public saety

are improved.

Though parole reorm will enhance the eective-ness o transitional programs, restrictions on parole

revocation may increase crime among some parol-

ees. Some technical violators who would otherwise

be reimprisoned would be let in the community

threatening public saety. Three pieces o evidence

suggest that restricting reimprisonment o techni-

cal parole violators in combination with graduated


22/38




sanctions and transitional services poses a small risk

to public saety. First, data rom the 1994 Bureau

o Justice Statistics (2002) show that the criminal

involvement o parole violators, counted by the

number o arrests, is no greater than the criminal-

ity o those who successully complete parole, andsubstantially less than the criminality o parolees

who are recommitted or new crimes. Certainly,

technical violations by themselves oten all short

o new imprisonable oenses. Drug use without

any aggravating circumstances, or example, does

not rise to the level o a misdemeanor and attracts

only a citation in many jurisdictions. Second, the

Bureau o Justice Statistics recidivism study also

shows that recently released prisoners account or

a small raction, about 5 or 6 percent, o all arrests

(Langan and Levin 2002). Reducing parole revoca-tion rates would reduce the overall level o public

saety by a very small raction. Restricting parole

revocation or technical violators would increase

the pool o recently released prisoners by about 20

percent, which would add 1 percent to the arrest

rate, assuming no reduction in recidivism due ei-

ther to programs or reormed parole supervision.

However, programs and parole reorm are likely

to reduce arrest rates, and the contribution o pa-

role revocation to overall arrest rates is likely to be

smaller.

Finally, several states have recently experimented

with community-based sanctions and reductions

in parole revocation with no great adverse impact

on crime. Analysis o Oregon parolees ound 20 to

75 percent less reoending among those receiving

community-based sanctions than those receiving

incarceration, controlling or scores on a risk-as-

sessment instrument (Oregon Department o Cor-

rections 2002). A preliminary evaluation o a Geor-

gia program suggests parolees who are allocated tograduated sanctions with a risk-assessment instru-

ment are no more likely to be rearrested than is a

control group (Meredith and Prevost 2008). More

inormally, large cuts in parole revocation rates in

Kansas (50 percent rom 200406) and New Jersey

(32 percent rom 200107) coincided with signi-

cant declines in index crime rates. In short, several

states have begun to adopt the kinds o parole re-

orms suggested here without negative eects on

crime, and evaluations suggest technical violators

who are sanctioned in the community or receiving

very short jail stays do better than those who are

reimprisoned.

How much will crime increase by restricting re-

imprisonment or technical violators? There were

about six hundred thousand arrests or violent

crimes in 2007. Recidivism statistics suggest that

about 6 percent o those arrested, or thirty-six thou-

sand, were recently released prisoners (Roseneld,

Wallman, and Fornango 2005, p. 87). Parolees have

a relatively low rearrest rate compared to unsuper-

vised releasees (Roseneld et al., p. 93), suggest-

ing that about 40 percent, or teen thousand, othose arrested or violence were on parole. Leaving

technical violators in the community increases the

parole population by about 30 percent (Glaze and

Bonczar 2007, p. 7). This implies that parole reorm

would increase violent arrests by two thousand ve

hundred each year, assuming higher arrest rates

among those committing new crimes and assuming

transitional services and graduated sanctions had no

crime-reducing eect. However, we expect the new

programs and the graduated sanctions together

will reduce crime by about 25 percent, a conserva-tive summation o the eects o prison education,

transitional jobs, and community-based sanctions.

Parole reorm under the reentry program thus will

add about one thousand seven hundred arrests or

violence each year. Calculations below will weigh

this eect against the gross reduction in recidivism

produced by the national reentry program. As we

will see, the gains to public saety substantially out-

weigh the costs.

Caa Cqc

To provide a supportive context or reentry and re-

integration, I also propose the elimination o bans

on ederal benets or people with criminal records.

Some classes o elony oenders can be denied Tem-

porary Assistance or Needy Families (TANF), ood

stamps (now the Supplemental Nutrition Assistance


23/38



DECEMBER 2008 21

Program, SNAP), postsecondary educational assis-

tance, and housing benets. In 1996, ederal welare

reorm created a lietime ban on TANF and ood

stamps or elony drug oenders, although states

could narrow the ban or opt out. Eighteen states,

including Caliornia and Texas, retain the ban onex-elons or drug-related crimes. Another twenty-

two states operate a modied ban, typically exempt-

ing ex-elons who are in substance-abuse treatment.

The Government Accountability Oce estimates

that about 15 percent o released drug oend-

ers would otherwise be eligible or TANF or ood

stamps (Government Accountability Oce [GAO]

2005). This would include at least orty thousand

parolees released in 2006, and a larger group o ex-

prisoners who are no longer under supervision.

Federal postsecondary educational benets are also

denied to prisoners and drug oenders. Prisoners

have been ineligible or Pell Grants since 1994,

greatly reducing the number o prison college

programs. Felony and misdemeanor drug convic-

tions also disqualiy students rom Pell Grants and

student loans. It is dicult to estimate how many

would apply or Pell Grants i the restriction on

drug oenders were to be removed. From 2001 to

2003, one hundred orty thousand applicants unsuc-

cessully applied or benets ater reporting a drugconviction or reusing to answer a question related

to drug convictions (GAO 2005). This gure does

not count those with drug convictions who did not

apply because they assumed they were ineligible.

Federally assisted housing benets are also restricted

or those with criminal records or drugs or violence.

Public housing tenants evicted or drug-related ac-

tivity also are given a mandatory ban or three years.

Public housing agencies are given wide discretion

to screen and evict tenants who have engaged indrug-related or violent criminal activity. Around 80

percent o public housing agencies surveyed by the

Department o Housing and Urban Development

(HUD) in 1999 reported that they conducted some

kind o background check, whether by sel-reports

rom applicants or, more commonly, by searching

a criminal record database (Devine, Haley, Rubin,

and Gray 2000). In 2002 and 2003, three thousand

one hundred public housing agencies surveyed by

HUD reported denying orty-nine thousand ap-

plications each year or criminal activity, which is

about 4.2 percent o all applications (GAO 2005,

p. 67). These gures likely understate the impacto rules against ex-prisoners because those coming

out o prison are less likely to apply or ederally

assisted housing.

Policymakers oer a punitive motivation or the

bans on welare and educational assistance, ex-

tending punishment by withholding help rom the

undeserving. This motivation should be balanced

by considerations o public saety and reintegra-

tion. People coming out o prison will be better

equipped to resume normal lie i they have accessto social supports. All bans on educational and wel-

are benets should be dropped, but there might be

a stronger argument or excluding some ex-prison-

ers rom ederal housing benets on the grounds o

public saety. Although current rules or excluding

drug oenders may be too wide, unless we under-

stand the public saety risk posed by ex-prisoners in

public housing we have no sound basis or policy.

The eects o restrictions on ex-prisoners on rates

o crime in ederally supported housing should be

evaluated beore a policy decision is made.

bad h Pa

The national prisoner reentry program ocuses on

one specic strategy or improving employment

among people released rom prison: transitional

jobs in the context o supportive programs and su-

pervision. A national reentry eort could be broad-

ened to advance the main goal o steady employ-

ment or ex-prisoners. Supplementary measures

might include community-based education andtraining. Therapeutic measures such as motivation-

al interviewing or a cognitive behavioral program

designed to develop impulse control might support

the object o developing the noncognitive skills o

reliability, motivation, and sociability.

Finally, the program might also target employers.


24/38




Training bonuses or employers and additional or

automatic bonding o paroled workers might in-

crease employers incentives or hiring workers

with criminal records. To prevent job applicants

rom being screened out, states might also relax hir-

ing restrictions on workers with criminal records.Opening employment to job seekers with criminal

records in the health-care industry, or example, or

restricting criminal background inormation may

reduce the stigma o incarceration or those leaving

subsidized jobs or the open labor market. Although

we have little concrete evidence rom existing eval-

uation research, supplementary measures such as

these may improve the eects o transitional jobs.

These measures would also broaden the test-bed

or program evaluation.

A Pah a naa Pa

While the national prisoner reentry program out-

lined here suggests the scale o the eort needed

to improve employment among released prisoners,

wholesale reductions in state prison populations

and large increases in post-release services can-

not be adopted overnight. Detailed eatures o the

programs are unspecied. States ace a wide variety

o challenges in adapting programs to local condi-

tions. A easible path or moving to a national plan will require incremental change in which policy

knowledge about implementation is accumulated

and disseminated.

The rst step to adopting a national prisoner re-

entry program will thus involve establishing a rela-

tively small number o demonstration states. These

states would adopt the three key elements o the

program: (1) transitional jobs and other services,

(2) parole reorm, and (3) expanded correctional

programs. The demonstration states would be se-lected through a competition in which applicants

would detail the programs and then demonstrate

their easibility.

Evaluation will be central to the demonstration.

States must build rigorous plans or data collec-

tion and analysis to evaluate the programs. The

evaluations will gauge the programs success at re-

ducing recidivism, and at increasing employment

and earnings. Standard reporting requirements

will help ensure that the program evaluations will

contribute to a cumulating body o results that can

easily be interpreted across jurisdictions. Althoughmany studies have evaluated programs or released

prisoners, only a ew are based on randomization

or similarly strong designs. The implementation o

the program thus provides a signicant opportu-

nity or dramatically expanding policy knowledge

through randomized evaluations. Prior evaluations

o reentry programs have relied heavily on admin-

istrative data to measure employment and earnings.

These data are likely quite inaccurate or those with

criminal convictions. Data collection should thus

draw widely rom dierent sources, including romsurveys o the parolees themselves. Reporting on

the evaluation should be prompt, and reports wide-

ly disseminated. In this way, policy learning will be

built into the implementation.

Implementation across states will proceed incre-

mentally. Several demonstration states will be und-

ed initially, and more states will be brought online

over time in successive competitions or ederal re-

entry unds. Program implementation will generate

a growing body o evaluation results, and this newknowledge should be refected in successive appli-

cations. The program will thus grow across juris-

dictions, producing a common and increasing und

o policy knowledge.


25/38



DECEMBER 2008 2

Table 3 summarizes the key components o the

national prisoner reentry program. These g-ures indicate that the total gross cost o the

program will equal about $8.5 billion dollars, with

transitional employment accounting or more than

hal. The unding scheme is guided by two consid-

erations. First, the reentry program is partly con-

ceived to reduce prison populations and shit cor-

rectional costs rom custody (which is expensive) to

community supervision and programming (which

is inexpensive). Second, there is great variability be-

tween the states criminal justice and social service

agencies, so states must have fexibility to apply re-entry unds to their own local conditions.

4. C ad bf

In the national prisoner reentry program, ederal

reentry unds are distributed to states that adoptnational standards or discharge planning, inten-

sive reentry programming, and parole supervision.

These measures will reduce incarceration rates

by reducing recidivism and parole revocations or

technical violations. To apply or ederal reentry

unds, states would develop local plans or transi-

tional services, parole supervision, and prison pro-

grams. States may opt out o the guidelines and still

obtain unds i they can otherwise link the expan-

sion o post-release services to a reduction in prison

populations. Depending on local conditions, states would then distribute unds to local authorities,

Annal Annal Annal ro cot

cot per participant participant (million $)

Decription

Tranitional ervice ater prionEmployment, p to one year $14,00/$15,400a 184,000 $5,464Hoin, p to one year $10,000 100,000 $1,000

Dr treatment $4,900 100,000 $490In-prion prorammin

240 hor o baic edcation $2,000 40,000 $680Wor in manal and clerical job n/a n/a

Elimination o collateral coneqenceRetore TANF eliibility $4,200 100,000 $420Retore Pell grant $2,800 140,000 $92

Total $8,446

sorce: Athor.

Note: Prion wor proram are amed to be el-ndin; cot will be incrred in ettin p the proram. TANF fre are calclated amin an averae monthly

beneft o $50. Pell grant are amed to provide an averae beneft o $2,800. All fre are in 2007 dollar.

a. Wae/overhead.

n/a = Not applicable.

tAble 3.

Dcp ad C h naa P r Pa


26/38




departments o correction, workorce development

agencies, and so on.

Are the benets o these measures greater than the

costs? The gross costs o the program are oset in

our main ways (Table 4). First, the employmentprogram provides benets in the orm o improve-

ments in inrastructure, cleaner parks, streets, pub-

lic grounds, and so on. The NSW Demonstration

valued the output o similar subsidized employ-

ment at about 45 percent o program costs (Kem-

per, Long, and Thornton 1984). I the output o

parolees is valued at 45 percent o its cost, the net

cost o transitional employment is reduced by about

$2.5 billion.

Second, the program also has large individual and

social benets or the people released rom prison,

their amilies, and their communities. By helping

ex-prisoners develop social and job skills, the pro-

gram will make them more employable even ater

the one-year employment placement has ended,increasing their uture earnings. To estimate that

aggregate post-program eect, we assume that

the transitional employment raises earnings by 15

percenta conservative estimate compared to the

ComALERT evaluation aboveand value untreat-

ed earnings at about $9,000 annually, approximately

the level o earnings o ex-prisoners in survey data.2

Under these assumptions, the annual aggregate

benet o the program is around $250 million each

tAble 4.

sca bf h naa P r Pa

Pv pac Annal beneft(million $)

Quantifable Benefts

Prodctivity rom tranitional employment 2,460

Increaed earnin rom the reentry proram (annally) 250[increaed pport to children o releaed prioner] [140]

Redction in crime amon proram participant 2,500Redction in correctional cot . . .

de to redction in crime 1,510

de to redction in parole revocation 4,050Total 10,770

Hard-to-quantiy benefts

Lietime increae in earnin

Lietime redction in crimeImproved child well-bein de to increaed earnin and redced crime

sorce: Athor.Note: Increaed pport to children in bracet i not conted amon the total beneft becae it i already conted in the annal increaed earnin. All fre are in

2007 dollar.

2. ComALERT participants earned 37 percent more than a matched comparison group with earnings measured by unemployment insurancerecords, and 29 percent more than a matched comparison group with sel-reported earnings.


27/38



DECEMBER 2008 25

year. The program eects may well decay, but even

over a three-year period total eect may exceed hal

a billion dollars.

Importantly, a portion o these wages will fow to

amilies o ex-prisoners. About 80 percent o stateprisoners are athers, so ertility estimates suggest

around two hundred ty thousand children would

obtain some economic benet rom these subsi-

dized wages. Some children may benet because

wages rom transitional jobs will make ex-prisoners

more-attractive partners or the mothers o their

children. Additional wages may thus contribute

to higher rates o coresidence and greater nan-

cial contributions o athers to children (e.g., Blau,

Kahn, and Waldogel 2000; Homan and Duncan

1995). In other cases, the wages o transitional jobsmay contribute to athers child support payments.

Beston (2006) estimates that 25 percent o household

spending is spent on the child in single-child ami-

lies, and recent analyses (Geller, Western, and Gar-

nkel 2008) suggest that ormerly incarcerated par-

ents contribute about $2,000 annually to each child.

I a ather in transitional employment contributes a

quarter o his earnings o $10,200 to his child, this

will yield an increase o $550 annually over his un-

subsidized contribution, passing on in the aggregate

around $140 million each year to poor children. Inaddition to increased earnings, ex-prisoners would

become more supportive spouses and parents due to

improved literacy and sobriety. These amily eects

are hard to quantiy but should be counted among

the social benets o the reentry program.

Third, the national reentry program will reduce

crime. The sequence o interventions proposed

here, including prison education, transitional ser-

vices, and parole reorm have not been evaluated

in combination. Evaluation studies show that tran-sitional jobs by themselves reduce recidivism by

20 percent (Bloom et al. 2007; Jacobs and Western

2007; or prisoners over age twenty-six see Uggen

2000). I we consider the eects o expanded prison

education and program eects under a system o

graduated parole sanctions, the national prisoner

reentry program could reasonably reduce arrests

by 25 percent. Arrest rates among released prison-

ers suggest new parole cohorts account or our

hundred ty-ve thousand arrests each year, so

the national reentry program would produce a 25

percent reduction o one hundred teen thousand

arrests and a somewhat smaller number o prisonadmissions each year.

The reentry program yields a reduction in crime,

but parole reorm may increase crime by expanding

the number o released prisoners in the community.

Above, I estimated that the increased number o

recently released prisoners would increase violent

arrests by one thousand seven hundred, which im-

plies an increase in total arrests o thirty-our thou-

sand. (Arrests or violence are about 5 percent o

all arrests.) The net eect o the reentry programon arrests is thus one hundred teen thousand less

thirty-our thousand, a reduction o about eighty

thousand arrests. Levitts (1996) estimates suggest

one out o seven crimes results in an arrest, suggest-

ing the reentry program will avert about ve hun-

dred sixty thousand crimes annually, given current

levels o crime and parole supervision.

Estimates o the economic costs o crime vary widely,

though a common estimate accounts or pecuniary

loss, medical costs, lost working time, and pain andsuering. Considering these actors yields an aver-

age cost o crime o $4,500 in 2007 dollars (DiIulio

and Piehl 1991; Freeman 1996; Levitt 1996). At this

price, the social benet o reduced crime is about

$2.5 billion.

About a third o those rearrested return to prison

or a new oense, suggesting that the eighty thou-

sand arrests averted will eliminate correctional costs

or twenty-our thousand ex-prisoners. Given me-

dian time served o twenty-eight months, reducedcrime will yield an incarceration reduction o ty-

six thousand prison years at the cost o $27,000 a

year. The total savings is $1.51 billion.

Finally, the program links the expansion o services

to reductions in the prison population through the

elimination o imprisonment or parole violations.


28/38




Parole violations account or more than a third o

state prison admissions, or two hundred thirty-two

thousand out o six hundred seventy-ve thousand

in 2005 (Sabol, Minton, and Harrison 2007). About

two thirds (or one hundred ty thousand) o these

parolees are returned to prison or technical viola-tions (Glaze and Bonczar 2007). Figures or specic

states suggest parole recommitments add an addi-

tional ourteen months o incarceration (Blumstein

and Beck 2005). The annual cost o a prison bed is

about $27,000, so annually diverting one hundred

ty thousand parolees saves about $4 billion each

year in correctional costs. (O course, average costs

o incarceration are not equal to marginal costs, so

departments o correction would need to distribute

the reductions in incarceration to reduce correc-

tional budgets.)

A list o the programs social benets is reported in

Table 4. Program benets slightly exceed the costs.

The social benets may be larger than those report-

ed here because the combined eects o new pro-

grams on earnings and crime reduction may well be

larger than assumed. Under current assumptions,

program eects on recidivism and employment are

short-lived, but they may be persistent and produce

large lietime gains in earnings and reductions in

crime. Finally, the analysis takes no account o the

reentry programs likely positive eect on childrenswell-being and lie chances.

The costs o the national reentry program are in-

curred by the ederal government, but the benets

fow to states (in reduced correctional costs) and

individuals (through reductions in crime and in-

creased employment). When costs and benets are

separated by levels o government there is a danger

o crowding out, where states spend less in antici-

pation o ederal support. In this scenario, prison-

er reentry measures would come to resemble theederal welare program, which takes the orm o

block grants to states. Despite this institutional re-

semblance, the eect o crowding out is likely to be

very small, because so little state spending currently

goes to prisoner reentry programs.


29/38



DECEMBER 2008 27

Critics may object that a national prisoner re-

entry program will displace private sectoremployment and is likely politically inea-

sible. Alternative proposals emphasizing a dierent

mix o services may oer more promise.

W h aa p pa

dpac pva-c p?

Transitional employment programs displace work in

the public sector and may crowd out private sector

employment by raising wages (Ellwood and Welty

2000). These disemployment eects appear to be

largest when public service employment broadly re-cruits rom the labor orce and when program em-

ployment is used to counter recession. The national

prisoner reentry program is unlikely to negatively

aect employment. In this case, the transitional em-

ployment is highly targeted and is not broad-based.

The program would not treat parolees as a whole,

but would treat only those who have trouble nding

work. There are ew competitors or these workers

in the open labor market. The employment pro-

gram is also conceived as a standing eature o the

process o reentry, paying minimum wages throughexpansions as well as recessions. As a result, the la-

bor-displacing eects o the national prisoner reen-

try program are likely to be small.

i pca a xpad vc

x-p?

Although transitional employment, on balance, will

positively aect the economic opportunities o pa-

rolees, any measure that provides social services to

criminal oenders is politically vulnerable. Few re-

cipients would seem less deserving. Traditional re-habilitation programs were built on a philosophy o

remediation, mending criminal oenders or return

to society. In contrast, the national reentry program

has as its key objective sustained public saety. The

program aims to reduce crime by expanding eco-

nomic opportunity while developing the rudimen-

tary skills o motivation, reliability, and sociability.

5. ojc ad Aav

By keeping parolees out o prison and providing

benets that will fow to amily members, the pro-gram also strengthens poor, high-crime commu-

nities. Finally, by oering a path back rom mass

incarceration, the program also provides states with

an alternative to mounting correctional budgets.

Public saety, community investment, and scal

prudence all recommend the national reentry pro-

gram as an improved and politically viable model

or criminal punishment.

Wha a aav appach?

Vocational education, job readiness training, and job placement eature in other, less-expensive, re-

entry programs. The largest obstacles to employ-

ment among ex-prisoners are the human capital

decits that are refected in noncognitive as much

as in cognitive skills. Because the work histories o

ex-prisoners are generally so poor, development o

basic job skills such as maintaining a regular work

schedule, ollowing directions, and cooperating

with coworkers can be enormously dicult. Pro-

grams that emphasize improving vocational skills

or connecting job seekers with employers oten ailto address these undamental deciencies o non-

cognitive skills. As a result, transitional employ-

ment provides more promising results than either

job placement or vocational training. Because the

decits o ex-prisoners are so acute, and programs

in many cases must undo the eects o the prison

time, a larger dose is needed to produce reductions

in unemployment and recidivism.


30/38




The American penal system has grown contin-

uously or the past thirty-ve years. Spend-ing on corrections now totals $70 billion

each year. Among men born since the late 1960s,

30 percent o blacks without college education and

6 percent o whites without college education have

spent time in prison, over hal serving more than

two years or a elony conviction. Ater release,

ex-prisoners experience reduced rates o employ-

ment, wages, and wage growth, and elevated risks

o divorce and separation. Two-thirds are rearrested

within three years, and one-ourth return to prison

during that time.

The emergence o mass incarceration presents pol-

icymakers with two challenges. First is the social

challenge o averting the ormation o a large class

o outsiders who have little contact with mainstream

institutions and who are deeply and enduringly in-

volved in the criminal justice system. Second is the

scal challenge o reining in correctional budgets

that divert resources rom education and other so-

cial investments.

The large decline in crime rates through the

1990swidely shared by other countries that did

6. Cc

not double their incarceration ratesoers an op-

portunity to meet the social challenge o reentryand the scal challenge o mounting correctional

costs. My proposal or a national prisoner reentry

program aims to link the social reintegration o ex-

prisoners to a reduction in prison populations. In

this proposal, the choice or policymakers is not

whether or not to spend money on reentry pro-

grams, but rather whether to spend money on re-

Date post:	06-Apr-2018
Category:	Documents
Upload:	brianosclt
View:	217 times
Download:	0 times

12 Prison to Work Western

Documents