Research Paper:Policy Issue and Solution Assignment
farra004
CostofbadarticleResearchP.pdf
Articles
Criminal Justice Policy Review 21(4) 391 –434
© 2010 SAGE Publications Reprints and permission: http://www. sagepub.com/journalsPermissions.nav
DOI: 10.1177/0887403409352896 http://cjp.sagepub.com
Estimating the Costs of Bad Outcomes for At-Risk Youth and the Benefits of Early Childhood Interventions to Reduce Them
Mark A. Cohen,1,2 Alex R. Piquero,3
and Wesley G. Jennings4
Abstract
Although it appears that there is growing interest in early childhood intervention as an effort to reduce crime, resources continue to be funneled toward punishment and incarceration. Considering this and acknowledging earlier cost-based empirical research, the question still remains as to the cost incurred by a lifetime of involvement in crime and experiencing a host of adverse noncrime outcomes. This study provides a review of the literature in search of well-designed early childhood interventions that address a series of socials ills, such as crime and delinquency, educational attainment, drug and alcohol abuse, smoking, child abuse and neglect, poor health outcomes, and teen pregnancy. Furthermore, building on the earlier framework and basic methodology developed by Cohen and recently updated by Cohen and Piquero, this study offers calculations of the present value of lifetime costs imposed on society for each of these various social ills—discounted to the date of birth to put them on comparable terms. The largest cost is imposed by the career criminal (US$2.1-US$3.7 million). Next, the present value costs associated with both drug abuse and alcohol dependence/abuse are roughly the same—about US$700,000 each—whereas child abuse and neglect costs an estimated US$250,000 to US$285,000. Health-related outcomes range from a low US$10,300 for the estimated present value cost of low birth weight to US$127,000, US$144,000, US$187,000, and US$260,000 for coronary heart disease, asthma, diabetes, and smoking, respectively. Finally, the present value cost of teen pregnancy is estimated to range from US$120,000 to US$140,000. Thus, properly designed programs and policies that focus on early childhood intervention
1Vanderbilt University, Nashville, TN 2Resources for the Future, Washington, DC 3University of Maryland, College Park 4University of Louisville, Louisville, KY
392 Criminal Justice Policy Review 21(4)
have the potential to produce significant social benefits. Study limitations and suggestions for future research are also discussed.
Keywords
early childhood interventions, parent training, home visitation programs, costs and benefits, willingness-to-pay
Programs targeting at-risk families and their young children from prenatal stages to early childhood have as their primary goal the promotion of the child’s healthy devel- opment, including good health and nutrition, proper social and emotional development, and early cognitive gains. They are also focused on the mother–child relationship, with the intent of improving those interactions and reducing child abuse. Longitudinal studies have identified possible longer-term benefits of such early childhood interven- tions that likely result from these initial improvements, including increased educational attainment later in life, reduced delinquency and criminal offending, reduced alcohol and drug abuse, and reduced teen pregnancy (for both the child in later years as well as the young mother herself, through reduced repeat pregnancies; Olds et al., 1998; Piquero, Farrington, Welsh, Tremblay, & Jennings, 2009).
Even if shown to be successful in reducing one or more social ill(s), however, a key policy question is whether the cost to society from that intervention program exceeds its benefits. This question has been difficult to answer because although the costs of intervention programs are often available, the benefits are more elusive (Greenwood, 2006). One approach is estimating the “costs of bad outcomes”—that is, what society likely will pay for a set of social ills experienced by at-risk youth throughout their lives. Such an estimate can then be linked to interventions associated with preventing or reducing them. Together, these costs and benefits provide an estimate of what soci- ety should pay to prevent the social ills and how much net benefit it might gain from such attempts.
In this same vein, Cohen’s (1998) seminal work on estimating the monetary costs of saving a high-risk youth has received considerable attention by at-risk youth pro- grams and foundations, and it has been used in particular to promote funding for interventions to help at-risk youth. Cohen provided one of the initial and most com- prehensive attempts to calculate the costs of crime, and his research estimated that the typical career criminal caused US$1.3 to US$1.5 million in external costs with a heavy drug user and a high school dropout incurring approximately US$370,000 to US$970,000 and US$243,000 to US$388,000 in costs, respectively. After eliminating the duplication between crimes committed by individuals who are both heavy drug users and career criminals, the results suggested an overall estimate of the “monetary value of saving a high-risk youth” being between US$1.7 and US$2.3 million.1 Cohen and Piquero (2009) recently extended Cohen’s original estimates by using a more comprehensive cost methodology and new data on career offenders and estimated the present value of saving a high-risk youth (as of birth) to range between US$2.6 and US$4.4 million.
Cohen et al. 393
This study expands the coverage of Cohen’s (1998) and Cohen and Piquero’s (2009) analysis by incorporating other social ills that have been determined to be associated with high-risk youth behaviors—especially delinquency/crime (Loeber & Farrington, 1998) and for which early childhood interventions might have an impact. In particular, in addition to the educational, alcohol and drug abuse, and criminal/delinquent out- comes dealt with in Cohen’s (1998) and Cohen and Piquero’s (2009) research, we will also focus on (a) physical health outcomes, (b) teenage pregnancy, and (c) child abuse and neglect.
Review of Effective Early Childhood Interventions Although there is a considerable amount of variation across early childhood interven- tion programs that focus on reducing delinquency/crime and antisocial behavior, there is relative consistency in the literature suggesting that most of these programs can be classified into three broad categories: (a) home visitation programs, (b) parent-training programs, and (c) daycare/school-based programs. The general components of these programs along with selected examples of each are reviewed below.
Home Visitation Programs Home visitation programs are typically designed with the intention of affecting mater- nal, child, and family functioning; however, they may vary in whether they focus on parent education, promotion of maternal health-related behavior, cognitive stimulation of the child, providing social support to the families, or linking the families with appro- priate services (Olds & Kitzman, 1993, p. 56). As the name implies, home visitation programs typically involve some kind of component where paraprofessionals, such as those with academic credentials in nursing, education, or social work, visit the families and assist them with navigating through parenthood. These programs generally target the mothers/families during pregnancy or on the birth of their child and provide some form of prenatal, postnatal, or pre-and-postnatal services. In addition, these programs also vary in the type(s) of outcomes that they assess such as a reduction in preterm deliveries and low-birth-weight babies, improving health-related behaviors (e.g., reducing or eliminating maternal cigarette, alcohol, and/or drug use during/after preg- nancy), promoting obstetrical health (e.g., lowering rates of epidural anesthesia and kidney infections), and supporting psychosocial functioning and partner engagement/ involvement with the pregnancy. Other outcomes include educating and increasing access of the expecting mothers/families to routine pre/postnatal care in the hopes of reducing the prevalence and frequency of hospitalizations during pregnancy or the infant’s time spent in the neonatal intensive care unit (NICU; Olds & Kitzman, 1993).
Although most of the outcomes reviewed above focus on pregnancy/birth-related factors, home visitation programs are also well-known for their emphasis on promot- ing the infant/child’s cognitive development and physical health. Furthermore, home visitation programs often attempt to educate the young mothers on the importance of maternal caregiving and establishing a warm and stimulating home environment for
394 Criminal Justice Policy Review 21(4)
the child. Research has demonstrated that mothers who receive home visits exhibit better interaction with their child and have more realistic attitudes toward child rearing. Mothers who experienced home visits also generally report that their children have less difficult temperaments compared to mothers of children who do not receive home visits (Olds & Kitzman, 1993). Finally, it is important to highlight that some home visi- tation programs evaluate their success based on reducing child behavior problems including delinquency/crime and antisocial behavior, reducing the number of later health problems for the mothers and their children, preventing teenage pregnancy as well as preventing the occurrence of child abuse and neglect (Olds et al., 1997; Olds & Kitzman, 1993).
The nurse home visitation program developed and evaluated by Olds et al. (Olds, Henderson, Chamberlin, & Tatelbaum, 1986; Olds et al., 1997, 1998) is certainly the most recognized example of home visitation programs in this literature for a number of reasons, most important for its solid research design and long-term follow-up. Car- ried out in Elmira, New York between April 1978 and September 1980, Olds et al. recruited 500 consecutive women with no previous births to enroll in a home visitation program where they were randomly assigned to receive home visits during pregnancy and home visits through the child’s second birthday. Comparatively, the control group received standard prenatal care and routine wellness visits in a clinic. Subsequent results demonstrated that the mothers who received the pre-and-postnatal visits had fewer instances of child abuse and neglect and their babies had less frequent visits to the emergency room and trips to the doctor for accidents and poisonings (Olds et al., 1986). Furthermore, Olds et al. (1997) revealed that the reduction in the incidence of child abuse and neglect persisted throughout the child’s life course (ages 4 through 15) and home-visited mothers also had less subsequent births and fewer self-reported and official arrests. The latest follow-up study of these Elmira youth suggested that those youths whose mothers received pre-and-postnatal visits reported fewer instances of running away, fewer arrests, fewer convictions and violations of probation, fewer life- time sex partners, fewer cigarettes smoked per day, and fewer days having consumed alcohol. Since this early pioneering study, Olds and his colleagues have replicated the majority of these findings insofar as the benefits of home visitation programs on child outcomes in Memphis, Tennessee (Kitzman et al., 1997, 2000; Olds et al., 2007) and Denver, Colorado (Olds, Robinson et al., 2004), although the follow-up has not been as long in these latter studies considering their recent implementation. In addition, other noteworthy home visitation programs such as the Busselton Project (Cullen & Cullen, 1996), the Comprehensive Child Development Program (St. Pierre & Layzer, 1999; St. Pierre, Layzer, Goodson, & Bernstein, 1997), and the Infant Health and Development Program (McCormick et al., 2006) also revealed differences between the home-visited children and the controls, with the intervention youth being more likely to have obtained a university degree (Cullen & Cullen, 1996) and being less likely to have been arrested as a juvenile (McCormick et al., 2006). In contrast, whereas the follow-up of the Comprehensive Child Development Program had not reached into adolescence or adulthood for the intervention children, the long-term
Cohen et al. 395
findings related to maternal outcomes (education, alcohol/drug use, etc.) suggested little to no program impact on these outcomes (St. Pierre et al., 1997; St. Pierre & Layzer, 1999).
Early Parent-Training Programs While home visitation programs could certainly be classified under the category of parent-training programs in a broad sense, there are some notable differences between the two types of programs. In general, early parent-training programs target parents with young children and offer parent training to strengthen the parent’s competencies in monitoring and appropriately disciplining their child’s behavior. More specifically, these types of programs usually also emphasize the importance of promoting the child’s social and emotional competence and its effect on reducing the occurrence/ discontinuation of the child’s behavior problems. These interventions are also typically administered by trained experts either directly or indirectly (e.g., through videotapes), and the sessions are conducted in a variety of settings including the home, the school, and/or the clinic in an individual or group-based form (Piquero et al., 2009).
The most common examples of these types of early parent-training programs are the Incredible Years Program (Webster-Stratton, 1982, 1984, 1990, 1992, 1998), the Triple P-Positive Parenting Program (Leung, Sanders, Leung, Mak, & Lau, 2003; Markie-Dadds & Sanders, 2006; Morawska & Sanders, 2006; Sanders, Markie-Dadds, Tully, & Bor, 2000; Sanders, Montgomery, & Brechtman-Toussaint, 2000), and Parent– Child Interaction Therapy (Brestan, Eyberg, Boggs, & Algina, 1997; Eyberg, Boggs, & Algina, 1995; Eyberg & Durning, 1994; McNeil, Eyberg, Eisenstadt, Newcomb, & Funderburk, 1991; Schuhmann, Foote, Eyberg, Boggs, & Algina, 1998; Zangwill, 1983). Although these programs are similar in their goals (e.g., reduce child behavior problems), there are some differences worth mentioning. For instance, the Incredible Years Program aims to provide comprehensive treatment regimes for young children with early onset conduct problems as well as the promotion of universal prevention programs that can serve as a mechanism to prevent the occurrence of conduct prob- lems among children in general (www.incredibleyears.com).
Comparatively, the Triple P-Positive Parenting Program offers a multilevel pre- vention program that trains the parents how to use positive and nonviolent techniques in response to their child’s behavior problems. The program is generally adminis- tered in various levels that begin with universal parenting information followed by one-on-one sessions with a health care provider. There are also additional levels of the program that may apply in situations with children who are demonstrating serious behavior problems, and these versions involve more intense and frequent one-on-one sessions with the health care provider (Sanders et al., 2004). Finally, the Parent–Child Interaction Therapy program focuses on fostering and developing a caring and respon- sive relationship between the parent and their child to effectively manage and reduce their child’s behavior problems. The program usually is administered by trained ther- apists who use instructions, modeling, and role-playing techniques to improve the
396 Criminal Justice Policy Review 21(4)
quality of the parent–child interactions and train the parent how to appropriately rec- ognize and reward positive child behavior and punish noncompliant behaviors (Eyberg et al., 1995).2
Although the majority of the studies that have evaluated the effectiveness of these types of early parent-training programs have demonstrated success in reducing child- hood behavior problems (Piquero et al., 2009), the outcome measures in virtually all of these studies are often limited to psychometric scales. Thus, although this feature makes it possible to determine the significance of the intervention and calculate an effect size for the intervention, it makes it difficult (if not impossible) to estimate how the changes translate into monetary benefits. Furthermore, despite the fact that a number of these studies are well designed methodologically, the majority of these programs and their evaluations do not provide long-term data and are thus limited to short-term evaluations that merely provide pre-post assessments with a considerable absence of long-term follow-up (particularly into the adolescent years of the child’s life). One notable exception is the study by Long, Forehand, Wierson, and Morgan (1994), who found that there were no significant differences between the children who received the parent-training program and a matched community-based sample with regards to academic achievement and a variety of delinquency/crime outcomes nearly 14 years following their participation in a parent-training program similar to those described above.
Daycare/School-Based Programs Although programs that fit into this particular category are often similar to those reviewed above with respect to the use of home visitation and/or parent training as a key program component, they are distinct in that the program itself is primarily admin- istered in the daycare/school setting. In general, these programs focus on the child’s intellectual and social development drawing from classical conditioning, cognitive- developmental, and psychoanalytic theories (Schweinhart, Berrueta-Clement, Barnett, Epstein, & Weikart, 1985). Furthermore, according to Farrington and Welsh (2003), the child-focused interventions emphasize intellectual enrichment and skills training sup- plemented with family support to prevent/reduce a number of adverse life outcomes.
There are a number of well-known early childhood intervention programs that are considered daycare/school-based programs such as the High/Scope Perry Preschool Program (Schweinhart et al., 1985), the Chicago Child-Parent Centers (Reynolds, Temple, Robertson, & Mann, 2001), the Carolina Abecedarian Study (Masse & Barnett, 2002), the Parent-Child Development Centers (Johnson, 2006), the Seattle Social Development Project (Hawkins, Kosterman, Catalano, Hill, & Abbott, 2005), and the Montreal Longitudinal Experimental Study (Tremblay, Pagani-Kurtz, Masse, Vitaro, & Phil, 1995). For example, the High/Scope Perry Preschool Program was a program that targeted at-risk youth (ages 3-4) in Ypsilanti, Michigan between 1967 and 1970. The long-term results of this randomized experiment have since suggested that the experimental children accumulated fewer arrests than the control children and
Cohen et al. 397
they were also more likely to graduate high school. These findings are rather robust considering the quality of the research design and the long-term follow-up, with Sch- weinhart (2007) reporting the consistency of these program effects observed through age 40.
Furthermore, the long-term results from evaluations of the Chicago Child-Parent Centers that emphasize high-quality, active learning have demonstrated that experi- mental children were less likely to be arrested for nonviolent or violent offenses in late adolescence/early adulthood and they also had a higher rate of high school completion compared with the control children (Reynolds et al., 2001, 2007; Temple & Reynolds, 2007). In addition, the long-term results from the Carolina Abecedarian Study (Masse & Barnett, 2002) and the Syracuse Family Development Program (Lally, Magoine, & Honig, 1987) were also consistent with those of the Perry Preschool Program and the Chicago Child-Parent Centers where the children that received the intensive center- based preschool program were more likely to have been enrolled in a 4-year college by age 21 (Masse & Barnett, 2002) and less likely to have been involved in delinquency (Lally et al., 1987). Yet despite these encouraging results, a long-term evaluation of the Parent-Child Development Centers did not find any significant differences between experimental and control children insofar as their involvement with the law in early/ late adolescence following their participation in an early preschool/school-based pro- gram designed to prevent school failure, enhance social competence, and prevent the development of behavior problems (Johnson, 2006; see also Johnson & Breckenridge, 1982; Johnson & Walker, 1987).
The Seattle Social Development Project is a large-scale school-based early interven- tion program that spanned across Grades 1 through 6. Beginning in 1981, the teachers in the randomly assigned intervention classrooms were trained in the core program elements that included the use of interactive teaching and cooperative learning. This teacher training and focus on child social and emotional skill development was also supplemented with parent-training exercises. The long-term results indicated that the experimental children demonstrated higher educational attainment, less drug and alco- hol use, and less involvement in a large variety of crimes by age 21 (Hawkins et al., 2005). Comparatively, Tremblay et al.’s (1995) Montreal Longitudinal Experimental Study is another well-recognized early childhood intervention program that focused primarily on two key components: (a) home-based parent training and (b) school- based social skills training. The results from a series of long-term follow-up studies evaluating this particular program have suggested that the experimental children had significantly lower rates of delinquency, were less likely to have a criminal record by age 24, and were more likely to have graduated high school (Boisjoli, Vitaro, Lacourse, Barker, & Tremblay, 2007; Tremblay et al., 1995).
Taken together, the literature reviewed above appears to suggest that early child- hood intervention—especially home visitation—is successful in preventing/reducing child behavior problems including crime and delinquency. Specifically, two recent meta-analyses on this topic have demonstrated that the inverse variance weighted effect size for early childhood interventions including home visitation programs, early
398 Criminal Justice Policy Review 21(4)
parent-training programs, and daycare/preschool programs is positive, moderate in strength, and relatively consistent across program type (Farrington & Welsh, 2003; Piquero et al., 2009). Furthermore, the literature reviewed above also suggests that early childhood interventions are effective in increasing educational attainment, pre- venting/reducing involvement with alcohol/drugs, along with affecting other child and maternal outcomes such as teenage pregnancy and child abuse and neglect. Consider- ing these empirical consistencies, we now provide a few brief examples of the costs of these particular high-risk activities.
Method Selection of Programs
Informed by evidence from a recent meta-analysis on early childhood interventions (Piquero et al., 2009), we conducted a literature search through a series of electronic research databases to identify potential early childhood prevention/intervention pro- grams that may be relevant for estimating the various costs described later. On locating these documents, we then reviewed and acquired other potentially relevant studies from the reference lists of large-scale narrative and systematic reviews in this line of research (Bernazzani, Cote, & Tremblay, 2001; Bernazzani & Tremblay, 2006; Bilukha et al., 2005; Farrington & Welsh, 2003, 2007; Gomby, Culross, & Behrman, 1999; Greenwood, 2006; McCart, Priester, Davies, & Razia, 2006; Serketich & Dumas, 1996). Considering the larger number of sources that this search yielded and to be consistent with our study’s intentions, we required that the program have long- term results from at least one rigorous evaluation on at least one of the outcomes of interest: (a) crime/delinquency; (b) education; (c) alcohol and drug abuse; (d) smoking; (e) child abuse and neglect; (f) physical health problems; and (g) teenage pregnancy. To assess whether a program affects one or more of these outcomes, we required the study to have a well-constructed comparison group whether randomly assigned or nonrandomly assigned provided that the study described evidence as to the direct comparability of the intervention and nonintervention groups.3 Based on this selection criterion, we did not include studies that lacked a comparison group, that is, pre-post research designs were excluded. Finally, the study had to provide quantifiable evidence to determine whether a program was successful at preventing/reducing the relevant outcomes of interest and it had to be a program that was capable of application/ replication in “real-world” settings.4
Following the exhaustive literature search and in light of the quality control criteria mentioned above, 14 well-designed studies with long-term follow-up results met the selection criteria (all of which have been reviewed earlier in this article). Table 1 details the programs and their respective studies that addressed the outcomes relevant to the cost estimates described below. Overall, the majority of these programs were classified as home-visitation programs specifically (although they may have other components) (Cullen & Cullen, 1996; McCormick et al., 2006; Olds et al., 1986; Olds, Kitzman,
399
T a b
le 1
. Pr
o je
ct s
w it h
Lo ng
-T er
m O
ut co
m es
A ut
ho r,
Pu bl
ic at
io n
D at
e
C ul
le n
an d
C ul
le n
(1 99
6)
H aw
ki ns
e t
al . (
19 99
, 20
05 );
Lo nc
za k,
A bb
o tt
, H
aw ki
ns , K
o st
er m
an ,
an d
C at
al an
o (
20 08
) Jo
hn so
n (2
00 6)
La lly
e t
al . (
19 87
)
Lo ng
e t
al . (
19 94
)
M as
se a
nd B
ar ne
tt (
20 02
)
M cC
o rm
ic k
et a
l. (2
00 6)
O ld
s et
a l.
(1 98
6, 1
99 7,
19
98 )
Sa
m pl
e Si
ze
E =
10 5/
C =
1 04
E =
15 6
(f ul
l- in
te rv
en ti o n)
/ C
= 2
20
E =
90 /C
= 8
1 (B
ir m
in gh
am );
E =
95 /C
= 1
76
(H o us
to n)
; E =
8 2
(c en
te r)
; E =
4 0
(h o m
e vi
si t)
/C =
75
( N
ew O
rl ea
ns )
E =
65 /C
= 5
4
E =
26 /C
= 2
6
E =
54 /C
= 5
1
E =
37 7/
C =
6 08
E =
79 (
pr en
at al
vi
si ts
); E =
97 (
pr e
an d
po st
na ta
l vi
si ts
)/ C
= 1
48
Pr
o je
ct N
am e
B us
se lt o n
Pr o je
ct
Se at
tl e
So ci
al
D ev
el o pm
en t
Pr o je
ct
Pa re
nt C
hi ld
D
ev el
o pm
en t
C en
te rs
Fa m
ily D
ev el
o pm
en t
R es
ea rc
h Pr
o gr
am Pa
re nt
-T ra
in in
g Pr
o gr
am A
be ce
da ri
an P
ro je
ct
In fa
nt H
ea lt h
an d
D ev
el o pm
en t
Pr o gr
am N
ur se
–F am
ily
Pa rt
ne rs
hi p
Pl
ac e
o f St
ud y
W es
te rn
A
us tr
al ia
Se at
tl e,
W
as hi
ng to
n
B ir
m in
gh am
, A
la ba
m a;
H o us
to n,
T ex
as ;
N ew
O rl
ea ns
, Lo
ui si
an a
Sy ra
cu se
, N ew
Yo
rk A
th en
s, G
eo rg
ia
C ha
pe l H
ill ,
N o rt
h C
ar o lin
a 8
pr o gr
am s
it es
El m
ir a, N
ew Y
o rk
D at
e o f
St ud
y
19 64
-1 97
3
19 81
19 70
-1 97
8
La te
1 97
0s
La te
1 97
0s
19 72
-1 97
7
19 84
-1 98
5
19 78
-1 98
0
A ge
a t
In te
rv en
ti o n
B ir
th
~ 7
ye ar
s o ld
2- 12
m
o nt
hs
o ld
B ir
th
~ 2-
7 ye
ar s
o ld
1- 4
m o nt
hs
o ld
B ir
th
B ir
th
Le ng
th o
f In
te rv
en ti o n
6 ye
ar s
6 ye
ar s
(G ra
de s
1- 6)
2- 3
ye ar
s
5 ye
ar s
~ 1
m o nt
h
5 ye
ar s
3 ye
ar s
2 ye
ar s
A ge
a t
Fo llo
w -U
p
27 -2
9 ye
ar s
o ld
21 y
ea rs
o ld
9- 16
y ea
rs
o ld
13 -1
6 ye
ar s
o ld
17 -2
2 ye
ar s
o ld
21 y
ea rs
o ld
18 y
ea rs
o ld
15 y
ea rs
o ld
(c
on tin
ue d)
400
T a b
le 1
. (c
o n
ti n
u e d
)
A ut
ho r,
Pu bl
ic at
io n
D at
e
O ld
s, K
it zm
an , e
t al
. (2
00 4)
; O ld
s et
a l.
(2 00
7) ; K
it zm
an e
t al
. (1
99 7,
2 00
0) O
ld s
et a
l. (2
00 2)
; O ld
s, R
o bi
ns o n,
e t
al . (
20 04
)
R ey
no ld
s et
a l.
(2 00
7) ;
Te m
pl e
an d
R ey
no ld
s (2
00 7)
Sc hw
ei nh
ar t
(2 00
7)
St . P
ie rr
e et
a l.
(1 99
7) ;
St . P
ie rr
e an
d La
yz er
(1
99 9)
B o is
jo li,
V it ar
o, L
ac o ur
se ,
B ar
ke r,
an d
Tr em
bl ay
(2
00 7)
; T re
m bl
ay e
t al
. (1
99 5)
Sa
m pl
e Si
ze
E =
21 3/
C =
4 77
E =
23 4
(p ar
ap ro
fe ss
io na
l) ;
E =
22 3
(n ur
se )/
C =
2 38
E =
88 8/
C =
4 80
E =
58 /C
= 6
5
E =
2, 21
3/ C
= 2
,1 97
E =
69 /C
= 1
81
Pr
o je
ct N
am e
N ur
se –F
am ily
Pa
rt ne
rs hi
p
N ur
se –F
am ily
Pa
rt ne
rs hi
p
C hi
ca go
L o ng
it ud
in al
St
ud y
H ig
h- Sc
o pe
P er
ry
Pr es
ch o o l P
ro gr
am C
o m
pr eh
en si
ve
C hi
ld D
ev el
o pm
en t
Pr o gr
am M
o nt
re al
L o ng
it ud
in al
Ex
pe ri
m en
ta l S
tu dy
Pl
ac e
o f St
ud y
M em
ph is
, Te
nn es
se e
D en
ve r,
C o lo
ra do
C hi
ca go
, I lli
no is
Y ps
ila nt
i, M
ic hi
ga n
21 p
ro gr
am s
it es
M o nt
re al
, Q
ue be
c,
C an
ad a
D at
e o f
St ud
y
19 90
-1 99
1
19 94
-1 99
5
19 84
-1 98
5
19 62
-1 96
7
19 90
-1 99
1
19 84
-1 98
6
A ge
a t
In te
rv en
ti o n
B ir
th
B ir
th
3 ye
ar s
o ld
3- 4
ye ar
s o ld
B ir
th
~ 7
ye ar
s o ld
Le ng
th o
f In
te rv
en ti o n
2 ye
ar s
2 ye
ar s
1. 6
ye ar
s (m
ea n)
1. 8
ye ar
s (m
ea n)
5 ye
ar s
2 ye
ar s
A ge
a t
Fo llo
w -U
p
9 ye
ar s
o ld
4 ye
ar s
o ld
22 -2
4 ye
ar s
o ld
40 y
ea rs
o ld
5 ye
ar s
o ld
24 y
ea rs
o ld
Cohen et al. 401
et al., 2004; Olds, Robinson, et al., 2004; Reynolds et al., 2007; St. Pierre et al., 1997), followed by daycare/school-based programs (Hawkins, Catalano, Kosterman, Abbott, & Hill, 1999; Hawkins et al., 2005; Johnson, 2006; Lally et al., 1987; Masse & Barnett, 2002; Reynolds et al., 2007; Schweinhart, 2007; Tremblay et al., 1995), and early parent- training programs (Long et al., 1994). Nearly all of the programs were implemented in the United States, with two exceptions: the Busselton Project (Western Australia) and the Montreal Longitudinal Experimental Study (Montreal, Quebec, Canada). The implementation of these programs span over four decades with the earliest of these programs being the High/Scope Perry Preschool Project (1962-1967) and the most recent program being the Nurse Family Partnership Project implemented in Denver, Colorado (1994-1995). The majority of these programs started during pregnancy or at birth and followed the children (and at times the mothers) for a considerable length of time, ranging from a 4-year follow-up (Olds, Robinson, et al., 2004) to more than 30 years of follow-up (Schweinhart, 2007).
Tables 2 through 6 describe the operationalization of the various outcomes addressed in the studies that are relevant for generating the cost estimates. The programs vary in their measurement of the specific outcomes of interest (self-report vs. official; prev- alence vs. incidence) and how many of the relevant outcomes for which they have quantifiable assessments.5
Estimating the Costs of Social Ills Next, we examine the costs of each social ill identified as being related to an at-risk childhood. We first conducted an extensive literature review on both the incidence of impacts and their costs. This review was similar to that done for the analysis of treat- ment programs, but required an even broader scope of literatures—including medical, education as well as criminal justice economics. To be included, a study must include a rigorous attempt to estimate marginal impacts and/or costs. Before providing details on the cost estimates, we provide some background on the underlying economic theory and methodologies employed to estimate the costs of social ills.
Cost Estimation Theory and Method Before one can estimate the costs of social ills, it is important to start with some simple but important economic concepts. There are many types of costs in economics, includ- ing private costs, taxpayer/public costs, social costs, and opportunity costs. A researcher who is interested in estimating the costs of a social ill should first identify the pur- pose of the costing exercise. This will dictate the type of costs to be estimated. For example, if one is only interested in the monetary burden to those afflicted with diabe- tes, we might only consider the out-of-pocket medical costs not covered by insurance and perhaps lost earnings if the individual has a reduced income due to this illness. However, a state budget officer might be interested primarily in the direct impli- cations (costs and benefits) to taxpayers. However, for most public policy purposes,
402 Criminal Justice Policy Review 21(4)
Table 2. Long-Term Crime/Delinquency Outcomes (Child)
Project Name
Seattle Social Development Project
Parent Child Development Centers Family Development Research
Program Parent-Training Program
Infant Health and Development Program
Nurse-Family Partnership (Elmira)
Chicago Longitudinal Study
High-Scope Perry Preschool Program
Montreal Longitudinal Experimental Study
Outcomes
• Lifetime violence • Lifetime nonviolent crime • Lifetime arrested • Involved in crime in past year • High variety of involvement in crime • Sold drugs • Court charge in past year • Court charge in lifetime • Trouble with the law • Juvenile delinquency (probation cases)
• Index offenses • Number of general delinquencies • Juvenile arrest • Been in jail • Incidence of times ran away • Incidence of times stopped by the police • Incidence of arrests • Incidence of convictions and probation violations • Incidence of times sent youth to corrections • Incidence of mother-reported arrests • Mean number of minor antisocial acts • Mean number of major delinquent acts • Incarceration or jail • Arrest • Felony arrest • Violent arrest • Conviction • Felony conviction • Violent conviction • Arrested 5 or more times as a juvenile • Arrested 5 or more times as a young adult (20-27 years old)
• Arrested 5 or more times by age 40 • Violent arrest • Drug arrest • Property arrest • Arrested for a felony • Delinquency (e.g., trespassing, stealing) • Criminal record
researchers are interested in a broader notion of costs—something akin to the burden on society. Economists would generally view this as the “social cost” of illness—anything that reduces the aggregate well-being of society.6 Note that under this definition, some
Cohen et al. 403
taxpayer costs such as social welfare are considered “transfers”7 from one segment of society to another, and not losses to society. For example, if we find that child abuse victims have lower lifetime earnings, we would value this lost productivity but would not count social welfare payments meant to offset some of those losses. In this exam- ple, lost productivity will always exceed the cost of social welfare payments meant to compensate for some of these losses, and to include social welfare would essentially double-count. However, we would want to include the administrative costs associated with those social welfare transfer payments. Our goal in this article is to estimate the total social costs associated with at-risk youth. Where possible, however, we attempt to
Table 3. Long-Term Educational and Alcohol and Drug Abuse Outcomes (Child)
Outcomes
• Attended a technical school • College degree • High school graduate • At least 2 years of college • Highest grade achieved • High school graduate or equivalent • Enrolled in 4-year college • Postsecondary training • High school dropout • Highest grade achieved • High school graduate • Special education • Attended college • Attended 4-year college • High school graduate • At least 2 years of college • High school graduate
• Alcohol use • Marijuana use • Other drug use • Substance use • Hard drug use • Incidence of days drank alcohol • Incidence of days used drugs • Alcohol impairment • Parent-reported alcohol and drug impairment • Substance use • Frequent substance use • Prescription drug abuse • Drug use
Project Name
Educational Busselton Project
Seattle Social Development Project
Parent-Training Program Abedcedarian Project
Infant Health and Development Program Chicago Longitudinal Study
High-Scope Perry Preschool Program
Montreal Longitudinal Experimental Study Alcohol and drug abuse
Seattle Social Development Project
Parent-Training Program Nurse–Family Partnership (Elmira)
Chicago Longitudinal Study
High-Scope Perry Preschool Program
404 Criminal Justice Policy Review 21(4)
identify components of costs such as medical costs, wage losses, and the like. Our abil- ity to do so depends on the underlying cost estimates that have been generated by other researchers.
Opportunity costs. Many costs are not observable as direct monetary exchanges. Economists have long recognized the notion of “opportunity costs” as being the con- ceptual approach to valuing such burdens. The opportunity cost of a good or service is simply its value in the next best alternative—or put differently, what must be given up in exchange for that good or service. For example, if a crime victim must spend on average 10 hours meeting with police, prosecutors, and attending court proceedings, it is important to include the opportunity cost of the victim’s time. Similarly, if a diabetes patient must make monthly visits to a physician for checkups, the value of travel and unproductive waiting time might be an important opportunity cost to consider. In these cases, the opportunity cost is generally based on the hourly earning capacity of the individuals affected under the theory that the individual values his/her time by at least the amount he/she could obtain in the next best alternative—which is at least the amount of wages he/she could earn in the workplace. Additional opportunity costs that should be estimated include the value of pain, suffering and lost quality of life to vic- tims as well as fear to the public at large for social ills such as crime.
Table 4. Long-Term Health, Teenage Pregnancy, and Child Abuse/Neglect Outcomes (Child)
Project Name
Health Busselton Project
Seattle Social Development Project Abecedarian Project Infant Health and Development Program Nurse–Family Partnership (Memphis) Chicago Longitudinal Study High-Scope Perry Preschool Program
Teenage pregnancy Parent Child Development Centers Seattle Social Development Project Nurse–Family Partnership (Elmira) Chicago Longitudinal Study Parent Child Development Centers
Child abuse/neglect Nurse–Family Partnership (Elmira)
Chicago Longitudinal Study
Outcomes
• Smoking • Obesity—Body mass index of ≥25 (child) • Cigarette use • Smoking • Physical Health Scale—reported by caretaker • Physical Health Scale—reported by child • Incidence of cigarettes smoked per day
• Daily tobacco use • Lost work for health problems
• Teenage pregnancy • Teenage/early adulthood pregnancy (<21 years old) • Teenage pregnancy • Teenage pregnancy (<18 years old) • Teenage pregnancy
• Child abuse in infancy • Child abuse and neglect between ages 4 and 15 • Child abuse and neglect
Cohen et al. 405
“Incidence-” versus “prevalence-” based costs. The health economics literature gener- ally distinguishes between “incidence-based” and “prevalence-based” costs (Hartunian, Smart, & Thompson, 1980). Prevalence-based cost estimates generally look at the costs imposed during 1 year and are thus of most interest to budget planners, public health economists attempting to estimate aggregate medical costs next year, and the like. In contrast, incidence-based cost estimates attempt to account for the entire lifetime of impacts on individuals. They are most appropriate for policy analysts who wish to compare the costs and benefits of a program designed to reduce or prevent long-term social ills. In this article, we only estimate incidence-based costs. Thus, for example, we attempt to estimate the lifetime costs of smoking and compute that amount for one individual smoker.
Present value of future costs. A dollar incurred in the future is not the same as a dollar spent today. Thus, costs we expect to incur (or benefits we expect to observe) 15 years from now, for example, must be discounted to present value when compared to the costs borne today.8 Although there is no general consensus on the appropriate discount rate for purposes of policy analysis, we use a 2% rate to be consistent with Cohen (1998) and Cohen and Piquero (2009), and all estimates are discounted to present value as of the date of birth and expressed in 2007 dollars.9 We also report estimates using a 3% discount rate.
The 2% discount rate is consistent with the “real” (i.e., net of inflation) discount rate for worker wages over time and the real consumer interest rate over time. This discount rate is also within the range most likely to be used in tort awards for lost wages. Furthermore, statistical modeling suggests workers apply a 2% discount rate when they trade off possible loss of future life years against extra earnings in the pres- ent (Moore & Viscusi, 1989). Finally, the Congressional Budget Office concluded from a review of the economic evidence that the most appropriate discount rate for public decision making was 2% (Hartman, 1990). Yet a similar consensus appears to have developed around a 3% net discount rate in health care economics (Gold, Siegel, Russell, & Weinstein, 1996) and Office of Management and Budget (2003) which recommends a 3% discount rate when using a “social” discount rate.10 Thus, we also report costs using a 3% discount rate.
Outcomes
• Incidence of arrests • Incidence of convictions • Incidence of days in jail • NYS arrests • NYS convictions • Number of arrests • Spent time in jail
Table 5. Long-Term Crime/Delinquency Outcomes (Mother)
Project Name
Nurse–Family Partnership (Elmira)
Nurse–Family Partnership (Memphis)
406 Criminal Justice Policy Review 21(4)
Table 6. Long-Term Educational, Alcohol and Drug Abuse, Health, and Teenage/At-Risk Births Outcomes (Mother)
Project Name
Educational Abecedarian Project Nurse–Family Partnership (Memphis) Nurse–Family Partnership (Denver) Comprehensive Child Development Program
Alcohol and drug abuse Nurse–Family Partnership (Elmira) Nurse–Family Partnership (Memphis)
Nurse–Family Partnership (Denver)
Comprehensive Child Development Program
Health Comprehensive Child Development Program
Teenage/at-risk births Abecedarian Project Nurse–Family Partnership (Elmira)
Nurse–Family Partnership (Memphis)
Nurse–Family Partnership (Denver)
Comprehensive Child Development Program
Outcomes
• Postsecondary training • High school graduate or equivalent • High school graduate or equivalent • Enrolled in vocational or job training program • High School Graduate or equivalent • Some college
• Substance use impairments • Number of substances used • Moderate/heavy drinker • Drug use • Alcohol use • Percentage who used alcohol • Percentage who used illegal drugs
• Smoked cigarettes
• Subsequent births • Subsequent pregnancies • Subsequent births • Subsequent pregnancies • Subsequent births • Subsequent miscarriages • Subsequent abortions • Subsequent low-birth-weight births • Subsequent still births • Subsequent births receiving special care in nursery • Subsequent pregnancies • Subsequent births • Subsequent miscarriages • Subsequent abortions • Subsequent low-birth-weight births • Subsequent low-birth-weight births • Subsequent premature births • Subsequent births receiving special care in nursery
Sorting out the marginal costs of a social ill. Another important concept used in prepar- ing our cost estimates was to focus on the “marginal” costs caused by each social ill under consideration. The marginal costs of a social ill include all costs that would not have been incurred in the absence of the underlying cause of interest. For example, when studying the cost of illness, a common approach in the medical literature is to
Cohen et al. 407
compare two population means—those who have the illness and those who do not. This might provide an estimate of the average medical costs of treating someone with the illness but might not inform policymakers of the expected benefit of reduc- ing one of the causes of the illness. For example, suppose teen mothers have a higher rate of birth complications than the population at large and that a higher percentage of teen mothers are represented in low-income populations. If one were interested in estimating the effect of teen pregnancy on birth complications, it would be inappropri- ate to take the average cost of birth complications in the United States as an estimate of the value of reducing teen pregnancy. Instead, we want to know the marginal cost of birth complications associated with teen pregnancy. To estimate this, a more appro- priate comparison would control for maternal income—perhaps through a multiple regression analysis or matched sample approach—to isolate the impact of teen preg- nancy from the effect of income. Given the way that many “health cost” studies are designed, this proved to be a challenge in preparing the cost estimates for this article and greatly reduced the number of studies we could rely on. More specifically, even when we could isolate the incremental effect that a social ill has on costs, oftentimes we are forced to make assumptions that costs do not vary by underlying cause. For example, although we have estimates of the cost of diabetes, we do not know if those costs vary by diabetes caused by genetic factors versus obesity. To the extent these costs vary, we might be under- or overestimating the costs of diabetes associated with childhood obesity.
Methods for Estimating Costs of Social Ills There are two types of studies that attempt to estimate costs for most of these social ills.11 The more common approach is a “bottom-up” estimate of each cost component such as medical costs, lost wages, and the value of time spent by caregivers. In some cases, these studies will include intangible costs—by valuing lost life years and/or lost quality of life due to disability. The other approach is a “top-down” method using survey techniques to estimate individuals’ willingness-to-pay (WTP12) to reduce the risk (or certainty) of a particular illness.
To date, this top-down approach has provided more comprehensive estimates than the bottom-up approach. Not only do the top-down estimates take into account the impacts endured by those who suffer from the social ill (such as lost productivity, pain, suffering, and lost quality of life) and government costs (such as police, courts, prisons, and social welfare agencies) but they also incorporate other social costs such as fear of crime, avoidance behaviors, and loss in neighborhood cohesion—that is, the public’s WTP to avoid the social ill.13
In theory, WTP estimates should include any costs expected to be incurred directly by respondents—that is, out-of-pocket expenses but might not include costs imposed on others (e.g., health care costs paid by insurance companies or taxpayers). Thus, in theory, to arrive at total social costs, one would add that portion of tangible costs that are not paid by the respondent to the WTP estimates. In practice, there have been few
408 Criminal Justice Policy Review 21(4)
attempts to assess the extent to which respondents on WTP surveys take these costs into account, and generally, WTP estimates are taken at face value without adding other costs not considered by the respondent.14 In some cases (such as the health care costs associated with coronary heart disease), WTP can be used to identify the private value of pain, suffering, and fear. In other cases, however, there are social costs borne beyond those who are directly affected by the social ill. For example, the cost of crime includes not only the pain and suffering to crime victims but also the cost to communities from lack of social cohesion, reduced economic activity, and the like.
Our goal is to build incidence-based estimates over the child’s lifetime following the approach used in Cohen (1998) and Cohen and Piquero (2009). Although there are many studies that attempt to total up the annual costs of specific illnesses in the United States, those studies are generally based on the prevalence of illness in the population—not the cost associated with a typical individual over their lifetime. Thus, most of these aggregate estimates are of limited value. Instead, we searched for studies that estimated the annual cost per person—by age if possible. However, in some cases, if we assume a steady-state percentage over time in a population, a reasonable approx- imation of average annual costs per person can be derived by dividing aggregate annual costs by the number of individuals affected. When possible, we have tried to obtain three estimates—tangible costs, intangible costs, and WTP estimates. Due to the scar- city of studies, we often fell short of that goal. In all cases, however, we specify what costs are available and provide some detail so the reader can understand the basic approach. Due to space limitations, we have not replicated the analysis in the source documents. In the remainder of this section, we examine the costs of each social ill identified as being related to an at-risk childhood. We consider in turn, (a) crime and delinquency, (b) educational attainment, (c) drug abuse, (d) alcohol abuse, (e) smoking, (f) child abuse and neglect, (g) poor health outcomes, and (h) teen pregnancy.15 Teen pregnancy costs are enumerated in Table 7, and all costs are summarized in Table 8. Whereas we report present value costs based on a 2% discount rate in the text, Tables 7 and 8 also estimate costs based on a 3% discount rate. Table 9 attempts to separate costs out by tangible, intangible, and WTP costs whenever possible. As explained above, as tangible and intangible costs are generally estimated using bottom-up approaches, they generally add up to less than the WTP amount.
Crime and delinquency. Cohen and Piquero (2009) estimate the cost imposed by individuals in a cohort of youth through age 26. Using a bottom-up approach as in Cohen (1998), they include estimates of victim costs (lost wages, medical costs, and pain and suffering), criminal justice costs (police, courts, prisons), and the opportunity costs of a criminal’s time while in prison. They also prepared estimates based on a top-down approach taken from a public survey of WTP to reduce crime (Cohen, Rust, Steen, & Tidd, 2004).
Cohen and Piquero estimate the present value of costs imposed by a career criminal (defined as an individual with six or more police contacts over their lifetime) to range from US$2.1 to US$3.7 million (discounted to the date of birth at a 2% discount
Cohen et al. 409
rate).16 Offense data are obtained from the Second Philadelphia Birth Cohort Study (born in 1958). Costs are based on society’s WTP for reduced crime.
Of course, only a small percentage of offenders become career criminals— approximately 4% of the Philadelphia cohort, as estimated by Cohen and Piquero (2009). They also estimate that about 23% of the cohort had at least one police con- tact by age 26. Cohen and Piquero (2009, Table 6) also report on the costs imposed by offenders who have only one police contact through age 26. Individual offenses range from loitering to murder. Averaged over all offenses, the present value as of age 8 was estimated to be US$90,268—or about US$78,500 in present value terms as of birth. In the summary tables, we report on the costs both of a “career criminal” and a “one-time offender.”
Educational attainment. The cost of dropping out of high school was estimated by Cohen and Piquero (2009) to range between US$300,000 and US$450,000 (discounted to present value as of the date of birth). The largest component is US$250,000 which includes lower earnings and fringe benefits. These estimates are taken from Census data comparing earnings by age for high school dropouts versus those who graduated or completed a GED.17 The additional loss estimated between US$50,000 and US$200,000 represents additional “nonmarket” social losses such as less social cohesion, more equitable income distribution, and charitable giving (see Cohen, 1998).
Table 7. Present Value of Lifetime Costs of Teen Pregnancy
Medical costs Child abuse and
neglect Foster care
High school dropout (child)
Juvenile and Adult Crime (child)
Father’s earnings Subtotal Second
Generation Costs
Total
Discount Rate 2% (US$)
1,800 8,800
6,600
21,000-32,000
16,000-24,000
62,000 116,200-135,200
2,600-3,000
118,800-138,200
Discount Rate 3% (US$)
1,400 8,100
6,000
14,000-21,000
13,000-20,000
41,000 83,500-97,500 1,900-2,200
85,400-99,700
Sources
Maynard and Hoffman (2008) Tangible and intangible costs (Miller,
Cohen, & Wiersema, 1996) Cost of providing service (Hoffman
& Scher, 2008a) Earnings plus benefits plus social
value of education (Cohen & Piquero, 2009), multiplied by 7% higher dropout rate (Hoffman & Scher, 2008a)
WTP (Cohen & Piquero, 2009)a
Earnings plus benefits
3.1% of first generation costs, discounted to present value
a. All costs in this table are tangible and intangible (i.e., bottom up) with the exception of the costs of juvenile and adult crime which are WTP costs.
410 Criminal Justice Policy Review 21(4)
Table 8. Summary of Present Value of Lifetime Costs of Social Ills
Child abuse and neglect
Health outcomesb
Career criminala
One-time offendera
High school dropout Drug abusea
Alcohol dependence and abuse
Smoking
Impact on abused child Subsequent offending behavior
Diabetes Asthma
Coronary heart disease Low birth weight Teen pregnancy
Present Value 2% (US$)
2.1-3.7 million
78,500
300,000- 450,000 650,000- 740,000
690,000
1,140,000
200,000
55,000-85,000
153,000 144,000
127,000
10,300
118,800- 138,200
Present Value 3% (US$)
1.8-3.1 million
53,000
200,000- 300,000 430,000- 490,000
465,000
740,000
180,000
45,000-70,000
103,000
111,000
78,000
10,300
85,400-99,700
Notes
WTP (from Cohen & Piquero, 2009, Table 12)
WTP (from Cohen & Piquero, 2009, Table 6)
WTP (from Cohen & Piquero, 2009, Table 12)
WTP (from Cohen & Piquero, 2009, Table 12)
US$230,000 tangible costs including productivity, motor vehicle crashes, social welfare, and fire-related costs. US$460,000 estimated intangible costs. Note: additional costs of crime as well as social costs associated with homelessness and the like might be attributable to alcohol abuse. See text.
WTP based on worker wage rates and smoking behavior. Likely excludes cost of second hand smoke or other costs not directly borne by smokers.
Tangible and intangible costs (Miller, Cohen, & Wiersema, 1996)
WTP. Note that some portion of these costs likely overlap with the cost of a career criminal.
WTP US$85,000 tangible costs, US$84,000
WTP. Estimated 30% of tangible costs paid by asthmatic, the rest through insurance or government agencies. Thus, 70% of tangible costs are added to WTP costs to arrive at total.
Medical costs, earnings, avoidance costs, and WTP
Hospital costs only
See Table 7 above.
Juvenile and adult behavior
a. These costs cannot be added together as a significant portion of the cost of heavy drug abuse involves criminal activity that are included in career criminal costs. Overlap between these two categories totals US$800,000, or US$460,000 in present value terms (Cohen & Piquero, 2009, Table 12). b. There may be some overlap between these health outcomes—especially diabetes and coronary heart disease; hence, adding them might result in some double-counting. This is also true with coronary heart disease and alcohol dependency and abuse.
411
T a b
le 9
. T an
gi bl
e, I nt
an gi
bl e,
a nd
W T
P C
o st
s
Pr
es en
t V al
ue (
di sc
o un
te d
to d
at e
o f bi
rt h
at 2
% r
at e)
Pr
es en
t V al
ue (
di sc
o un
te d
to d
at e
o f bi
rt h
at 3
% r
at e)
Ta
ng ib
le (
U S$
) In
ta ng
ib le
( U
S$ )
W T
P (U
S$ )
Ta ng
ib le
( U
S$ )
In ta
ng ib
le (
U S$
) W
T P
(U S$
)
Ju ve
ni le
a nd
a du
lt b
eh av
io r
C
ar ee
r cr
im in
al a
28 5,
00 0-
46 0,
00 0
43 0,
00 0-
69 0,
00 0
2. 1-
3. 7
m ill
io n
24 5,
00 0-
38 5,
00 0
37 0,
00 0-
58 0,
00 0
1. 8-
3. 1
m ill
io n
O ne
-t im
e o ffe
nd er
b 20
,8 00
13
,8 00
78
,5 00
9,
30 0
14 ,0
00
53 ,0
00 H
ig h
sc ho
o l d
ro po
ut c
25 0,
00 0
50 ,0
00 -2
00 ,0
00
? 17
0, 00
0 35
,0 00
-1 35
,0 00
?
D ru
g ab
us ed
25
0, 00
0- 34
0, 00
0 10
0, 00
0 65
0, 00
0- 74
0, 00
0 21
5, 00
0- 29
0, 00
0 80
,0 00
54
5, 00
0- 62
0, 00
0 A
lc o ho
l d ep
en de
nc e
an d
ab us
e 23
0, 00
0 46
0, 00
0 ?
15 5,
00 0
31 0,
00 0
? Sm
o ki
ng e
? ?
26 0,
00 0
? ?
26 0,
00 0
Im pa
ct o
n ab
us ed
c hi
ld f
22 ,0
00
17 8,
00 0
20 0,
00 0
20 ,0
00
16 0,
00 0
18 0,
00 0
Su bs
eq ue
nt o
ffe nd
in g
be ha
vi o rg
8,
80 0-
13 ,2
00
13 ,2
00 -1
9, 80
0 50
,0 00
-7 5,
00 0
6, 70
0- 10
,2 00
10
,0 00
-1 5,
30 0
38 ,0
00 -5
8, 00
0 D
ia be
te s
15 3,
00 0
27 ,0
00
18 0,
00 0
10 3,
00 0
18 ,0
00
12 1,
00 0
A st
hm ah
85
,0 00
59
,0 00
84
,0 00
65
,0 00
46
,0 00
64
,0 00
C o ro
na ry
h ea
rt d
is ea
se
11 2,
00 0
15 ,0
00
12 7,
00 0
69 ,0
00
9, 00
0 78
,0 00
Lo w
b ir
th w
ei gh
ti 10
,3 00
?
? 10
,3 00
?
? Te
en p
re gn
an cy
j 97
,5 00
-1 10
,0 00
12
,3 00
-1 4,
50 0
? 67
,2 00
-7 5,
70 0
10 ,9
00 -1
2, 80
0 ?
N o te
: ? =
e st
im at
es n
o t
av ai
la bl
e.
a. W
T P
fo r
cr im
e co
nt ro
l i nc
lu de
s a
si gn
ifi ca
nt p
o rt
io n
o f b
en ef
it s
to p
o te
nt ia
l v ic
ti m
s (i .e
., fe
ar o
f c ri
m e)
. T hu
s, W
T P
is s
ig ni
fic an
tl y
hi gh
er t
ha n
th e
su m
o f t
an gi
bl e
an d
in ta
ng ib
le v
ic ti m
c o st
s. C
o he
n an
d Pi
qu er
o (
20 09
) al
so r
ep o rt
o n
co st
s us
in g
a “b
o tt
o m
u p”
a pp
ro ac
h, w
hi ch
y ie
ld s
es ti m
at es
o f b
et w
ee n
31 %
( o f t
he la
rg er
e st
im at
e) a
nd 3
4% (
o f t
he s
m al
le r
es ti m
at e)
o f W
T P
co st
s. H
o w
ev er
, t ho
se c
o st
s ca
n be
fu rt
he r
di vi
de d
in to
t an
gi bl
e an
d in
ta ng
ib le
c o st
s us
in g
th e
ra ti o o
f t an
gi bl
e to
in ta
ng ib
le c
o st
s fr
o m
C o he
n (2
00 0)
, w hi
ch fi
nd s
th at
in ta
ng ib
le q
ua lit
y o f l
ife c
o st
s to
v ic
ti m
s ac
co un
ts fo
r ap
pr o xi
m at
el y
75 %
o f v
ic ti m
c o st
s an
d th
at c
ri m
in al
ju st
ic e
sy st
em c
o st
s ar
e ab
o ut
e qu
al t
o t
an gi
bl e
vi ct
im c
o st
s. T
hu s,
in ta
ng ib
le c
o st
s ar
e ab
o ut
6 0%
o f t
he t
o ta
l “ bo
tt o m
u p”
c o st
e st
im at
e. b.
C o he
n an
d Pi
qu er
o (
20 09
) al
so r
ep o rt
o n
co st
s us
in g
a “b
o tt
o m
u p”
a pp
ro ac
h, w
hi ch
y ie
ld s
es ti m
at es
o f ab
o ut
4 4%
o f th
e W
T P
co st
fo r
a o ne
-t im
e o ffe
nd er
. T hu
s, th
e “b
o tt
o m
u p”
c o st
s ar
e es
ti m
at ed
t o
be a
bo ut
U S$
34 ,5
40 (
U S$
78 ,5
00 ×
4 4%
). U
si ng
t he
s am
e m
et ho
d as
a bo
ve , t
an gi
bl e
co st
s ar
e es
ti m
at ed
t o b
e 40
% (
U S$
13 ,8
00 )
an d
in ta
ng ib
le c
o st
s 60
% (
U S$
20 ,8
00 ).
Si m
ila r
es ti m
at es
a re
m ad
e us
in g
a 3%
di
sc o un
t ra
te .
c. W
e ar
e un
aw ar
e o f an
y W
T P
es ti m
at es
. T he
“ in
ta ng
ib le
c o st
s” m
ig ht
a ct
ua lly
in cl
ud e
so m
e ta
ng ib
le c
o st
s bu
t ha
ve n
o t
be en
s ys
te m
at ic
al ly
e st
im at
ed in
a m
an ne
r th
at a
llo w
s fo
r su
ch a
b re
ak do
w n.
S ee
C
o he
n (1
99 8)
fo r
de ta
ils o
n th
is li
te ra
tu re
. d.
C o he
n an
d Pi
qu er
o (
20 09
) ex
cl ud
e an
y in
ta ng
ib le
c o st
s as
so ci
at ed
w it h
dr ug
a bu
se —
w it h
th e
o nl
y ex
ce pt
io n
be in
g th
e in
ta ng
ib le
c o st
s to
c ri
m e
vi ct
im s
w ho
a re
v ic
ti m
iz ed
b y
dr ug
a bu
se rs
. S im
ila rl
y, th
er e
ar e
no W
T P
es ti m
at es
fo r
dr ug
a bu
se o
th er
t ha
n fo
r “d
ru g
at tr
ib ut
ab le
” cr
im es
. T hu
s, al
th o ug
h th
e sa
m e
pr o ce
du re
a s
ab o ve
h as
b ee
n us
ed t
o e
st im
at e
in ta
ng ib
le a
nd W
T P
co st
s, th
e W
T P
co st
s re
pr es
en t
th e
co st
t o v
ic ti m
s o nl
y. e.
T he
se W
T P
es ti m
at es
a re
b as
ed o
n a
13 .9
% d
is co
un t
ra te
a s
re po
rt ed
in S
ch ar
ff an
d V
is cu
si (
20 08
). Se
e te
xt .
f. B
as ed
o n
M ill
er e
t al
. ( 19
96 ),
ta ng
ib le
c o st
s o f ch
ild a
bu se
a re
o nl
y 11
% o
f th
e to
ta l c
o st
s. g. T
he s
am e
pr o ce
du re
d es
cr ib
ed a
bo ve
in e
st im
at in
g th
e co
st o
f a
ca re
er c
ri m
in al
is u
se d
he re
. T hu
s, “b
o tt
o m
u p”
c o st
s ar
e es
ti m
at ed
t o b
e ab
o ut
4 4%
o f th
e W
T P
co st
. T an
gi bl
e co
st s
ar e
es ti m
at ed
t o b
e 40
% o
f th
is a
m o un
t, w
he re
as in
ta ng
ib le
c o st
s ar
e 60
% .
h. I nt
an gi
bl e
co st
s ar
e es
ti m
at ed
f ro
m t
he p
ri va
te W
T P.
A s
it is
e st
im at
ed t
ha t
30 %
o f ta
ng ib
le c
o st
s ar
e pa
id b
y as
th m
at ic
s (t
he r
es t
th ro
ug h
in su
ra nc
e o r
go ve
rn m
en t)
, 7 0%
o f ta
ng ib
le c
o st
s ar
e ad
de d
to
W T
P to
a rr
iv e
at t
o ta
l c o st
s. i.
H o sp
it al
c o st
s o nl
y. j. A
ll co
st s
ar e
ta ng
ib le
e xc
ep t
(a )
ch ild
a bu
se , w
hi ch
is o
nl y
11 %
t an
gi bl
e co
st s
an d
89 %
in ta
ng ib
le c
o st
s ba
se d
o n
M ill
er e
t al
. ( 20
06 )
an d
(b )
ju ve
ni le
a nd
a du
lt c
ri m
e co
m m
it te
d by
t he
c hi
ld . I
n th
e la
tt er
ca
se , a
s no
te d
in N
o te
b , t
he “
bo tt
o m
u p”
e st
im at
e is
o nl
y 44
% o
f W T
P co
st s
an d
ta ng
ib le
c o st
s re
pr es
en t
40 %
o f th
is a
m o un
t, w
he re
as in
ta ng
ib le
c o st
s ar
e 60
% .
412 Criminal Justice Policy Review 21(4)
Drug abuse. Cohen and Piquero (2009, Table 10) estimate the costs of heavy drug abuse, defined to be the 3.5 million Americans who are regular users of cocaine, crack, heroin, or methamphetamine. They estimated seven different cost components: (a) resources devoted to the drug market, (b) drug treatment for abusers, (c) reduced productivity, (d) medical costs, (e) premature death, (f) criminal justice costs associ- ated with “drug-defined crimes” such as possession and trafficking, and (g) additional crimes committed by drug abusers to support their habit or otherwise “attributable” to their drug abuse. They estimate the present value of the cost of a heavy drug abuser to range between US$650,000 and US$740,000 as of the date of birth. However, US$460,000 of this amount is estimated to be the cost of crimes committed by drug abusers to support their addiction. Thus, this amount partially overlaps with the cost of a career criminal above.18
Alcohol abuse. Harwood, Fountain, and Livermore (1998) estimated the annual aggregate cost in the United States from alcohol abuse, including health care costs, lost productivity, criminal justice system and victim costs, welfare costs, motor vehicle crashes, and fire damage costs. In deriving these aggregate estimates, however, they report on detailed regression analyses of individual data and report average losses per individual. In our estimates, we exclude the costs of crime attributable to alcohol abuse in the Harwood et al. (1998) study because the causal connection between alcohol and crime is not well established.19 Moreover, we have separately estimated the costs of crime imposed by career criminals—which presumably would overlap considerably with alcohol abuse.
According to Harwood et al. (1998, Section 2.2.1), 7.4% of adults in 1992 were either alcohol dependent or alcohol abusers. Based on 1992 Census estimates of 188.9 million Americans aged 18 and older, this translates into about 14 million individuals. In cases where they provide annual costs per person, we can use these directly to estimate present value costs of a lifetime of abuse. However, in many cases, they only report aggregate annual costs—which are prevalence-based instead of incidence-based costs.
Health care expenditures for alcohol treatment were estimated to be US$4.046 billion in 1992, whereas the cost of alcohol-related medical consequences (e.g., hospitalization, nursing homes) totaled US$13.247 billion (Harwood et al., 1998, Table 4.1). They also estimate that about US$7.6 billion was spent on social welfare programs; however, only US$683 million of that amount was spent on administration—and is thus considered a social cost (Harwood et al., 1998, Section 6.3.1). The remaining US$7 billion is a cost to taxpayers, but it is transfer, not a net loss to society.20 Combined, health care expenditures are thus estimated to be US$17.976 billion—or US$1,284 per person—annually.
Harwood et al. (1998, Section 6.4) estimate the annual cost of motor vehicle crashes attributed to alcohol abuse to be US$13.6 billion in 1992. In contrast, Miller, Cohen, and Wiersema (1996, Table 5) estimated the annual tangible costs of drunk driving crashes to be US$20.6 billion in 1993 dollars. About two thirds of these costs are for nonfatal injuries, whereas the remaining one third involved fatal crashes. As it is dif- ficult to reconcile the two different estimates, we take the average of these two studies, US$17 billion for tangible losses (about US$1,214 per person).
Cohen et al. 413
Productivity losses are estimated from two sources—mortality and morbidity. Fatalities include direct alcohol-related illnesses, such as alcoholic cirrhosis, and an estimated fraction of alcohol “attributed” causes, such as 15% of deaths by accidental falls where alcohol was involved (Harwood et al., 1998, Table 5.5). Overall, they esti- mate a total of 107,360 deaths in 1992 due to alcohol abuse, which would represent 0.77% of the estimated 14 million alcohol abusers annually. Harwood et al. (1998, Section 5.11) also report the present value of lost productivity of US$290,000 per person (based on a 3% discount rate)—which translates into an average annual cost per alcohol abuser of about US$2,230 (US$290,000 × .0077). Based on a 2% discount rate, total costs would be approximately US$3,325 per person annually.
For morbidity, Harwood et al. (1998, Section 5.3.7) estimate monthly earnings losses of US$260 (in 1992 dollars) for “ever-alcohol-dependent” males (females are found to have no statistically significant difference in earnings). This amount is mul- tiplied by 1.4 to account for fringe benefits and 1.1206 to account for lost household productivity.21 Thus, annual losses are estimated to average US$3,846 in 1992 dollars. Based on an estimate that there are twice as many males with alcohol disorders as females (Harwood et al., 1998, Table 5.8), the average cost per person is estimated to be US$2,564.
Finally, Harwood et al. (1998, Section 6.5) estimate the cost of fire damage and fire protection attributed to alcohol abuse to be US$1.59 billion—or about US$113 per alcohol abuser—annually. This is primarily due to inebriated smokers who fall asleep while smoking. Although this loss could also be attributed to cigarette smoking, we have included it here as not all of these costs are related to smoking.22
Combined, the Harwood et al. (1998) estimates imply annual tangible costs associ- ated with alcohol abuse of US$5,175 per person in 1992 dollars (US$1,284 health care, US$1,214 motor vehicle crashes, US$2,564 lost productivity due to morbidity, and US$113 fire damage). They also estimate approximately US$3,325 per year in lost productivity due to mortality. To estimate the present value of costs throughout the lifetime of an individual alcohol abuser requires that we account for this reduced life expectancy. A recent study by the Center for Disease Control (CDC, 2004) estimates that deaths attributable to alcohol abuse result in an average of 30 years of life lost. This is approximately 50% of the life expectancy for a typical 18-year-old in the United States. Thus, in estimating the present value of future nonfatality-related costs, we assume a 50% uniform reduction of life expectancy from age 18 onward. Updating to 2007 dollars and discounting to present value from the time of birth (assuming alcohol abuse starts at age 18), we estimate the costs of morbidity (through lost pro- ductivity) to be US$130,000 per alcohol abuser, and the total costs associated with health care costs, lost productivity due to morbitity, motor vehicle crashes, and fire damage to be US$100,000. Combined, the tangible cost of alcohol abuse is thus esti- mated to be US$230,000 per person.
Harwood et al. (1998) did not estimate intangible losses from alcohol abuse. We are aware of only a handful of studies that have done so. Miller et al. (1996, Table 5) esti- mated the annual tangible costs of drunk driving crashes to be US$61 billion—US$20.6
414 Criminal Justice Policy Review 21(4)
billion in tangible losses and US$39.4 billion in lost quality of life (1993 dollars). Thus, the ratio of intangible to tangible costs was estimated to be about 1.9 to 1 (US$39.4/ US$20.6). Jeanrenaud, Pellegrini, and Widmer (2003) conducted a WTP survey in Switzerland eliciting valuation for avoiding the physical and psychological damage caused by alcohol abuse and estimated total costs of 13,700 Swiss Francs in 1998 dol- lars (the equivalent of approximately US$12,300 at the time). Of this amount, one third (US$4,100) is estimated to be tangible losses and two thirds (US$8,200) WTP for reduced physical and psychological effects of alcohol abuse. This results in a ratio of intangible to tangible costs of about 2.0 to 1 (US$8,200/US$4,100). However, the tan- gible costs exclude lost productivity. Although it is problematic to compare across countries given different health care systems and standards of living, these figures are not out of line with the estimates by Harwood et al. Moreover, they provide us with another rationale for multiplying tangible costs by two to arrive at intangible costs (although the multiple in this study would be smaller if productivity losses were included). Thus, if we include intangible costs, the estimated present value of the cost of alcohol abuse is US$690,000.
Estimating the social costs of alcohol abuse is more problematic, as we are unaware of any studies of the public’s WTP for reduced alcohol abuse. The intangible costs included above are based on an assumption that the individuals (and perhaps their fami- lies) suffer beyond the pure monetary losses associated with alcohol-related illnesses and death. However, it is quite plausible that the public at large would be willing to pay an additional amount to avoid some of the consequences of serious alcohol abuse such as homelessness, vagrancy, panhandlers, and general social decay in certain neighbor- hoods. Additional costs might be incurred by families of alcohol abusers and the general public through fear and avoidance behavior associated with drunk driving. Thus, the estimates presented here are likely an understatement of the value of reduced alcohol abuse, and we do not present WTP estimates here.
Smoking. Viscusi (1995) estimated that the external costs associated with smoking (e.g., higher taxpayer funded medical expenses, second hand smoke, fires, etc.) are completely offset if not more than offset by the savings that society receives through lower outlays for social security, Medicare, and so on. In other words, he finds that the net external cost imposed by smokers is essentially zero. However, this excludes the private costs that smokers themselves endure. While economists would normally exclude private costs that are voluntarily incurred (as opposed to those that are consid- ered externalities and not voluntary) from a social welfare analysis, the situation is not so clear in the case of youth who have not yet begun smoking. If there are private costs associated with smoking, then preventing a child from becoming a smoker will result in a private benefit to that youth over his or her lifetime—something that society values and would be considered a social benefit.
Viscusi and Hersch (2008) estimate the private cost of smoking by estimating the statistical value of life by age and smoking status using hedonic wage equations from age 20 on. They estimate the private mortality cost of smoking to be US$222 per pack for men and US$94 for women in 2006 based on a 3% discount rate. However, Scharff
Cohen et al. 415
and Viscusi (2008, p. 20) estimate that a smoker’s implicit discount rate is 13.9%, which reduces the “per pack” cost to US$24 for men and US$6 for women. Updated to 2007 dollars and discounted to present value as of the date of birth yields an esti- mated cost of US$260,000. This is an estimate of the present value of the long-term private benefits of not becoming a smoker based on a 13.9% discount rate.23 These estimates include intangible losses and WTP as they are based on actual worker wage rates and smoking behavior.
Child abuse. Miller et al. (1996, p. 16) estimated the average cost of child abuse (including physical, sexual, and emotional abuse) to be US$74,000 in 1993 dollars. Although the largest cost component of this was pain and suffering, tangible losses still totaled more than US$8,000 per victim and include medical costs, mental health care costs, and the cost of social service agencies (including foster care). Note that this was an average over all categories of child abuse and represent lifetime costs including long-term physical and mental health consequences. The cost of child rape or sexual assault was higher (US$87,000), whereas nonsexual child abuse with injury was lower (US$24,000). It is important to note that although the Miller et al. (1996) estimates include pain and suffering to the victim, these calculations are based on specific inju- ries associated with child abuse and do not necessarily reflect society’s broader WTP for reduced child abuse. Cohen et al. (2004) estimated WTP for reduced crime—including assault and rape—and found that the WTP for these two crimes was approximately twice the cost estimated by Miller et al. (1996).
Updating to 2007 dollars and using a multiple of two to account for WTP, we esti- mate the cost of child abuse to be US$236,000 per child. Because child abuse might occur at any point prior to age 18, we have estimated the age distribution of victims and used that distribution to further reduce the present value estimates to the date of birth.24 Using this approach, the present value cost of child abuse from birth is estimated to be US$200,000.
Widom and Maxfield (2001, p. 3; see also Maxfield, Weiler, & Widom, 2000) com- pared juvenile and adult records for a sample of 908 abused and neglected children to a comparison group selected for their sociodemographic characteristics. Results indi- cated that 17.2% of the comparison group was arrested as juveniles, compared to 27.4% of those who were abused. Similarly, 32.5% of the comparison group was arrested as adults compared to 41.6% of the child abuse victims. If their comparison group adequately accounts for other factors, an additional 10% of child abuse victims will ultimately be arrested. Thus, we estimate that an additional 10% of child abuse and neglect victims will themselves become juvenile and/or adult criminal offenders as a result of the abuse.25 However, we do not know whether these additional offenders are career criminals or one-time offenders.
Cohen and Piquero (2009, Table 6) estimated the present value WTP costs associ- ated with a “high-risk youth” who has at least one police contact through age 26 to range between US$580,000 and US$880,000.26 This figure was discounted as of age 8 and is reported in 2007 dollars. Discounting further to birth and applying the 10% excess offending rate from Widom and Maxfield (2001), the present value costs of
416 Criminal Justice Policy Review 21(4)
additional crimes committed by abused children are estimated to be approximately US$50,000 to US$75,000. Since we do not know what portion of these offenders are “career criminals,” it would not be appropriate to add these costs to the career criminal costs above.
Health Outcomes In this section, we estimate the costs of four negative health states often associated with poor parenting, childhood development, and even antisocial behavior—asthma, diabetes, CHD, and low birth weight (e.g., Jackson & Dickinson, 2009; Piquero, Daigle, Gibson, Piquero, & Tibbetts, 2007). Although we have attempted to estimate only the costs associated with each of these illnesses, there will inevitably be some overlap—especially between diabetes and CHD. For example, an obese smoker might have a higher risk of both—and certain symptoms such as circulatory problems may also be attributed to both chronic illnesses. Thus, it would neither be appropriate to add all three illnesses together nor be appropriate to add these costs to the cost of smoking and alcohol or drug abuse. Also, unlike the costs of crime and alcohol and drug abuse, it is likely that most of the costs associated with these illnesses are borne either directly by the individual, family members, or the public at large through higher insurance and/or tax payments for health care costs. However, there are unlikely to be significant additional social costs such as fear or lack of social cohesion normally asso- ciated with crime. Thus, accounting for tangible and intangible costs of the illnesses themselves should be relatively comprehensive.
Asthma. Smith et al. (1997) studied the costs of asthma based on the 1987 National Medical Expenditure Survey (NMES), a survey of 35,000 individuals representing the noninstitutionalized, civilian U.S. population. They estimated the per capita costs for those diagnosed with asthma to be US$1,254 in 1994 dollars,27 or about US$2,164 in 2007 dollars. Included in their cost estimates were direct medical costs including pre- scriptions, physician and emergency room visits, and hospitalization costs. In addition, indirect costs such as housekeeping, lost work and school days (e.g., the value of care- giver time for students who miss school), restricted activity, and death were included. Excluded, however, were the costs associated with any increased mortality due to asthma. A study by Sullivan et al. (2002) of children with asthma, aged 5 to 11, found similar cost estimates of approximately US$1,200 annually in 1995. Thus, there does not appear to be a significant age difference in costs when averaged over the entire population. Over an estimated life expectancy of 78.2 years (U.S. Department of Health and Human Services, 2007), total costs are thus US$170,000. Discounted to birth at a 2% rate, total lifetime costs are estimated to be US$85,000.
Excluded from the above estimates, however, is the lost quality of life to asthmat- ics. A U.S. Environmental Protection Agency (1999, p. 97) report of WTP studies estimated that the lifetime value of avoiding an asthmatic patient was US$25,000 in 1990 dollars (based on a 5% discount rate). Adjusting to 2007 dollars and the 2% dis- count rate used herein, the present value of the WTP to avoid becoming an asthmatic is estimated to be US$84,000.
Cohen et al. 417
It would not be appropriate to add these WTP estimates to the direct and indirect cost estimates above, since many of these costs are borne by the asthmatics themselves and presumably would be part of their own valuation calculus. Thus, to add them together would double count to the extent costs are borne by asthmatics. However, we estimate that only about 30% of the tangible costs of asthma are born by patients— with the remainder paid by insurance or government health care agencies. Thus, about US$60,000 (70% of US$85,000) of the cost of being an asthmatic is borne by others and could be added to the WTP figure.28 Thus, we estimate the total cost of being an asthmatic to be US$144,000.
Diabetes. A recent study by the American Association of Clinical Endocrinologists (2007) estimated the annual health care costs for individuals diagnosed with diabetes (age 20 or older) were US$8,039 compared to US$2,848 for individuals without dia- betes (2006 dollars). Thus, annual costs are US$5,191 higher for diabetes patients. The data for this study are based on the 2000-2004 Medical Expenditure Panel Survey. However, because of copayments and deductibles, individuals with diabetes pay only US$888 more in health care costs than those who do not have this disease. These figures are direct costs only and exclude the value of caregiver time and any nonmedical expenditure; thus, the actual costs are likely to be higher. In an earlier study using similar data, the American Diabetes Association (1998) estimated the average diabetes patient aged 18 to 64 lost 8.3 days of work compared to 1.7 days for people without diabetes. Thus, adult diabetes patients are estimated to lose 6.6 days annually. At an average hourly wage rate of US$17.41 plus 25.8% fringe benefits, this loss is estimated to be US$1,160 annually (US$920 in wages and US$240 benefits).29
In a study of high-risk individuals for Type 1 diabetes, respondents on average were willing to pay US$2,380 over 3 years for a 30% reduction in risk and US$4,400 for a 50% reduction (Johnson et al., 2006, p. 1354). These amounts were based on a “mini- mal” intervention program requiring minimal diet restrictions, only 3-hr weekly exercise, no medication, and the like (lower WTP values were found for more demand- ing treatment programs). Thus, this appears to be a reasonable (although conservative) approximation of the WTP to reduce the risk of Type 1 diabetes. To move from 50% risk reduction to 100% reduction would likely more than double the WTP because of risk aversion. However, due to wealth effects, it might actually be less than double. For current purposes, we double the amount, to US$8,800 over 3 years, US$2,900 per year in 2004 dollars or US$3,170 in 2007 dollars.
To combine these figures, we first exclude any out-of-pocket medical costs, wage losses, and benefits from the individual WTP estimates to arrive at an estimate of intangible costs, as individuals presumably would factor in these private benefits when thinking about their WTP. Thus, intangible costs are US$1,122 (US$3,170 – US$1,160 – US$888). Tangible costs of US$6,350 (US$5,190 + US$1,160) are added to intangible costs to arrive at total annual costs of diabetes of US$7,472 in 2007 dol- lars. Assuming diabetes begins at approximately age 18, the present value of costs as of the date of birth is estimated to be US$153,000 and would be even higher if the onset of diabetes occurs earlier in the life course (Koopman, Mainous, Diaz, & Geesy,
418 Criminal Justice Policy Review 21(4)
2005). About 85% of these costs are tangible, whereas the remaining 15% are intan- gible costs of pain, suffering, and lost quality of life.
CHD. Russell, Huse, Drowns, Hamel, and Hartz (1998) estimated the direct medical costs associated with CHD over 1-, 5-, and 10-year periods following diagnosis. Based on an estimated 616,900 cases, 1st-year treatment costs were estimated to be US$5.54 billion—or US$8,753 per person (all costs are in 1995 dollars). Cumulative 5-year costs were estimated to be US$9.2 billion; hence a total of US$3.66 billion over years 2 through 5—or US$1,483 per person annually in years 2 through 5. Cumulative 10-year costs were estimated to be US$12.5 billion, which implies an annual cost per person of US$1,070 in years 6 through 10. Note that these costs appear to be lower than full costs, as Russell et al. use wholesale prices for prescription drugs and the amount paid (not charged) for physician and hospital bills. They also ignore all indirect costs (including wage losses) associated with CHD. Russell et al. also note that heart disease is rare in individuals below age 35.
A study commissioned by Environmental Protection Agency in 1986 examined actual annual expenditures by heart disease patients as well as their WTP to avoid angina episodes (Chestnut et al., 1988). The study involved 50 men in California who were screened on the basis of prior chest pain and physician diagnosis of angina pec- toris. All of the patients had experienced angina symptoms for at least 2 years, 68% had previously suffered from heart attacks and on average suffered from about one episode weekly (Chestnut et al., 1988, Table 3.2-1). Average costs included US$256 in out-of-pocket medical costs and US$4,523 medical costs paid for by others; lost wages of US$9,581; and US$903 in defensive expenditures such as hiring others to do lawn maintenance.30 Adding 25.8% fringe benefits to lost wages to approximate the value of lost productivity increases the wage loss to US$12,052. In addition, respondents reportedly would be willing to pay an additional US$50 to US$100 monthly to avoid an additional angina episode. Given the fact that the average respondent experienced four episodes monthly, a conservative estimate of WTP to avoid angina would be US$200 monthly, or US$2,400 annually (it should be noted that the average income in this sample was US$22,021). Combined, the total annual cost for these patients is thus US$20,134 (US$256 + US$4,523 + US$12,052 + US$903 + US$2,400). Tangible costs (US$17,734) represent 88.1%, whereas intangible costs (US$2,400) are 11.9% of this amount.
Although the Chestnut et al. (1988) study is not based on a representative sample of all CHD patients (as a large majority had already suffered heart attacks), it does provide a basis for estimating wage losses and intangible costs associated with diabetes by assuming that the ratio of these losses to medical costs are similar to those found in Russell et al. (1998). In the Chestnut et al. study, medical costs represent 23.4% of total costs. Moreover, direct medical costs in the two studies are not dramatically different (US$4,523 annually in Chestnut et al., compared to US$8,753 in year 1 and US$1,483 in years 2 through 5 in Russell et al., 1998). For current purposes, we have thus used the baseline medical costs in Russell et al. (assuming that CHD first begins at age 35) and multiplied by 4.2 (1/.234) to arrive at an estimate of the total social costs of CHD.
Cohen et al. 419
Total lifetime costs in 2007 dollars are estimated to be US$516,000. Discounted to present value as of birth, total costs are US$127,000.
Low birth weight. Almond, Chay, and Lee (2005, p. 1062) estimated the additional hospital costs associated with a low-birth-weight baby. Using fixed effects regression analysis comparing twin populations, they estimated that a baby weighing between 2,000 and 2,500 g (5 lbs., 8 oz.) at birth will cost US$604 (2000 dollars) more in hos- pital costs than an average baby born in excess of 2,500 g, whereas a baby weighing between 800 and 1,000 g costs US$36,846 more. Based on the frequency of birth weight categories in the population, the weighted average cost for an underweight baby is US$8,300 or US$10,300 in 2007 dollars. These costs are limited to hospitalization and do not include any long-term costs associated with lower quality of life, lower earn- ings, or long-term health effects. To the extent any of the effects of low birth weight carry over into later years, these costs would have to be included.
Teen Pregnancy Teen pregnancy may impose costs on society through both the parents and the child. Teen mothers might be less likely to complete high school or college, have reduced productivity (and hence earn less and pay less taxes), and require more social services paid for by government or nonprofit agencies. They might also have a higher risk of medical complications during birth. The children of teen parents might also have higher medical costs, social service agency costs (including foster care), reduced edu- cational attainment and quality of life (e.g., they have a higher rate of child abuse and neglect), and be at higher risk of becoming juvenile and/or adult offenders and sub- stance abusers.
Hoffman (2006) estimated the costs imposed directly on taxpayers from teen preg- nancy to be US$9.1 billion—about US$1,430 per child born. He includes health care costs to mother and child, child welfare and public assistance programs as well as an estimate of the cost of incarceration of sons of teen mothers who subsequently become criminal offenders. However, these estimates are not based on social costs. For exam- ple, although we would consider the entire value of lost productivity due to dropping out of high school, Hoffman (2006) would only count tax revenue losses. Similarly, he would only count government-sponsored social services, excluding nonprofit agencies and family caregivers. Although these two examples suggest that Hoffman underesti- mates the costs of teen pregnancy, in some cases he overestimates them. For example, to the extent teen mothers rely on public assistance for medical care as opposed to paying for it privately or through insurance, Hoffman counts these costs as additional public expenditures. For our purposes, they would simply be a transfer from the private to public sector and thus would not represent social costs. However, administrative costs associated with making these transfer payments are considered a social cost.
Maynard and Hoffman (2008, p. 382, Table 10.5) estimate annual social costs per teen parent to be US$5,502 in 2004 dollars, which includes any additional medical costs and lost earnings to parents and child, additional administrative costs associated
420 Criminal Justice Policy Review 21(4)
with public transfer payments, foster care, and incarceration costs—all discounted to present value at a rate of 5% as of birth. However, Maynard and Hoffman do not con- vert these figures into a lifetime cost. Below, we review the studies included by Maynard and Hoffman and elsewhere to arrive at our own, more comprehensive esti- mates and provide lifetime costs for teen pregnancy.
Medical costs. Wolfe and Rivers (2008) conclude that although children of teen mothers generally cost the public more in medical expenditures during their childhood years, this cost is largely a shift from the private to public sectors, and the evidence suggests that children of teen mothers are (on average) as healthy as those of older mothers. Moreover, in multivariate regression analyses controlling for other determi- nants of child’s health, they find that children of teen mothers actually have lower acute health conditions and higher but statistically insignificant differences in chronic conditions (Wolfe & Rivers, 2008, p. 238). In fact, they conclude that overall, children of teen mothers have lower total medical expenses than children of older mothers (p. 247). However, teen mothers themselves apparently have slightly higher costs. Maynard and Hoffman (2008, p. 396, Table A.10.2) estimate a combined net medical cost increase of US$42 annually in 2004 dollars. In 2007 dollars, the present value of these medical costs over a lifetime is US$1,800.
Child abuse and neglect. Goerge, Harden, and Lee (2008) compare reported child abuse and neglect rates in the state of Illinois between 1982 and 1998. Controlling for demographic characteristics, they find that children of teen mothers are at significantly higher risk of being the victim of child abuse or neglect. For example, they find a 39% higher rate for those born to 18- or 19-year-old mothers compared to those born to 20-or 21-year-old mothers (p. 277), with an overall base rate of substantiated child abuse or neglect of 53.4 per 1,000. Even higher rates are found for younger mothers (81% higher for age 16-17 and 132% higher for age 15 and below). Based on the dis- tribution of births by age of mother, we estimated the underlying rate (adjusted for demographics) to be 2.6% for children born to mothers aged 20+ and 7.0% for children born to teen mothers. Thus, an estimated 4.4% of all children born to teen mothers (7.0%-2.6%) are child abuse or neglect victims who would not have been victims if they had not been born to teen mothers. With an estimated cost of US$200,000 per child abuse case, this would imply an additional cost associated with teen pregnancy of about US$8,800 (US$200,000 × .044).
In addition, Maynard and Hoffman (2008, p. 397, Table A10.2) estimated the cost of foster care (which is primarily although not exclusively associated with cases of child abuse or neglect) to be US$395 annually per teen birth. Assuming foster care ends at age 18, the present value of foster care costs per teen birth are estimated to be US$6,600.31
High school dropout. In an analysis of teen mothers from the 1960s, Haveman, Wolfe, and Peterson (1997) found no educational achievement loss to children of teen mothers after controlling for other demographic factors. However, Hoffman and Scher (2008a) apply the same methodology to a more recent cohort of children born to teen mothers between 1970 and 1981, and after controlling for other demographic characteristics, they estimate a 7 percentage point reduction in the high school graduation rate of
Cohen et al. 421
children born to teen mothers (p. 347). Based on Cohen and Piquero’s (2009) estimate of US$300,000 to US$450,000 for the value of completing a high school education, teen pregnancy increases the expected present value cost of reduced educational achievement of the child by US$21,000 to US$32,000.
Crime and delinquency. Hoffman and Scher (2008b) estimate the probability that a child born to a teen mother will report ever being incarcerated, based on the NLSY79 (following up through their late 30s and early 40s). They find a 10.6% higher incarcera- tion rate from a base rate of 13.92%. Thus, an additional 1.5% of children (10.6% × 13.92%) born to teen mothers are estimated to be incarcerated at some point during their life who would not otherwise be incarcerated if they had not been born to a teen mother. Based on this analysis, Maynard and Hoffman (2008, p. 397) estimate the cost of incarceration per teen child to be US$364 annually in 2004 dollars.
Of course, incarceration is not the only cost associated with criminal offending. In addition to the cost of the criminal justice system (police, judges, etc.), crimes impose their largest cost on victims in the form of medical costs, lost wages, pain and suffering, and reduced quality of life. A more comprehensive approach would be to estimate the cost of crimes—in addition to criminal justice–related costs such as the cost of incar- ceration. This is the approach taken in Cohen and Piquero (2009).
Since Cohen and Piquero do not report incarceration rates, it is not possible to directly estimate the additional crime cost imposed by children of teen mothers. At one extreme, if they all were to become career criminals (imposing present value life- time costs of US$2.1 to US$3.7 million according to Cohen and Piquero, Table 6), this would represent a cost associated with teen mothers of US$31,500 to US$55,500. At the other extreme, Cohen and Piquero also estimate the present value of lifetime costs for all individuals who have only one police contact through age 26 to range from US$64,000 to US$240,000, which would imply an expected future cost of only US$1,000 to US$3,600. A more realistic assumption is that at a minimum, the average offender has two or more police contacts (as all of the offenders in the Hoffman and Scher (2008a, 2008b) study were observed during a period of incarceration), with costs ranging from US$1.1 to US$1.6 million. Applying the 1.5% increased rate of offending behavior to this figure results in an estimated expected present value cost of US$16,000 to US$24,000.
Earnings of parents. In a statistical analysis of the evidence comparing teen mothers to similar women whose teen pregnancies ended in miscarriage, Hotz, McElroy, and Sanders (2008) conclude that the effect of teen pregnancy on the lifetime earnings of the mother is actually positive. Their explanation is that
women who begin motherhood as teens come from less-advantaged back- grounds, are less likely to be successful in school, and, as such, are less likely to end up in occupations that require higher education than are women who post- pone motherhood . . . (f)or such women, concentrating their childbearing at early ages may prove more compatible with their likely labor-market career options than would postponing motherhood. (p. 71)
422 Criminal Justice Policy Review 21(4)
Hoffman (2008) updated this analysis by extending the sample through 2000 and age 36. He finds similar, but smaller impacts—about US$6,000 higher earnings over an entire lifetime (p. 84), compared to about US$40,000 in Hotz et al. (2008). More- over, Hoffman notes that the estimated higher earnings for teen mothers is not statis- tically significant.
Interestingly, the impact of children born to teen mothers does have a significant negative effect on the earnings of the father. Brien and Willis (2008) use the National Longitudinal Survey of Youth (NLSY) data of 14- to 21-year-olds as of 1979, with follow-up through 1992. They conclude that “in general, early entry into fatherhood is associated with lower levels of schooling, lower actual occupational income, and fewer hours worked in the labor market” (p. 148). For example, they estimate the present value of average earnings for White or Hispanic single fathers of children born of teen moth- ers at age 16 was US$195,659, compared to US$215,308 when the mother was aged 20 to 21 or US$221,943 when the mother was 22 or older. These figures are reported for both single and married fathers by age category of mother. Based on the percentage of single versus married fathers by race, we estimate that fathers of children born to teen mothers earn 3.8% less on average than fathers of children born to mothers aged 20 or older. Based on median earnings for men in the population, we estimate the pres- ent value of reduced earnings over their lifetime to be US$62,000 in 2007.32
Summary of Teen Pregnancy Costs Table 7 summarizes the lifetime costs associated with teen pregnancy, including medi- cal costs, reduced productivity for the father and child, higher risk of being a child abuse or neglect victim and foster child, and higher risk of becoming a career criminal. Combined, the present value of these costs is estimated to range between US$116,200 and US$135,200.
Hoffman and Scher (2008a, p. 348) also estimate that daughters of teen mothers have a 3.1% higher likelihood of becoming teen mothers themselves. This intergenera- tional transfer of teen pregnancy can itself be valued. Assuming the average teen pregnancy occurs at age 16, discounting to present value reduces the loss to between US$85,000 and US$98,000. However, as the excess teen birth rate for children of teen mothers is only 3.1%, the present value of the cost of this second generation of teen pregnancy is estimated to be approximately US$2,600 to US$3,000. Combined, the total cost of teen pregnancy is estimated to range between US$118,800 and US$138,200.
Policy Implications and Future Research Directions Within the backdrop of other studies, the present study calculated the lifetime cost of negative outcomes for at-risk youth and parent(s) of at-risk youth. As summarized in Table 8, the largest cost is imposed by the career criminal (US$2.1 to US$3.7 million). Next, the costs associated with both drug abuse and alcohol dependence/abuse are roughly the same—about US$700,000 each. Child abuse and neglect results in approx- imately US$250,000 to US$285,000 in total costs—most of which (US$200,000) is
Cohen et al. 423
the cost of abuse or neglect to the child, with the remaining amount (US$55,000 to US$85,000) being an estimate of the costs imposed by increasing the risk that the abused child will subsequently become a criminal offender. Health-related outcomes range from a low of US$10,300 for the estimated cost of low birth weight, to US$127,000 for coronary heart disease, US$144,000 for asthma, US$187,000 for diabetes, and US$260,000 for smoking. Finally, the cost of teen pregnancy is estimated to range from US$120,000 to US$140,000. Note that as these social ills often overlap (e.g., crime and drug abuse) and are partly causal themselves (e.g., smoking increases the risk of heart disease), these costs cannot simply be added together.
At first glance, a review of the well-designed home visitation, parent training, and daycare school–based early childhood intervention programs that addressed the long- term outcomes for the various social ills described in this research suggested that programs such as these are, for the most part, effective. However, although effective- ness is typically determined and discussed in terms of statistical significance, we proceeded to estimate the “monetary value of saving a high-risk youth” from a lifetime of adverse outcomes such as crime and delinquency, educational attainment, drug and alcohol abuse, smoking, child abuse and neglect, poor health outcomes, and teen preg- nancy via their participation in an effective childhood intervention. Drawing from the most recent cost estimates of some of these social ills provided by Cohen and Piquero (2009) as well as generating new comprehensive estimates for other relevant outcomes, the results from this study demonstrated the substantial costs that are incurred for at-risk children and their parent(s) over the life course. Thus, properly designed pro- grams and policies that focus on early childhood intervention have the potential to produce significant social benefits.
In a perfect world, early child intervention programs would be evaluated on the basis of their long-term effectiveness in reducing these social ills. Once the effective- ness was known, one could compare the cost of the program against the benefits estimated in this study. However, due to the difficulty of obtaining large-scale longi- tudinal data that identify such outcomes, a more likely approach is to ask the reverse question—how effective must a program be before it pays for itself? For example, suppose a program was found to reduce obesity in young children with the expectation that this will reduce the rate of diabetes and coronary heart disease over their lifetime. If the program covered 1,000 children at a cost of US$1 million, one could calculate the break-even point at which the program paid for itself. Given an average cost of approximately US$150,000 per case of heart disease or diabetes, the program could be judged a success if it prevented about seven cases—or about 0.7% of children in the treatment program. Of course, any childhood intervention program will likely target a diverse set of social ills and outcomes. Thus, depending on the program, additional benefits are likely to accrue. What this example makes clear, however, is that well- designed early childhood interventions are very likely to provide significant quantifiable social benefits across a variety of life domains.
Going forward, it will be important to identify and estimate a value for each of the social ills examined in the current study, and in those cases in which interventions point to evidence of a cause-and-effect reduction in later social ills, to calculate the point at
424 Criminal Justice Policy Review 21(4)
which interventions pay for themselves. Such an analysis can then comment more forc- ibly on the level of effectiveness required before interventions pay for themselves. At the same time, this particular analysis, which is reminiscent of a meta-analysis as com- pared to a cost-benefit study, is difficult to carry out because of the large variability in program types, program costs, and the length of interventions as well as a general lack of randomized and well-designed studies providing long-term results on the various social ills. Nevertheless, it remains an important piece of the puzzle in better document- ing the monetary value of early childhood interventions.
Declaration of Conflicting Interests
One or more authors has declared a potential conflict of interest (e.g. a financial relationship with the commercial organizations or products discussed in this article) as follows:
Funding
The authors disclosed that they received the following support for their research and/or author- ship of this article: This research was supported by the Partnership for America’s Economic Success, The Pew Charitable Trusts. The views expressed are those of the authors and not neces- sarily those of the sponsor or any of the author’s respective affiliated organizations.
Notes
1. DeLisi and Gatling (2003) and Welsh et al. (2008) have also examined the costs associated with career criminals and have yielded comparable estimates to those provided by Cohen (1998).
2. It is important to note that Eisner’s (2009) evaluation of a controlled trial of the Triple P program in Sweden yielded no positive impact and several negative results for the program.
3. One may ask whether it makes sense to include programs with nonrandomly assigned con- trol groups. Our intention was to err on the side of inclusion for a more expansive overview of the interventions and, in addition, because setting a more stringent inclusion criteria, that is, only experimental programs, may be too restrictive. At the same time, we did require that the program offer a means of direct comparability of the intervention and nonintervention groups so as to afford a more concrete interpretation of the program. Yet all of the studies we included that had nonrandomly assigned control groups constructed their control groups using a controlled matched sampling design (e.g., Lally et al., 1987; Long et al., 1994). Clearly, experimentally based studies represent the gold standard, but such efforts are the exception and not the rule in the prevention/intervention literature. Of course, readers can choose to focus their attention on only those experimentally based studies. In addition, the programs selected had to also provide long-term results defined as results beyond mere posttest assessments that were assessed sometime (e.g., at least more than 1 year) after the intervention was administered.
4. An anonymous reader correctly noted that this study does not present a systematic review of this literature in a meta-analytic sense; instead, it is only providing “proof of concept,” that is, some early childhood interventions work.
5. It is important that readers do not assume “effectiveness” from the studies reviewed in Tables 2 through 6.
Cohen et al. 425
6. Although we do not want to enter the debate or confuse the reader over the distinction between external and social costs, suffice it to say that this article uses the terms inter- changeably. Technically, some of the estimates in this article might be considered external costs instead of social costs. Cohen (1998) argues that the relevant concept for analysis of crime control programs is that of external costs, as under some definitions, social costs might exclude harms such as stolen property. We do not exclude these “unauthorized” wealth transfers from our cost estimates. See Cohen (2005) for more detail.
7. Unlike “involuntary” property transfers from victims to criminals, these transfers are vol- untarily made by society and thus do not reduce aggregate social welfare. Although this distinction might be subtle, it is important to keep in mind that social welfare transfer pay- ments are not “externalities” imposed by one party on another—they are voluntary govern- mental transfers like many taxes and subsidies. It is also important to realize that we have already included the cost of lost productivity as a social cost that reduces aggregate welfare.
8. The concept of “present value” is fundamental to economics and is relatively easy to understand. A dollar today is worth more than a dollar tomorrow in purchasing power due to inflation. Similarly, a dollar next year is worth less than having a dollar today, since one could just as easily take that dollar and invest it at current interest rates and have more than a dollar next year. Thus, when economists talk about the “present value” of a future income stream, they are simply computing the amount of money today that would be the equiva- lent to the amount needed in future years, after accounting for the fact that (a) prices and wages increase over time and (b) today’s dollars can be invested and interest compounded. Except in rare circumstances, present value is always less than future value. In this article, present value is always less than future value; hence the use of the phrase discounted to present value.
9. All dollars are updated to 2007 using average wage rate changes by year as reported in the Economic Report of the President.
10. OMB recommends a 7% discount rate when a regulation will replace private capital, such as requiring industries to invest in pollution control equipment. However, a 3% rate is gen- erally used for valuing health benefits.
11. Cohen (2009) contains a detailed discussion and comparison of these approaches in the context of the costs of crime.
12. The WTP methodology estimates the economic value of various policies by determining how much individuals are willing to pay for each of them (Cohen, 2009).
13. In theory, “bottom up” estimates could include the other cost components such as fear of crime. However, to date, they have not done so (Cohen, 2009).
14. One exception is Donohue (2009) who explicitly adds the cost of the criminal justice sys- tem to individual WTP for reduced crime—under the assumption that individual respon- dents would not consider the reduced taxes they would pay if criminal justice expenditures were reduced.
15. Many of these noncrime outcomes are also inextricably linked to crime either indirectly or directly.
16. Note that this figure differs from the US$2.6 to US$4.4 million value of saving a “high-risk youth” in Cohen and Piquero (2009), as the latter includes the costs of drug abuse and drop- ping out of high school in addition to the costs of crime.
426 Criminal Justice Policy Review 21(4)
17. Recent evidence by Heckman and LaFontaine (2006) suggests that the lifetime earnings for GED graduates is significantly lower than it is for high school graduates. By including GED in our “high school graduate” pool, our estimates might be overly conservative.
18. Although we do not know the extent to which this overlap occurs, it is likely to be substan- tial as heavy drug abusers are often likely to have six or more police contacts over their lifetime (the definition of career criminal in Cohen & Piquero, 2009).
19. Both Cohen (1999) and Miller, Levy, Cohen, and Cox (2006) suggest that the crime esti- mates in Harwood et al. (1998) underestimate the cost per crime since they ignore intan- gible losses but overestimate costs due to attributions that are not necessarily causal.
20. As noted earlier, we have not included social welfare transfers. In this case, the transfer cost to taxpayers is approximately US$500 per person annually in 1992 dollars. Updating to 2007 dollars and discounting to present value, this represents approximately US$10,000 in present value costs compared to the US$230,000 total social costs estimated for an alcohol abuser.
21. Note that the fringe benefit rate of 40% differs from the 28.5% rate used elsewhere in this article. However, this is appropriate because Harwood et al. (1998) did not include all mon- etary compensation in their base earnings loss calculation. In particular, they counted paid sick time and bonuses as “fringe benefits,” whereas other studies would generally include those as income.
22. Note that Viscusi (1995) includes the cost of smoking-related fires in his estimate of the external costs of smoking.
23. To arrive at this estimate, we have discounted to present value as of age 20 using a 13.9% discount rate, and then further discounted that amount by 2% annually to the date of birth. Note that using a 2% discount rate throughout the entire period instead of the 13.9% rate would increase the present value of these lifetime benefits as of the date of birth to approxi- mately US$1.1 million, or US$740,000 at a 2% discount rate. An interesting question left for others to assess is what the appropriate discount rate is for estimating the present value of such benefits. Using a social discount rate of 2% to 3% would essentially imply that a smoker’s 13.9% discount rate is irrational or ill-informed and that a paternalistic govern- ment wishes to substitute a lower discount rate. This might be appropriate in the context of young children.
24. The age distribution of child abuse victims is based on a weighted average of physical and sexual abuse victims, as reported by child welfare agencies (see Finkelhor & Ormrod, 2001, Table 1).
25. It is also important to note that this is an “excess” or “marginal” rate of child abuse. Hence, it might be the case that the rate of offending among child abuse victims is higher, but many of those offenders may have done so even in the absence of their abuse.
26. This figure is derived from Cohen and Piquero (2009, Table 6 and accompanying text) as follows: They report a total of 6,157 offenders, 2,827 of whom had only one police contact with average WTP (including offenses not involving police contacts) estimated to range between US$173,140 and US$241,950. In addition, 3,330 offenders had two or more police contacts with total WTP estimated to range between US$1,074,124 and US$1,627,736. The weighted average cost—including all offenders regardless of number of police contacts—is thus estimated to range between US$660,000 and US$1.0 million.
Cohen et al. 427
27. To be consistent, we have added fringe benefits to wage loss estimates in Smith et al. (1997), which increases total costs by approximately US$16, as wage losses account for only a small fraction of their estimated costs of asthma.
28. Note that the majority of costs are health care expenditures, which we estimate 80% are paid by insurance and government programs and only 20% out of pocket. The remaining costs are largely borne by patients or their families, such as lost wages.
29. Average wage rates are taken from the 2008 Economic Report of the President. The value of fringe benefits is estimated from the Bureau of Labor Statistics (2007).
30. Of the 50 participants interviewed, 21 reported defensive expenditures averaging US$2,151. Averaged over the entire 50 participants, annual expenditures are thus US$903.
31. Note that the cost of child abuse in Miller et al. (1996) includes US$1,800 in 1993 dollars for the cost of victim services (primarily foster care). In 2007 dollars, this would be approximately US$2,900. Multiplied by the 4.4% excess rate of teen mothers whose children are victims of abuse or neglect, total costs are estimated to only be US$130. This compares to the US$395 annual cost (2004 dollars) per teen mother estimated by Maynard and Hoffman (2008, p. 397, Table A.10.2). It appears that the key difference between these two estimates is that the Miller et al. estimate is based on the cost per child abuse or neglect victim—irrespective of whether this is reported to authorities and the child ends up in foster care. However, Maynard and Hoffman’s estimates are based on actual foster care cases (and the age of mother).
32. Average earnings for men in the United States were taken from Census data, by age, using a 39.5 year average worklife from age 18 and a net discount rate of 2%. Benefits are esti- mated to be 25.8% of earnings based on BLS data.
References
Almond, D., Chay, K. Y., & Lee, D. S. (2005). The costs of low birth weight. Quarterly Journal of Economics, 3, 1031-1083.
American Association of Clinical Endocrinologists. (2007). State of diabetes complications. Retrieved May 1, 2009 from http://harrisschool.uchicago.edu/News/press-releases/media/ Diabetes%20Complications%20Report_FINAL.PDF
American Diabetes Association. (1998). Economic consequences of diabetes mellitus in the U.S. in 1997. Diabetes Care, 21, 296 -309.
Bernazzani, O., Cote, C., & Tremblay, R. E. (2001). Early parent training to prevent disruptive behavior problems and delinquency in children. ANNALS, 578, 90-103.
Bernazzani, O., & Tremblay, R. E. (2006). Early parent training. In B. C. Welsh & D. P. Far- rington (Eds.), Preventing crime: What works for children, offenders, victims, and places. Dordrecht, Netherlands: Springer.
Bilukha, O., Hahn, R. A., Crosby, A., Fullilove, M. T., Liberman, A., Moscicki, E., et al. (2005). The effectiveness of early childhood home visitation in prevention violence: A systematic review. American Journal of Preventive Medicine, 28, 11-39.
Boisjoli, R., Vitaro, F., Lacourse, É., Barker, E. D., & Tremblay, R. E. (2007). Impact and clini- cal significance of a preventive intervention for disruptive boys: A 15 year follow-up. British Journal of Psychiatry, 191, 415-419.
Brestan, E. V., Eyberg, S. M., Boggs, S. R., & Algina, J. (1997). Parent–child interaction therapy: Parents’ perceptions of untreated siblings. Child and Family Behavior Therapy, 19, 13-28.
428 Criminal Justice Policy Review 21(4)
Brien, M. J., & Willis, R. J. (2008). Costs and consequences for fathers. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed., pp. 119-160). Washington, DC: The Urban Institute Press.
Bureau of Labor Statistics. (2007, September 20). Employer costs for employee compensation— June 2007. Washington, DC: Author.
Center for Disease Control. (2004, September 24). Alcohol-attributable deaths and years of potential life lost—United States, 2001. MMWR Weekly, 53, 866-870. Retrieved October 25, 2009, from http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5337a2.htm
Chestnut, L. G., Colome, S. D., Keller, L. R., Lambert, W. E., Ostro, B., Rowe, R. D., et al. (1988). Heart disease patients’ averting behavior, cost of illness, and willingness to pay to avoid angina episodes (Final Report to Office of Policy Analysis). Washington, DC: U.S. Environmental Protection Agency.
Cohen, M. A. (1998). The monetary value of saving a high risk youth. Journal of Quantitative Criminology, 14, 5-33.
Cohen, M. A. (1999). Alcohol, drugs and crime: Is “crime” really one-third of the problem? Commentary on “The economic costs of alcohol and drug abuse in the United States, 1992,” Report to NIH by Harwood et al. (1998). Addiction, 94, 644-647.
Cohen, M. A. (2000). Measuring the costs and benefits of crime and justice. In Measuring and analysis of crime and justice (NIJ Publication No. 182411, pp. 263-316). Washington, DC: National Institute of Justice. Available at http://www.ncjrs.org/criminal_justice2000/ vol_4/04f.pdf
Cohen, M. A. (2005). The costs of crime and justice. New York: Routledge. Cohen, M. A. (2009). Valuing crime control benefits using stated preference approaches. In
T. Dunworth (Ed.), Cost and benefits of crime. Washington, DC: Urban Institute Press. Cohen, M. A., & Piquero, A. R. (2009). New evidence on the monetary value of saving a high-
risk youth. Journal of Quantitative Criminology, 25, 25-49. Cohen, M. A., Rust, R., Steen, S., & Tidd, S. (2004). Willingness-to-pay for crime control pro-
grams. Criminology, 42, 89-109. Cullen, K. J., & Cullen, A. M. (1996). Long-term follow-up of the Busselton six-year controlled
trial of prevention of children’s behavior disorders. Journal of Pediatrics, 29, 136-139. DeLisi, M., & Gatling, J. M. (2003). Who pays for a life of crime? An empirical assessment of
the assorted victimization costs posed by career criminals. Criminal Justice Studies: A Criti- cal Journal of Crime, Law and Society, 16, 283-293.
Donohue, J. J., III. (2009). Assessing the relative benefits of incarceration: The overall change over the previous decades and the benefits on the margin. Do prisons make us safer? The benefits and costs of the prison boom. Retrieved October 25, 2009, from http://works .bepress.com/john_donohue/65
Eisner, M. (2009). No effects in independent prevention trials: Can we reject the cynical view? Journal of Experimental Criminology, 5, 162-183.
Eyberg, S. M., & Durning, P. (1994). Parent–child interaction therapy: Procedures manual. Unpublished manuscript.
Eyberg, S. M., Boggs, S. R., & Algina, J. (1995). New developments in psychosocial, pharma- cological, and combined treatments of conduct disorders in aggressive children. Psycho- pharmacology Bulletin, 31, 83-91.
Cohen et al. 429
Farrington, D. P., & Welsh, B. C. (2003). Family-based prevention of offending: A meta-analysis. Australian and New Zealand Journal of Criminology, 36, 127-151.
Farrington, D. P., & Welsh, B. C. (2007). Saving children from a life of crime: Early risk factors and effective interventions. New York: Oxford University Press.
Finkelhor, D., & Ormrod, R. (2001, May). Child abuse reported to the police: Comparing NIBRS and Child Welfare System data. Juvenile Justice Bulletin. Retrieved October 25, 2009, from http://www.ncjrs.gov/html/ojjdp/jjbul2001_5_1/page3.html
Goerge, R. M., Harden, A., & Lee, B. J. (2008). Consequences of teen childbearing for child abuse, neglect, and foster care placement. In R. A. Maynard & S. Hoffman (Eds.), Kids hav- ing kids (Rev. ed., pp. 257-288). Washington, DC: The Urban Institute Press.
Gold, M. R., Siegel, J. E., Russell, L. B., & Weinstein, M. C. (1996). Cost-effectiveness in health and medicine. New York: Oxford University Press.
Gomby, D. S., Culross, P. L., & Behrman, R. E. (1999). Home visiting: Recent program evaluations—analysis and recommendations. Future of Children, 9, 4-26.
Greenwood, P. W. (2006). Changing lives: Delinquency prevention as crime-control policy. Chicago: University of Chicago Press.
Hartman, R. (1990). One thousand points of light seeking a number: A case study of CBO’s discount rate policy. Journal of Environmental Economics and Management, 18, S3-S7.
Hartunian, N. S., Smart, C. N., & Thompson, M. S. (1980). The incidence and economic costs of cancer, motor vehicle injuries, coronary heart disease, and stroke: A comparative analysis. American Journal of Public Health, 70, 1249-1260.
Harwood, H., Fountain, D., & Livermore, G. (1998). Economic costs of alcohol abuse and alcohol- ism. In M. Galanter (Ed.), Recent developments in alcoholism, volume 14: The consequences of alcoholism (pp. 307-329). New York: Plenum Press.
Haveman, R., Wolfe, B., & Peterson, E. (1997). Children of early childbearers as young adults. In R. A. Maynard (Ed.), Kids having kids: Economic costs and social consequences of teen pregnancy. Washington, DC: The Urban Institute Press.
Hawkins, J. D., Catalano, R. F., Kosterman, R., Abbott, R., & Hill, K. G. (1999). Preventing adolescent health-risk behaviors by strengthening protection during childhood. Archives of Pediatrics and Adolescent Medicine, 153, 226-234.
Hawkins, J. D., Kosterman, R., Catalano, R. F., Hill, K. G., & Abbott, R. D. (2005). Promoting positive adult functioning through social development intervention in childhood: Long-term effects from the Seattle Social Development Project. Archives of Pediatrics and Adolescent Medicine, 159, 25-31.
Heckman, J. J., & LaFontaine, P. (2006). Bias corrected estimates of GED returns. Journal of Labor Economics, 24, 661-700.
Hoffman, S. D. (2006). By the numbers: The public costs of teen childbearing. Washington, DC: The National Campaign to Prevent Teen and Unplanned Pregnancy.
Hoffman, S. D. (2008). Consequences of teen childbearing for mothers. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed., pp. 51-118). Washington, DC: The Urban Institute Press.
Hoffman, S. D., & Scher, L. S. (2008a). Consequences of teen childbearing for the life chances of children, 1979-2002. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed., pp. 342-358). Washington, DC: The Urban Institute Press.
430 Criminal Justice Policy Review 21(4)
Hoffman, S. D., & Scher, L. S. (2008b). Incarceration-related costs of early childbearing: Updated estimates. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed; pp. 95-143.). Washington, DC: The Urban Institute Press.
Hotz, J. V., McElroy, S. W., & Sanders, S. G. (2008). The impacts of teen childbearing on the mother and the consequences of those impacts for government. In R. A. Maynard & S. Hoffman (Eds.) Kids having kids: Revised edition (pp. 51-74 ). Washington, DC: The Urban Institute Press.
Jackson, C., & Dickinson, D. M. (2009). Developing parenting programs to prevent child health risk behaviors: A practice model. Health Education Research. Prepublished August 6, 2009; DOI: 10.1093/her/cyp039
Jeanrenaud, C., Pellegrini, S., & Widmer, G. (2003). The social cost of alcohol abuse in Switzerland. Neuchâtel, Switzerland: University of Neuchâtel.
Johnson, D. L. (2006). Parent–child development center follow-up project: Child behavior problem results. Journal of Primary Prevention, 27, 391-407.
Johnson, D. L., & Breckenridge, J. N. (1982). The Houston parent–child development center and the primary prevention of behavior problems in young children. American Journal of Community Psychology, 10, 305-316.
Johnson, D. L., & Walker, T. (1987). Primary prevention of behavior problems in Mexican- American children. American Journal of Community Psychology, 15, 375-385.
Johnson, F. R., Manjunath, R., Mansfield C. A., Clayton, L. J., Hoerger, T. J., & Zhang, P. (2006). High-risk individuals’ willingness to pay for diabetes risk reduction programs. Dia- betes Care, 29, 1351-1356.
Kitzman, H., Olds, D. L., Henderson, C. R., Hanks, C., Cole, R., Tatelbaum, R., et al. (1997). Effect of prenatal and infancy home visitation by nurses on pregnancy outcomes, childhood injuries, and repeated childbearing: A randomized controlled trial. Journal of the American Medical Association, 278, 644-652.
Kitzman, H., Olds, D. L., Sidora, K., Henderson, C. R., Hanks, C., Cold, R., et al. (2000). Enduring effects of nurse home visitation on maternal life course: A 3-year follow-up of a randomized trial. Journal of the American Medical Association, 283, 1983-1989.
Koopman, R. J., Mainous, A. G., III, Diaz, V. A., & Geesy, M. E. (2005). Changes in age diagnosis of type 2 diabetes mellitus in the United States, 1988 to 2000. Annals of Family Medicine, 3, 60-63.
Lally, J. R., Mangoine, P. L., & Honig, A. S. (1987). The Syracuse University Family Devel- opment Research program: Long-range impact of an early intervention with low-income children and their families. New York: Grant (W.T.) Foundation.
Leung, C., Sanders, M. R., Leung, S., Mak, R., & Lau, J. (2003). An outcome evaluation of the implementation of the Triple P-Positive Parenting Program in Hong Kong. Family Process, 42, 531-544.
Loeber, R., & Farrington, D. P. (1998). Serious and violent juvenile offenders: Risk factors and successful interventions. Thousand Oaks, CA: Sage.
Lonczak, H. S., Abbott, R., Hawkins, J. D., Kosterman, R., & Catalano, R. F. (2008). Effects of the Seattle Social Development Project on sexual behavior, pregnancy, birth, and sexu- ally transmitted disease outcomes by age 21 years. Archives of Pediatrics and Adolescent Medicine, 156, 438-437.
Cohen et al. 431
Long, P., Forehand, R., Wierson, M., & Morgan, A. (1994). Does parent training with young non- compliant children have long-term effects? Behavioral Research and Therapy, 32, 101-107.
Markie-Dadds, C., & Sanders, M. R. (2006). Self-directed triple p (positive parenting program) for mothers with children at-risk of developing. Behavioral and Cognitive Psychotherapy, 34, 259-275.
Masse, L. N., & Barnett, W. S. (2002). A benefit cost analysis of the Abecedarian Early Child- hood Intervention. New Brunswick, NJ: National Institute for Early Education Research.
Maxfield, M. G., Weiler, B., & Widom, C. S. (2000). Comparing self-reports and official records of arrest. Journal of Quantitative Criminology, 16, 87-110.
Maynard, R. A., & Hoffman, S. D. (2008). The costs of adolescent childbearing. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed., pp. 359-402). Washington, DC: The Urban Institute Press.
McCart, M. R., Priester, P. E., Davies, W. H., & Razia, A. (2006). Differential effectiveness of behavioral parent-training and cognitive behavioral therapy for antisocial youth: A meta- analysis. Journal of Abnormal Psychology, 34, 527-543.
McCormick, M., Brooks-Gunn, J., Buka, S. L., Goldman, J., Yu, J., Salganik, M., et al. (2006). Early intervention in low birth weight premature infants: Results at 18 years of age for the Infant Health and Development Program. Pediatrics, 117, 771-780.
McNeil, C. B., Eyberg, S., Eisenstadt, T. H., Newcomb, K., & Funderburk, B. (1991). Parent– child interaction therapy with behavior problem children: Generalization of treatment effects to the school setting. Journal of Clinical Child Psychology, 20, 140-151.
Miller, T. R., Cohen, M. A., & Wiersema, B. (1996). Victim costs and consequences: A new look (National Institute of Justice Research Report, NCJ-155282). Washington, DC: U.S. Department of Justice.
Miller, T. R., Levy, D. T., Cohen, M. A., & Cox, K. L. C. (2006). Costs of alcohol and drug- involved crime. Prevention Science, 7, 333-342.
Moore, M. J., & Viscusi, W. K. (1989). Discounting environmental health risks: New evidence and policy implications. Journal of Environmental Economics and Management, 18, S51-S62.
Morawska, A., & Sanders, M. R. (2006). Self-administered behavioral family intervention for parents of toddlers: Part I. Efficacy. Journal of Consulting and Clinical Psychology, 74, 10-19.
Office of Management and Budget. (2003, September 17). Circular A-4. Available from http:// www.whitehouse.gov/OMB/circulars/
Olds, D. L., Eckenrode, J., Henderson, C. R., Jr., Kitzman, H., Powers, J., Cole, R., et al. (1997). Long-term effects of home visitation on maternal life course and child abuse and neglect: Fifteen-year follow-up of a randomized trial. Journal of the American Medical Association, 278, 637-643.
Olds, D., Henderson, C. R., Cole, R., Eckenrode, J., Kitzman, H., Luckey, D., et al. (1998). Long-term effects of nurse home visitation on children’s criminal and antisocial behavior: 15-year follow-up of a randomized controlled trial. Journal of the American Medical Asso- ciation, 280, 1238-1244.
Olds, D. L., Henderson, C. R., Jr., Chamberlin, R., & Tatelbaum, R. (1986). Preventing child abuse and neglect: A randomized trial of nurse home visitation. Pediatrics, 78, 65-78.
432 Criminal Justice Policy Review 21(4)
Olds, D. L., & Kitzman, H. (1993). Review of research on home visiting for pregnant women and parents of young children. Future of Children, 3, 53-92.
Olds, D. L., Kitzman, H., Cole, R., Robinson, J., Sidora, K., Luckey, D., et al. (2004). Effects of nurse home-visiting on maternal life-course and child development: Age 6 follow-up results of a randomized trial. Pediatrics, 114, 1550-1559.
Olds, D. L., Kitzman, H., Hanks, C., Cole, R., Anson, E., Sidora-Arcoleo, K., et al. (2007). Effects of nurse home visiting on maternal and child functioning: Age 9 follow-up of a ran- domized trial. Pediatrics, 120, e832-e845.
Olds, D. L., Robinson, J., O’Brien, R., Luckey, D., Pettitt, L., Henderson, C. R., Jr., et al. (2002). Home visits by paraprofessionals and by nurses: A randomized, controlled trial. Pediatrics, 110, 486-496.
Olds, D. L., Robinson, J., Pettitt, L., Luckey, D., Holmberg, J., Ng., R. K., et al. (2004). Effects of home visits by paraprofessionals and by nurses: Age 4 follow-up results of a randomized trial. Pediatrics, 114, 1560-1568.
Piquero, A. R., Daigle, L. E., Gibson, C., Piquero, N. L., & Tibbetts, S. G. (2007). Are life-course-persistent offenders at risk for adverse health outcomes? Journal of Research in Crime and Delinquency, 44, 185-207.
Piquero, A. R., Farrington, D. P., Welsh, B., Tremblay, R., & Jennings, W. G. (2009). Effects of early family/parent training programs on antisocial behavior and delinquency. Journal of Experimental Criminology, 5, 83-120.
Reynolds, A. J., Temple, J. A., Robertson, D. L., & Mann, E. A. (2001). Long-term effects of an early childhood intervention on educational achievement and juvenile arrest: A 15-year follow-up of low-income children in public schools. Journal of the American Medical Asso- ciation, 285, 2339-2346.
Reynolds, A. J., Temple, J. A., Ou, S., Robertson, D. L., Mersky, J. P., Topitzes, J. W., et al. (2007). Effects of a school-based, early childhood intervention on adult health and well- being. Archives of Pediatrics and Adolescent Medicine, 161, 730-739.
Russell, M. W., Huse, D. M., Drowns, S., Hamel, E. C., & Hartz, S. C. (1998). Direct medical costs of coronary artery disease in the United States. American Journal of Cardiology, 81, 1110-1115.
Sanders, M. R., Markie-Dadds, C., Tully, L. A., & Bor, W. (2000). The Triple P-Positive Parent- ing Program: A comparison of enhanced, standard, and self-directed behavioral family inter- vention for parents of children with early onset conduct problems. Journal of Consulting and Clinical Psychology, 68, 624-640.
Sanders, M. R., Montgomery, D. T., & Brechtman-Toussaint, M. L. (2000). The mass media and the prevention of child behaviour problems: The evaluation of a television series to promote positive outcome for parents and their children. Journal of Child Psychology and Psychiatry, 41, 939-948.
Sanders, M. R., Pidgeon, A. M., Gravestock, F., Connors, M. D., Brown, S., & Young, R. W. (2004). Does parental attributional retraining and anger management enhance the effects of the Triple P-Positive Program with parents at-risk of child maltreatment? Behavior Therapy, 35, 513-535.
Scharff, R. L., & Viscusi, W. K. (2008). Heterogeneous rates of time preference and the decision to smoke (Vanderbilt University Law School Law and Economics Working Paper Number 08-31). Retrieved October 25, 2009, from www.ssrn.com/abstract_id=1319389
Cohen et al. 433
Schuhmann, E. M., Foote, R. C., Eyberg, S. M., Boggs, S. R., & Algina, J. (1998). Efficacy of parent–child interaction therapy: Interim report of a randomized trial with short-term maintenance. Journal of Clinical Child Psychology, 27, 34-45.
Schweinhart, L. J. (2007). Crime prevention by the High/Scope Perry Preschool program. Paper presented at the 2007 American Society of Criminology Conference, Atlanta, GA.
Schweinhart, L. J., Berrueta-Clement, J. R., Barnett, W. S., Epstein, A. S., & Weikart, D. P. (1985). Effects of the Perry Preschool program on youths through age 19: A summary. Top- ics in Early Childhood Special Education, 5, 26-35.
Serketich, W. J., & Dumas, J. E. (1996). The effectiveness of behavioral parent training to modify antisocial behavior in children: A meta-analysis. Behavior Therapy, 27, 171-186.
Smith, D. H., Malone, D. C., Lawson, K. A., Okamoto, L. J., Battista, C., & Saunders, W. B. (1997). A national estimate of the economic costs of asthma. American Journal of Respira- tory Critical Care, 156, 787-793.
St. Pierre, R. G., & Layzer, J. I. (1999). Using home visits for multiple purposes: The compre- hensive child development program. Future of Children, 9, 134-151.
St. Pierre, R. G., Layzer, J. I., Goodson, B. D., & Bernstein, L. S. (1997). National impact evalu- ation of the Comprehensive Child Development Program (Prepared for Administration of Children, Youth and Families, U.S. Department of Health and Human Services). Cambridge, MA: Abt Associates.
Sullivan, S., Weiss, K., Lynn, H., Mitchell, H., Kattan, M., Gergen, P., et al. (2002). The cost- effectiveness of an inner-city asthma intervention for children. Journal of Allergy and Clini- cal Immunology, 110, 576-581.
Temple, J. A., & Reynolds, A. J. (2007). Benefits and costs of investments in preschool educa- tion: Evidence from the child-parent centers and related programs. Economics of Education Review, 26, 126-144.
Tremblay, R. E., Pagani-Kurtz, L., Masse, L. C., Vitaro, F., & Phil, R. O. (1995). A bimodal pre- ventive intervention for disruptive kindergarten boys: Its impact through mid-adolescence. Journal of Consulting and Clinical Psychology, 63, 560-568.
U.S. Department of Health and Human Services. (2007, December 28). United States life tables, 2004. National Vital Statistics Reports, 56(9). Retrieved October 24, 2009, from http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_09.pdf
U.S. Environmental Protection Agency. (1999). The benefits and costs of the Clean Air Act, 1990 to 2010. Washington, DC: United States Environmental Protection Agency.
Viscusi, W. K. (1995). Cigarette taxation and the social consequences of smoking. Tax Policy and the Economy, 9, 51-101.
Viscusi, W. K., & Hersch, J. (2008). The mortality cost to smokers. Journal of Health Economics, 27, 943-958.
Webster-Stratton, C. (1982). Teaching mothers through videotape modeling to change their children’s behavior. Journal of Pediatric Psychology, 7, 279-294.
Webster-Stratton, C. (1984). Randomized trial of two parent-training programs for families with conduct-disordered children. Journal of Consulting and Clinical Psychology, 52, 666-678.
Webster-Stratton, C. (1990). Enhancing the effectiveness of self-administered videotape parent training for families with conduct-problem children. Journal of Abnormal Child Psychology, 18, 479-492.
434 Criminal Justice Policy Review 21(4)
Webster-Stratton, C. (1992). Individually administered videotape parent training: Who ben- efits? Cognitive Therapy and Research, 16, 31-35.
Webster-Stratton, C. (1998). Preventing conduct problems in Head Start children: Strengthen- ing parent competencies. Journal of Consulting and Clinical Psychology, 66, 715-730.
Welsh, B. C., Loeber, R., Stevens, B. R., Stouthamer-Loeber, M., Cohen, M. A., & Farrington, D. P. (2008). Costs of juvenile crime in urban areas: A longitudinal perspective. Youth Violence and Juvenile Justice, 6, 3-27.
Widom, C. S., & Maxfield, M. G. (2001). An update on the “cycle of violence” (Research in brief). Washington, DC: National Institute of Justice.
Wolfe, B., & Rivers, E. M. (2008). Children’s health and health care. In R. A. Maynard & S. Hoffman (Eds.), Kids having kids (Rev. ed., pp. 221-256). Washington, DC: The Urban Institute Press.
Zangwill, W. M. (1983). An evaluation of a parent training program. Child and Family Behavior Therapy, 5, 1-16.
Bios
Mark A. Cohen, PhD, is a professor of management and law at the Owen Graduate School of Management, Vanderbilt University, and vice president for research and senior fellow at Resources for the Future. Previously, he served as a staff economist at the U.S. Environmental Protection Agency, the U.S. Federal Trade Commission, and the U.S. Sentencing Commission. He has published widely in the fields of criminology, economics, and public policy.
Alex R. Piquero, PhD, is a professor in the College of Criminology and Criminal Justice at Florida State University, and coeditor of the Journal of Quantitative Criminology. His research interests include criminal careers, criminological theory, and quantitative research methods, and he is a past recipient of the American Society of Criminology’s Young Scholar and E-Mail Mentor of the Year Award as well as the University of Florida’s College of Arts & Sciences Teacher of the Year Award.
Wesley G. Jennings, PhD, is an assistant professor in the Department of Justice Administration at the University of Louisville and holds a PhD in criminology from the University of Florida. His research interests expand over a number of criminological, psychological, and criminal justice-related areas, but his primary focus is applying semiparametric group-based modeling techniques to examine similarities/differences among groups that demonstrate distinct behav- ioral trajectories over time. In addition, another area of his research involves investigating the effect of early childhood interventions on reducing crime and delinquency. Some of his recent publications have appeared in the Journal of Youth and Adolescence, Journal of Interpersonal Violence, Criminology & Public Policy, Justice Quarterly, and the Journal of Experimental Criminology. He is also a recent recipient of the 2009 Outstanding Article Award for the Ameri- can Journal of Criminal Justice and a 2009 recipient of the Academy of Criminal Justice Sciences and SAGE Junior Faculty Professional Development Teaching Award.
