The Taxpayer Costs of
Divorce and Unwed Childbearing
First-Ever Estimates for the Nation
and All Fifty States
A Report to the Nation
Benjamin Scafidi, Principal Investigator
Institute for American Values
Institute for Marriage and Public Policy
Georgia Family Council
Families Northwest
On the cover: Man and Woman Splitting Dollar by
Todd Davidson, Stock Illustration RF, Getty
Images.
© 2008, Georgia Family Council and Institute for
American Values. No reproduction of the materi-
als contained herein is permitted without the
written permission of the Institute for American
Values.
ISBN: 1-931764-14-X
Institute for American Values
1841 Broadway, Suite 211
New York, New York 10023
Tel: (212) 246-3942
Fax: (212) 541-6665
Website: www.americanvalues.org
Email: [email protected]
M
OST OF THE PUBLIC DEBATE o
ver marriage focuses on the role of marriage as
a social, moral, or religious institution. But marriage is also an economic
institution, a powerful creator of human and social capital. Increases in
divorce and unwed childbearing have broad economic implications, including
larger expenditures for the federal and state governments. This is the first-ever
report that attempts to measure the taxpayer costs of family fragmentation for
U.S. taxpayers in all fifty states. Among its findings: Even programs that result in
very small decreases in divorce and unwed childbearing could yield big savings
for taxpayers.
The report’s principal investigator is Benjamin Scafidi, an economist in the
J. Whitney Bunting School of Business at Georgia College & State University. The
co-sponsoring organizations are the Institute for American Values, the Institute for
Marriage and Public Policy, Georgia Family Council, and Families Northwest.
The co-sponsoring organizations are grateful to Chuck Stetson and Mr. and Mrs.
John Fetz for their generous financial support of the project. The principal investi-
gator is grateful to Deanie Waddell for her expert research assistance.
Page 3
Project Advisors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
I. Why Should Government Care about Marriage? . . . . . . . . . . . . . . . . . . . . . .7
II. How Might Marriage Affect Taxpayers? Empirical Literature Review . . . . . . .9
How Much Does Marriage Reduce Poverty? . . . . . . . . . . . . . . . . . . . . . . .10
Does Family Fragmentation Increase Crime? . . . . . . . . . . . . . . . . . . . . . .11
III. Is the Methodology Used in This Estimate Reasonable? . . . . . . . . . . . . . . . .12
What Costs Are Associated with Means-Tested Government Programs? . . .13
What Costs Are Associated with the Justice System? . . . . . . . . . . . . . . . . .16
How Are Foregone Tax Revenues Estimated? . . . . . . . . . . . . . . . . . . . . . .16
IV. What Is the Total Estimated Cost of Family Fragmentation? . . . . . . . . . . . . .17
V. What Are the Policy Implications? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .20
Appendix A: Testing the Analysis: Is the Estimate of $112 Billion
Too High or Too Low? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
Appendix B: Explaining the Methodology for State-Specific Costs . . . . . . . . . . .31
Tables
Table 1: U.S. Children Residing in Two-Parent Families . . . . . . . . . . . . . . .7
Table 2: Percent of U.S. Children in a Single-Parent Household that Has... .7
Table 3: Persons and Children Lifted out of Poverty via Marriage . . . . . . .14
Table 4: Household Income and Usage of Food Stamps . . . . . . . . . . . . . .14
Table 5: Household Income and Usage of Cash Assistance . . . . . . . . . . . .15
Table 6: Household Income and Usage of Medicaid . . . . . . . . . . . . . . . . .15
Table 7: Estimated Costs of Family Fragmentation for U.S. Taxpayers . . . .18
Table A.1: Sub-Calculations of State and Federal Taxpayer Costs . . . . . . .32
Notes to Table A.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
Table A.2: Sub-Calculations for EITC and Justice System Estimates . . . . . .35
Table A.3: Total Poverty and Family Structure by State . . . . . . . . . . . . . . .36
Table A.4: Child Poverty and Family Structure by State . . . . . . . . . . . . . .37
Table A.5: Estimates of State and Local Taxpayer Costs of
Family Fragmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
Endnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39
Contents
The Taxpayer Costs of
Divorce and Unwed Childbearing
First-Ever Estimates for the Nation and All Fifty States
Page 4
Project Advisors
Project advisors provided expert review but are not authors of the report. Affiliations
are listed for identification purposes only. Any errors or omissions in this report are
the responsibility of the principal investigator and not of the project advisors.
James Alm
Andrew Young School of Policy Studies at Georgia State University
Obie Clayton
Morehouse College
Ron Haskins
The Brookings Institution
Brett Katzman
Kennesaw State University
Robert Lerman
Urban Institute
Theodora Ooms
Center for Law and Social Policy
Roger Tutterow
Mercer University
Matt Weidinger
U.S. House Ways and Means Committee
W. Bradford Wilcox
University of Virginia
About Benjamin Scafidi
Ben Scafidi is an associate professor in the J. Whitney Bunting School of Business
at Georgia College & State University. His research has focused on education and
urban policy. Previously he served as the Education Policy Advisor for Georgia
Governor Sonny Perdue and served on the staff of both of Georgia Governor Roy
Barnes’ Education Reform Study Commissions. He received his Ph.D. in Economics
from the University of Virginia and his bachelor’s degree in Economics from the
University of Notre Dame. Ben was born and raised in Richmond, Virginia. Ben and
Lori Scafidi and their four children reside in Milledgeville, Georgia.
Page 5
Executive Summary
T
HIS STUDY PROVIDES THE FIRST RIGOROUS ESTIMATE of the costs to U.S. taxpayers
of high rates of divorce and unmarried childbearing both at the national and
state levels.
Why should legislators and policymakers care about marriage? Public debate on
marriage in this country has focused on the “social costs” of family fragmentation
(that is, divorce and unwed childbearing), and research suggests that these are
indeed extensive. But marriage is more than a moral or social institution; it is also
an economic one, a generator of social and human capital, especially when it
comes to children.
Research on family structure suggests a variety of mechanisms, or processes,
through which marriage may reduce the need for costly social programs. In this
study, we adopt the simplifying and extremely cautious assumption that all of the
taxpayer costs of divorce and unmarried childbearing stem from the effects that
family fragmentation has on poverty, a causal mechanism that is well-accepted and
has been reasonably well-quantified in the literature.
Based on the methodology, we estimate that family fragmentation costs U.S. tax-
payers at least $112 billion each and every year, or more than $1 trillion each
decade. In appendix B, we also offer estimates for the costs of family fragmenta-
tion for each state.
These costs arise from increased taxpayer expenditures for antipoverty, criminal jus-
tice, and education programs, and through lower levels of taxes paid by individuals
who, as adults, earn less because of reduced opportunities as a result of having been
more likely to grow up in poverty.
The $112 billion figure represents a “lower-bound” or minimum estimate. Given the
cautious assumptions used throughout this analysis, we can be confident that cur-
rent high rates of family fragmentation cost taxpayers at least $112 billion per year.
The estimate of $112 billion per year is the total figure incurred at the federal, state,
and local levels. Of these taxpayer costs, $70.1 billion are at the federal level, $33.3
billion are at the state level, and $8.5 billion are at the local level. Taxpayers in
California incur the highest state and local costs at $4.8 billion, while taxpayers in
Wyoming have the lowest state and local costs at $61 million.
If, as research suggests is likely, marriage has additional benefits to children, adults,
and communities, and if those benefits are in areas other than increased income lev-
els, then the actual taxpayer costs of divorce and unwed childbearing are likely
much higher.
Page 6
How should policymakers, state legislators, and others respond to the large taxpayer
costs of family fragmentation? We note that even very small increases in stable mar-
riage rates as a result of government programs or community efforts to strengthen
marriage would result in very large savings for taxpayers. If the federal marriage
initiative, for example, succeeds in reducing family fragmentation by just 1 percent,
U.S. taxpayers will save an estimated $1.1 billion each and every year.
Because of the modest price tags associated with most federal and state marriage-
strengthening programs, and the large taxpayer costs associated with divorce and
unwed childbearing, even modest success rates would be cost-effective. Texas, for
example, recently appropriated $15 million over two years for marriage education
and other programs to increase stable marriage rates. If this program succeeds in
increasing stably married families by just three-tenths of 1 percent, it will be cost-
effective in its returns to Texas taxpayers.
This report is organized as follows: Section I explains why policymakers may have
an interest in supporting marriage. Sections II and III explain the methods used to
estimate the taxpayer cost of family fragmentation by using evidence about the rela-
tionship between family breakdown and poverty. Section IV reveals the national
estimate of the taxpayer cost. Estimated costs for individual states are found in
appendix B.
Finally, a note to social scientists: Few structural estimates exist of the relationships
needed to estimate the taxpayer costs of family fragmentation. Therefore, we have
used indirect estimates based on the assumption that marriage has no independent
effects on adults or children other than the effect of marriage on poverty.
Page 7
I. Why Should Government Care about Marriage?
O
VER THE LAST FORTY YEARS, marriage has become less common and more frag-
ile, and the proportion of children raised outside intact marriages has
increased dramatically. Between 1970 and 2005, the proportion of children
living with two married parents dropped from 85 percent to 68 percent, according
to Census data. About three-quarters of children living with a single parent live with
a single mother.
These important changes in family structure stem from two fundamental changes in
U.S. residents’ behavior regarding marriage: increases in unmarried childbearing
and high rates of divorce.
1
More than a third of all U.S. children are now born out-
side of wedlock, including 25 percent of non-Hispanic white babies, 46 percent of
Hispanic babies, and 69 percent of African American babies.
2
In 2004, almost 1.5
million babies were born to unmarried mothers.
3
Divorce rates, by contrast, after
increasing in the 1960s and 1970s, appear to have declined modestly in recent
years. The small decline in divorce after 1980, however, seems to have been offset
by increases in unwed childbearing, as the percentage of children living with one
parent increased steadily between 1970 and 1998 with only a small drop after 1998.
Overall, divorce rates remain high relative to the period before 1970. Today’s young
adults in their prime childbearing years are less likely to get married, and many
more U.S. children each year are
born to unmarried mothers. Should
U.S. taxpayers be concerned about
these increases in family fragmen-
tation, and if so, why?
Public debate on marriage in this
country has focused on the “social
costs” of increases in divorce and
unmarried childbearing. Research
suggests that the social costs are
indeed extensive. When parents
part, or fail to marry, their children
seem to suffer from increased risks
of poverty, mental illness, infant
mortality, physical illness, juvenile
delinquency and adult criminality,
sexual abuse and other forms of
family violence, economic hard-
ship, substance abuse, and educa-
tional failure, such as increased
risk of dropping out of school.
4
Table 1. U.S. Children Residing in Two-Parent Families
2005
85.2%
68.1%
68.3%
1970
1998
(Source:U.S.Bureau of the Census)
One Female Parent
21.5%
78.5%
One Male Parent
(Source:2005 American Community Survey)
Table 2. Percent of U.S.Children
in a Single-Parent Household that Has . . .
Page 8
But marriage is more than a moral or even social institution; it is also an economic
one, a generator of social and human capital, especially when it comes to children.
Not much attention has been focused to date on the hard, economic costs of family
fragmentation, by which we mean not only the economic costs to affected individ-
uals and families but also to the public purse.
There are good reasons for suspecting that taxpayer costs associated with family
fragmentation are substantial: To the extent that the decline of marriage increases
the number of children and adults eligible for and in need of government services,
costs to taxpayers will grow. To the extent that increases in family fragmentation
also independently drive social problems faced by communities—such as crime,
domestic violence, substance abuse, and teen pregnancy—the costs to taxpayers of
addressing these increasing social problems are also likely to be significant. Pointing
out these concerns is not to “blame the victim,” but rather to launch a serious effort
to determine what these costs are. If these costs are deemed substantial, then it is
worth thinking carefully about how these costs can be lowered so that resources
can be freed for other useful purposes.
In 2000, a group of more than one hundred family scholars and civic leaders noted
the range of public costs associated with family breakdown, concluding:
Divorce and unwed childbearing create substantial public costs, paid by tax-
payers. Higher rates of crime, drug abuse, education failure, chronic illness,
child abuse, domestic violence, and poverty among both adults and children
bring with them higher taxpayer costs in diverse forms: more welfare expen-
diture; increased remedial and special education expenses; higher day-care
subsidies; additional child-support collection costs; a range of increased direct
court administration costs incurred in regulating post-divorce or unwed fami-
lies; higher foster care and child protection services; increased Medicaid and
Medicare costs; increasingly expensive and harsh crime-control measures to
compensate for formerly private regulation of adolescent and young-adult
behaviors; and many other similar costs.
While no study has yet attempted precisely to measure these sweeping and
diverse taxpayer costs stemming from the decline of marriage, current research
suggests that these costs are likely to be quite extensive.
5
In response to public concerns about the negative consequences of divorce and
unmarried childbearing for child well-being, the federal government and many
states have modestly funded programs aimed at strengthening marriage.
Since the mid-1990s, at least nine states have publicly adopted a goal of strength-
ening marriage, and seven states have dedicated funding (often using a very small
Page 9
portion of their federal TANF, or welfare, funds) to various programs designed to
strengthen marriage.
6
For example, Oklahoma offers marriage skills classes throughout the state, provid-
ing the courses at no charge to low-income participants. In 2007, Texas legislators
mandated that a minimum of 1 percent of the federal TANF block grant to the state
be spent on marriage promotion activities, providing an estimated $15 million per
year for two years.
7
In addition to the TANF block grants, the Deficit Reduction Act of 2005 provided an
additional $150 million annually for a Healthy Marriage and Responsible
Fatherhood Program, administered by the Administration for Children and Families
of the Department of Health and Human Services. These monies were specifically
allocated for programs designed to help couples form and sustain healthy marriage
relationships, with up to $50 million available for responsible fatherhood promo-
tion.
8
Overall, less than 1 percent of TANF dollars are spent annually on healthy
marriage programming.
Evaluation is under way to determine the effectiveness of these programs. In the
meantime, this study provides the first rigorous estimate of the costs to taxpayers of
the decline of marriage, both at the national level and the state level.
9
II. How Might Marriage Affect Taxpayers?
Empirical Literature Review
R
ESEARCH SUGGESTS THAT MANY of the social problems and disadvantages
addressed by federal and state government programs occur more frequently
among children born to and/or raised by single parents than among children
whose parents get and stay married.
10
The potential risks to children raised in
fragmented families that have been identified in the literature include poverty,
mental illness, physical illness, infant mortality, lower educational attainment
(including greater risk of dropping out of high school), juvenile delinquency, con-
duct disorders, adult criminality, and early unwed parenthood. In addition, family
fragmentation seems to have negative consequences for adults as well, including
lower labor supply, physical and mental illness, and a higher likelihood of commit-
ting or falling victim to crime.
11
To the extent that family fragmentation causes negative outcomes for children and
adults, it also leads to higher costs to taxpayers through higher spending on
antipoverty programs and throughout the justice and educational systems, as well
as losses to government coffers in foregone tax revenues.
Page 10
A crucial issue for this study is to ascertain to what extent the associations between
family fragmentation and these negative outcomes are causal. There are of course
powerful selection effects into marriage, divorce, and unwed childbearing, and
some portion of the negative outcomes for children in nonmarital families are
caused by habits, traits, circumstances, and disadvantages among adults that may
also lead to divorce and nonmarital childbearing.
12
For example, a dating couple facing an unexpected pregnancy may choose not to
marry because the man is unemployed. Depending on how one looks at it, the
out-of-wedlock birth may be said to result from the father’s low-earnings or the
mother and child’s poverty may be said to result from the out-of-wedlock birth.
Untangling “what causes what” is a challenge faced by many researchers who
study the family.
How Much Does Marriage Reduce Poverty?
Researchers respond to this challenge by using a variety of methods to control for
unobserved selection effects (that is, to account for other factors that could be
explaining the finding) and to tease out causal relationships (that is, to untangle
“what causes what”). In this case, the idea that family fragmentation contributes to
child poverty has been studied extensively and is widely accepted.
13
Marriage can
help to reduce poverty because there are two potential wage earners in the home,
because of economies of scale in the household, and possibly also because of
changes in habits, values, and mores that may occur when two people marry.
14
In addition, there is recent, intriguing research that uses naturally occurring evi-
dence to examine whether family fragmentation causes poverty. Elizabeth Ananat
and Guy Michaels, for example, use an unusual predictor of whether a married
couple will stay married (the predictor is whether their firstborn child is a male,
since research has shown that divorce is less likely when this is the case). With this
predictor they are able to study married couples who do and do not divorce and
conclude that “divorce significantly increases the odds that a woman with children
is poor.”
15
Their analysis suggests that almost all of the increase in poverty
observed among divorced mothers is caused by the divorce. Less than 1 percent
of these women and children live in poverty if their first marriage is intact, while
more than 24 percent of divorced women with children are living in poverty.
16
Another area of research uses national data to simulate changes in family structure.
For example, Robert Lerman used the Current Population Survey (CPS) and simu-
lated “plausible” marriages by matching single mothers to single males who were
the same race and were similar in age and education levels. He found that if these
theoretical marriages occurred they would reduce poverty by 80 percent among
these single-mother households.
17
Adam Thomas and Isabel Sawhill used a similar
Page 11
approach and concluded that marriage would reduce poverty among single moth-
ers substantially, by about 65 percent.
18
Both Lerman’s and Thomas and Sawhill’s estimates assume that getting married has
no effect on men’s labor supply (and therefore male earnings). Most research on
this topic, by contrast, finds that marriage leads to a modest increase in male labor
supply, which would further reduce poverty rates. David Ribar did a useful survey
of the literature on the impact of marriage on men’s earnings.
19
Other research that seeks to analyze the impact of marriage on poverty consists of
studies that conduct a “shift-share analysis,” which show what poverty rates would
be if the proportion of households in different family structures remained constant
over a given time period. Examples of this research include work done by Hilary
Hoynes, Marianne Page, and Ann Stevens and by Rebecca Blank and David Card.
These studies find that over 80 percent of poverty is related to changes in family
structure, such as increases in households headed by single mothers.
20
One cautionary note, however, is that these studies could overstate the impact of
family structure on income because they do not account for the likelihood that, as
Thomas and Sawhill say, “single-parent families possess characteristics that dispro-
portionately predispose them to poverty.”
21
For example, persons struggling with
mental illness, substance abuse, or criminal records might find it difficult both to
hold a job and to get or stay married. Nevertheless, even studies that attempt to
control for these factors strongly suggest that family fragmentation negatively affects
the income available to single parents and their children.
Does Family Fragmentation Increase Crime?
In addition to poverty, family fragmentation also appears to have large effects on
rates of crime, according to three separate bodies of literature.
For example, research that considers entire communities has found a strong associ-
ation between the percent of single-parent households and crime rates. In one case,
Robert O’Brien and Jean Stockard found that increases in the proportion of adoles-
cents born outside of marriage were linked to significant increases in homicide
arrest rates for fifteen to nineteen year olds.
22
A second large body of literature—investigations of individual families using vari-
ables, such as parent-child relationships or mothers’ education levels—finds that a
child raised outside of an intact marriage is more likely to commit crimes as a teen
and young adult. In one study Cynthia Harper and Sara McLanahan control for a
large number of demographic and other characteristics and find that boys reared in
single-mother households and cohabitating (or “living together”) households are
Page 12
typically more than twice as likely to commit a crime that leads to incarceration,
when compared to children who grow up with both their parents.
23
Finally, a body of literature that analyzes future crime rates of juvenile offenders
shows that stable marriages reduce the likelihood that adult males will commit addi-
tional crimes. With a unique data set of former juvenile offenders spanning several
decades, Robert Sampson and his colleagues find evidence that marriage leads
these former juvenile offenders to commit fewer crimes as adults, even when con-
trolling for unobserved selection effects.
24
Overall, research on family structure suggests a variety of ways marriage might
reduce the demand for costly public services. A stable marriage might reduce the
likelihood of domestic violence, alcohol abuse, and parental depression, and might
increase the human and social capital available to children in the home in ways that
(independent of income) improve children’s educational and other outcomes. Two
parents in the home might provide more effective supervision of adolescents,
reducing the risk of delinquent activities. At the same time, divorce may be some-
times desirable. For example, about one-third of marriages ending in divorce are
“high conflict” marriages, and children, on average, appear to be better off when
those marriages end.
25
In this analysis, however, we adopt the simplifying and extremely cautious
assumption that all of the taxpayer costs of divorce and unmarried childbearing
stem solely from the negative effects family fragmentation has on poverty in female-
headed households. We make this simplifying assumption because the effect of
marriage on poverty has been established, is widely accepted, and can be reason-
ably well-quantified based on existing data.
III. Is the Methodology Used in This Estimate Reasonable?
T
HIS STUDY USES SEVERAL CALCULATIONS to estimate the taxpayer costs of family
fragmentation. These estimates include calculations of foregone tax revenue
in income taxes, FICA (Social Security and Medicare) taxes, and state and
local taxes as a result of family fragmentation. They also include the direct costs to
taxpayers as a result of increased expenditures on local, state, and federal taxpayer-
financed programs in the following areas:
• Temporary Assistance for Needy Families (TANF) cash assistance
• Food Stamps
• Housing Assistance
• Medicaid
• State Children’s Health Insurance Program (SCHIP)
• Child Welfare programs
Page 13
• Women, Infants, and Children (WIC) assistance
• Low Income Home Energy Assistance Program (LIHEAP)
• Head Start
• School Lunch and Breakfast Programs
• The Justice System
26
As noted previously, we assume taxpayer costs are driven exclusively by increases
in poverty; that is, we used the most widely accepted and best quantified conse-
quence of divorce and unmarried childbearing. It is important to recognize that if
family fragmentation has additional negative effects on child and adult well-being
that operate independently of income—and if these effects increase the numbers of
children or adults who need and are served by taxpayer-funded social programs—
then our methodology will significantly underestimate taxpayer costs. For example,
if family fragmentation increases the number of children who suffer from chronic
illnesses,
27
these additional costs to taxpayers would not be reflected in the esti-
mates provided by this study.
To put it another way, the methodology we use assumes that marriage would not
improve the habits, mores, or other behaviors of adults or children in ways that lead
to reduced social problems or increased productivity.
What Costs Are Associated with Means-Tested Government Programs?
To obtain an estimate of the taxpayer costs of family fragmentation, this study uses
the literature and information already described to make three key assumptions:
• Assumption 1: Marriage lifts zero households headed by a single male out
of poverty.
• Assumption 2: Marriage lifts 60 percent of households headed by a single
female out of poverty.
• Assumption 3: The share of expenditures on government antipoverty
programs that is due to family fragmentation is equal to the percent of
poverty that results from family fragmentation.
28
Taken as a group, these assumptions err on the side of caution. They are more likely
to lead to an underestimate of the actual taxpayer costs of family fragmentation
rather than an overestimate. Assumption 1 leads us to understate taxpayer costs
because marriage might bring a second wage earner into single-father households
and/or allow men to focus more effort on labor market activities that would
increase household earnings. Assumption 2 is based on the discussion on pages
10–11 of this report, specifically the empirical results provided by Ananat and
Michaels and Thomas and Sawhill.
29
Page 14
Assumption 3 implies that the proportion of poverty that can be attributed to fam-
ily fragmentation is equal to the proportion of expenditures on a variety of govern-
ment programs that are caused by family fragmentation. As shown in table 3, if mar-
riage would lift 60 percent of single-mother households out of poverty, then the
total number of persons in poverty would decline by 31.7 percent and the total
number of children in poverty would decline by 36.1 percent.
30
By virtue of
assumption 3, marriage would reduce the costs of some government programs by
31.7 percent and the costs of government programs that are exclusively for children
by 36.1 percent. Put another way, this assumption suggests that family fragmenta-
tion is responsible for 31.7 percent of the costs of government antipoverty programs
and is responsible for 36.1 percent of the costs of government programs that are
exclusively for children.
31
This crucial assumption seems cautious not only because single-parent households
have higher rates of poverty and other negative outcomes but also because at the
same income level single-parent households are much more likely than married
households to make use of government benefits.
In the cautious assumptions used in this analysis, we assume no behavioral effects
from marriage on the likelihood of choosing to use government programs, even
though (as shown in tables 4, 5, and 6) single-mother households use the Food
Stamp, cash assistance, and Medicaid programs at much higher rates than married
households with similar incomes.
Percent Receiving
Percent Receiving Food Stamps
Food Stamps Families Earning < 200%
Family Type All Income Levels of Poverty Level
Married 3.9% 16.2%
Male head no spouse present 8.6% 21.2%
Female head no spouse present 26.1% 42.5%
(Source: 2006 CPS)
Table 4.Household Income and Usage of Food Stamps
Number Lifted Out of
Poverty via Marriage
(thousands)
Total U.S.
@60% of female-headed
Poverty 2006
households in poverty Percent Lifted Out of
(thousands)
are lifted out of poverty Poverty via Marriage
Total Persons 36,460 11,554 31.7%
Children 12,827 4,629 36.1%
(Source:2006 CPS)
Table 3.Persons and Children Lifted out of Poverty via Marriage
Page 15
Assumption 1 means that our analysis focuses on female-headed households only;
that is, taxpayer costs associated with single-father households are excluded.
Assumptions 2 and 3 allow us to make cautious and straightforward estimates of
increased government expenditures on TANF, Food Stamps, housing assistance,
Medicaid, SCHIP, child welfare programs, WIC, LIHEAP, Head Start, and school
breakfast and lunch programs that result from family fragmentation. (See more
details in “Notes to Table A.1” on page 33.) For government programs that serve
both adults and children (TANF, Food Stamps, housing assistance, Medicaid, WIC,
and LIHEAP), we assume that 31.7 percent of these costs are due to family fragmen-
tation. We make this assumption because existing data (as shown in table 3) sug-
gests that family fragmentation is responsible for 31.7 percent of overall poverty,
and assumption 3 suggests that family fragmentation is responsible for 31.7 percent
of taxpayer costs on these programs.
For government programs that serve only or predominantly children (such as SCHIP,
child welfare programs, Head Start, and school breakfast and lunch programs), we
assume that 36.1 percent of these costs are due to family fragmentation.
32
We offer one cautionary note: The taxpayer costs associated with family fragmen-
tation may be real, but this link does not mean that taxpayers would necessarily
choose to realize all the tax savings from reductions in family fragmentation.
Many transfer programs, such as Head Start, Section 8 housing vouchers, and
LIHEAP are not entitlements. That means not all individuals or households poten-
tially eligible to receive funding or services under these programs receive them.
For non-entitlement programs, savings realized from increases in marriage and
marital stability could be reaped by taxpayers or the savings might be passed on
to other poor people.
33
But if such savings were to occur, legislators and voters
could either choose to change the rules and use the money for other government
purposes or return it to taxpayers.
The next steps in our study were finding ways to estimate any increased costs to
the justice system caused by family fragmentation and any foregone tax payments
that would result from eliminating family fragmentation. These two sets of calcula-
tions require some discussion.
What Costs Are Associated with the Justice System?
Evidence suggests that boys raised in single-parent households are likely to commit
crimes at much higher rates than boys raised in married households.
34
Further, mar-
riage reduces the likelihood that adult men will commit crimes.
35
For the purposes of calculating the impact of family fragmentation on increased
costs to the justice system, however, we use the following cautious and simplifying
assumption: All of the effects of family fragmentation on crime operate through their
impact on childhood poverty rates. In this analysis, we are following Harry Holzer
and his colleagues. They created a methodology to estimate the impact of eradicat-
ing childhood poverty on costs to the U.S. economy. One cost they consider is the
cost to the justice system—which includes courts, police, prisons, and jails.
Essentially, based on several assumptions taken from the empirical literature on
crime, they report that 24 percent of crime is caused by childhood poverty.
36
Using
this result, we estimate that if marriage were to reduce childhood poverty rates by
36.1 percent, then costs for the justice system would be reduced by approximately
$19 billion. (See details of this calculation in table A.1.)
How Are Foregone Tax Revenues Estimated?
To estimate the impact of family fragmentation on foregone tax revenues, we must
estimate the increase in taxable income that would result from marriage. We again
make the simplifying assumption that marriage has no behavioral effect; in other
words, marriage would not increase the labor supply of men and would therefore
have no impact on the taxable earnings of single parents who marry. Again, given
the rich literature on how marriage impacts male labor supply,
37
this is a cautious
assumption, which increases our confidence that our analysis does not overestimate
the actual taxpayer costs of the decline of marriage.
Page 16
Similarly, we assume that all of the effects of family fragmentation on children’s
future earnings capacity operate only through their impact on rates of childhood
poverty. Given the rich but difficult-to-quantify body of evidence that married par-
ents contribute to increasing the human and social capital of their children in other
ways (in addition to income),
38
this decision represents another simplifying but cau-
tious assumption, which increases our confidence that our results will not overesti-
mate the taxpayer costs.
There is good evidence on the impact of childhood poverty on future productivity.
Holzer and his colleagues estimate that childhood poverty reduces income nation-
ally by $170 billion per year. That is, they find that if children in poverty had instead
grown up in households that were not in poverty, then these children would as
adults earn $170 billion more each year.
39
Using Holzer’s estimate of total costs of
childhood poverty on adult annual earnings and the estimate that marriage would
reduce childhood poverty by 36.1 percent, we estimate that marriage would
increase taxable earnings by over $61 billion per year.
To translate this data into an estimate of tax losses from losses in future productiv-
ity, we must make simplifying assumptions about tax rates. For this analysis, we
assume that all of the increase in earnings is taxed at the 10 percent rate for U.S.
income taxes and that all of this increase in earnings is taxed at 15.3 percent for
FICA (as tax economists generally find that employees bear the burden of FICA tax-
ation through lower wages). To estimate losses in state and local taxation, we use
the national average percentage of income that is paid in state and local taxes—11
percent—as reported by the Tax Foundation on April 4, 2007.
40
(The details of this
calculation are shown in table A.1.)
IV. What Is the Total Estimated Cost of Family
Fragmentation?
H
OW MUCH DO HIGH RATES of divorce and unmarried childbearing cost U.S. tax-
payers? Here is our estimate:
Family fragmentation costs U.S. taxpayers at least $112 billion each year, or
over $1 trillion dollars per decade.
41
This $112 billion annual estimate includes the costs of federal, state, and local gov-
ernment programs and foregone tax revenues at all levels of government. Table 7
shows an itemized estimate.
To find the cost of family fragmentation in your state, turn to page 38.
Page 17
Table A.5 (page 38) reveals state-by-state estimates for the costs of family fragmen-
tation, and appendix B (page 31) describes the methods used to estimate the costs
of family fragmentation for state and local taxpayers. These state-by-state estimates
are a subset of the $112 billion total taxpayer cost.
We are confident this is a minimum figure because of the uniformly cautious
assumptions built into our methodology. For those who would like to dig deeper,
appendix A (page 22) provides a detailed response to possible arguments that we
have overestimated or underestimated taxpayer costs. For example, here are four
potential underestimates:
First, our estimate focuses exclusively on female-headed households; that is, we
assume the taxpayer costs of single-father families are zero. This assumption almost
certainly leads to an underestimate.
Second, we have excluded from analysis several expensive government programs
(because existing data does not allow us to quantify them with confidence), which
nonetheless very likely include significant marriage-related taxpayer costs. The tax-
payer-funded programs excluded from analysis include the Earned Income Tax
Credit (EITC), public education,
42
and Medicare and Medicaid benefits for older
adults. The EITC alone is a $40 billion taxpayer-funded program. Estimating the
effect of marriage on the EITC involves making complex judgments about who mar-
ries whom, and how their income shifts as a result. Since we lack hard data to make
Page 18
these judgments with the precision necessary to quantify them, we left this program
out of the analysis. While some fraction of currently cohabiting taxpayers might pay
a marriage penalty if they were to marry, the overall poverty-reducing effects of
marriage are likely to move many more families off the EITC rolls. (See appendix
A for more detail.)
Similarly, some fraction of public school budgets is likely spent in dealing with social
problems created by divorce and unmarried childbearing. Children whose parents
stay married are less likely to repeat a grade, exhibit conduct disorders requiring
special education outlays, or require expensive special education services generally.
Again, none of these costs are reflected in our analysis.
We have also excluded one of the largest taxpayer costs on the book: Medicaid for
the elderly and Medicare for unmarried adults. They are excluded partly because
most people do not think of older single adults when they think of “fragmented
families.” But high rates of divorce and failure to marry mean that many more
Americans enter late middle-age (and beyond) without a spouse to help them man-
age chronic illnesses, or to help care for them if they become disabled.
43
Through
the Medicare and Medicaid system taxpayers are picking up a large, but difficult to
quantify, part of the costs as a result.
Third, we have ignored for the purposes of this analysis any behavioral effects of
marriage. We have assumed that all the benefits of marriage come solely from
reduced rates of poverty for children, ignoring the evidence that stably married par-
ents provide human and social capital to their children other than income in ways
that increase children’s well-being and reduce the likelihood they will need or incur
expensive government services, from repeating grades at school to ending up in the
child protective system or the juvenile justice system.
Similarly, we have assumed no behavioral effects of marriage on fathers’ earning
capacity. If stable marriage increases men’s earnings, as the literature suggests, and/or
decreases the likelihood that they will commit crimes as adults, our methodology
most likely underestimates the taxpayer costs associated with unmarried parenthood.
Fourth, there is one other major reason we believe $112 billion each year repre-
sents a cautious, minimum estimate: For the purposes of this analysis, we assume
that households that marry will “take up” or use government benefits for which they
are eligible at the same rate as single-mother households. In reality, existing data
shows that lower-income married couples are far less likely to choose to use gov-
ernment benefits for which they are eligible than single-mother households.
Overall, single mothers are roughly twice as likely to take advantage of government
benefits for which they are eligible than are low-income married couples.
44
Page 19
Many more details, including a discussion of the empirical literature on which our
conclusions are based, are found in appendix A.
V. What Are the Policy Implications?
H
OW SHOULD POLICYMAKERS, state legislators, and others respond to these new,
rigorous estimates of the large taxpayer costs of family fragmentation?
First, public concern about the decline of marriage need not be based only on the
important negative consequences for child well-being or on moral concerns, as impor-
tant as these concerns may be. High rates of family fragmentation impose extraordi-
nary costs on taxpayers. Reducing these costs is a legitimate concern of government,
policymakers, and legislators, as well as civic leaders and faith communities.
Second, even very small increases in stable marriage rates would result in very large
returns to taxpayers. For example, a mere 1 percent reduction in rates of family
fragmentation would save taxpayers $1.12 billion annually.
Given the modest cost of government and civic marriage-strengthening programs,
even more modest success rates in strengthening marriages would be cost-effective.
Texas, for example, recently appropriated $15 million over two years for marriage
education and other programs to increase stable marriage rates. If such a program
succeeded in increasing stably married families by just three-tenths of 1 percent, it
would still save Texas taxpayers almost $9 million per year. Efforts are currently
underway to evaluate the impact of these programs.
Conclusion
E
ACH YEAR, FAMILY FRAGMENTATION costs American taxpayers at least $112 billion
dollars. These costs are recurring—that is, they are incurred each and every
year—meaning that the decline of marriage costs American taxpayers more
than $1 trillion dollars over a decade.
These costs are due to increased taxpayer expenditures for antipoverty, criminal jus-
tice and school nutrition programs, and through lower levels of taxes paid by indi-
viduals whose adult productivity has been negatively influenced by growing up in
poverty caused by family fragmentation.
This figure represents a minimum or “lower-bound” estimate. If, as research sug-
gests is likely, marriage has additional economic and social benefits to children,
adults, and communities—benefits that reduce the need for government services
and that operate through mechanisms other than increased income—then the actu-
al taxpayer costs of the retreat from marriage are likely much higher.
Page 20
Given the cautious assumptions used throughout this analysis, we can be confident
that current high rates of family fragmentation cost taxpayers at least $112 billion a
year, or more than $1 trillion over a decade. Finding new ways to strengthen mar-
riage and reduce unnecessary divorce and unmarried childbearing is a legitimate
and pressing public concern.
Because of the very large taxpayer costs associated with high rates of divorce and
unmarried childbearing, and the modest price tags associated with most marriage-
strengthening initiatives, state and federal marriage-strengthening programs with
even very modest success rates will be cost-effective for taxpayers.
Page 21
Appendix A: Testing the Analysis:
Is the Estimate of $112 Billion Too High or Too Low?
In this appendix, we consider in detail four arguments that suggest the estimate that
the total taxpayer cost of family fragmentation of $112 billion is too high and four
arguments that it is too low.
Is $112 Billion Too High?
In this section, we consider four arguments that suggest the $112 billion estimate is
too high:
1. If cohabitating households were to marry, most of them receiving transfer
payments would receive a marriage bonus from the federal tax and trans-
fer system.
2. The use of Holzer’s research in this study exaggerates the actual impact of
low income on childhood poverty.
3. The use of Thomas and Sawhill’s research in this study overestimates the
impact of marriage on reducing poverty.
4. The main assumption of this study—that the percentage of government
program costs due to family fragmentation is proportional to the amount
of poverty due to family fragmentation—overestimates the taxpayer costs.
1. If cohabitating couples were to marry, they would receive a marriage bonus from
the federal tax and transfer system.
The earned income tax credit (EITC) is an approximately $40 billion antipoverty
program that provides cash subsidies to low-income working adults, and Temporary
Assistance to Needy Families (TANF) is a $16 billion cash assistance program for
low-income families. Gregory Acs and Elaine Maag report that of the 1.1 million
people living in cohabitating households earning less than 200 percent of the
poverty level and receiving the EITC and/or TANF, about three-fourths of these
households would receive an average marriage bonus of approximately $2,423,
while about 10.5 percent of these households would receive a marriage penalty of
$1,742 in 2008. Thus, if the cohabitating adults in these households were to marry,
taxpayers would face increased expenditures for these two social programs. Taken
together, these estimates from Acs and Maag suggest that taxpayer expenditures for
the EITC, TANF, and the child tax credit would increase by about $0.5 billion if all
cohabitating couples were to marry.
45
Given the results in Acs and Maag, is our estimate of the taxpayer cost of family
fragmentation too high by $0.5 billion? In table 7, we ignore any costs of family frag-
mentation on the EITC. We do so because of complications such as the one pointed
out by Acs and Maag that some households would get higher EITC payments if they
became married households and nothing else about them (e.g., labor supply)
changed. Nevertheless, it appears likely that the net taxpayer costs of family
Page 22
fragmentation on the EITC would be positive because marriage would render mil-
lions of EITC recipients ineligible for any EITC benefits by adding a second wage
earner to the household and/or possible positive effects on male labor supply. If
marriage, as discussed in the next subsection, would decrease net expenditures on
the EITC, then the approach used in this study would underestimate the taxpayer
costs of family fragmentation by ignoring the EITC.
2. The use of Holzer’s results in this study overestimates the impact of low income
in childhood on adult outcomes.
We rely on results from Harry Holzer’s research to make two calculations—costs to
the justice system and increased tax payments. Holzer and his colleagues make a
case that the assumptions that underlie their estimates are cautious, and we find
their case persuasive. However, they make a broad interpretation of the relationship
between childhood poverty and life outcomes as follows:
[W]e interpret the causal effects of childhood poverty quite broadly. They
include not only the effects of low parental incomes, but also of the entire range
of environmental factors associated with poverty in the U.S., and all of the per-
sonal characteristics imparted by parents, schools, and neighborhoods to chil-
dren who grow up with or in them. We define “poverty” broadly in this way in
part because researchers have been unable to clearly separate low income from
other factors that affect the life chances of the poor, and also because the set
of potential policy levers that might reduce the disadvantages experienced by
poor children go beyond just increasing family incomes. Of course, in defining
poverty this way, we also assume that the entire range of negative influences
associated with low family incomes would ultimately be eliminated if all poor
children were instead raised in non-poor households.
46
(Holzer’s emphasis)
Childhood poverty may be a proxy for “environmental factors” that may or may
not be improved by the income gains from marriage. Thus, reliance on Holzer’s
estimates of the costs of childhood poverty could potentially overstate any bene-
fits of marriage that come from marriage reducing childhood poverty. One way to
think about this issue carefully is to list the broad pathways (i.e., “the environ-
mental factors”) through which growing up in poverty is associated with child-
hood disadvantage:
a. Low income may mean lower quality food, shelter, transportation, and
medical care for children.
b. Low income may necessitate living in worse neighborhoods (fewer parks,
more crime, less social trust), and poor neighborhood quality may
adversely affect child well-being.
c. Low income may mean attending worse schools.
d. Low income may negatively affect parenting processes (warmth, monitor-
ing, discipline) because of the stress economic hardship places on other-
wise competent parents.
Page 23
e. Conversely, parents may have low incomes because they are less
motivated and skilled and this lesser competence may be exhibited in
parenting as well.
f. Low income may hurt children because low-income families are more
likely to have only one parent present, and therefore only half the social
and human capital available to the child.
If absence of money itself is the root cause of the negative effects of childhood
poverty, then any strategy that increases income will increase child well-being.
With more money, parents can provide better nutrition, education, housing, and
medical care. They can move to better neighborhoods and enjoy better schools. It
may be the case, however, that not all the potential pathways through which child-
hood poverty negatively affects child well-being can be “treated” with more
money. More money may lower the income stress but not the emotional stress of
single parenting, for example. In addition, there may be less human and social cap-
ital that results when one parent—only half the potential talent pool for parent-
ing—is available.
If marriage increases household income, then marriage would ameliorate the neg-
ative effects of childhood poverty that operate through pathways labeled a, b, c,
and d above. Marriage may also address at least some of the other pathways
through which childhood poverty is associated with relative deprivation (f). But
marriage does not address all of the pathways: What if low-income parents are sim-
ply less competent generally in parenting as in other domains of life (e)? What if
they are less motivated to help their children succeed or have fewer of the skills
needed to help their children manage school or work?
To the extent that childhood poverty is caused by living with adults who have per-
sistent personality traits or skill deficits that lessen child well-being, neither income
supports nor increased marriage alone will “treat” these problems.
Holzer and his colleagues make an adjustment for genetic factors that may be pres-
ent in the ability to generate labor market earnings that they believe errs on the side
of caution. But not all the selection effects may be understood to be genetic in
nature. What if single moms or dads simply are people who have lower average
parenting skills and less motivation to begin with? If this is the case, then the use
of Holzer’s results for two calculations leads to an overestimate of the taxpayer cost
of family fragmentation.
Like Holzer, this analysis is constrained by the availability and quality of relevant
empirical evidence. We believe, however, that the way we use Holzer’s results does
not lead to an overestimate for at least three reasons:
• If the assumptions Holzer and his colleagues used are cautious, then that
offsets at least some of the “environmental” effects of having a mother or
father with less motivation, whether married or not.
Page 24
• Our analysis assumes no effect of marriage on the labor supply of parents.
The best evidence, as reported in Ribar’s extensive literature review in
2004, indicates that marriage increases male labor supply and seems not to
depress the average female labor supply in the more recent groups of
women studied.
47
• Our analysis assumes no behavioral effect of marriage on parenting skills. If
marriage reduces stress on parents, which leads to better parenting, then this
approach underestimates the true taxpayer costs of family fragmentation.
Given the limits of the available empirical evidence, we implicitly assume that these
three reasons exactly offset any of the effects of childhood poverty that are due to
unobserved lower motivation and/or skills present among single parents. To the
extent this assumption is wrong and it leads to overstating taxpayer costs because
of the use of Holzer’s research, the magnitude of the overestimate would have to
be viewed in light of the magnitude of our underestimates as described in the next
subsection. As suggested below, these underestimates are likely quite substantial.
3. The use of Thomas and Sawhill’s research overestimates the impact of marriage
on reducing poverty.
Thomas and Sawhill estimate that marriage would lift 65.4 percent of single-
mother households out of poverty.
48
In their microsimulation they place individu-
als in the March 1999 CPS in “plausible” marriages until they obtain a marriage rate
similar to 1970. Attempting to marry all single-mother households would likely fall
short because of a lack of marriageable men—prisoners are disproportionately men,
as are the unemployed, and men have lower life expectancies than women. The
dearth of marriageable men is one reason that we use a 60 percent figure instead
of the 65.4 percent estimate from Thomas and Sawhill. In addition, Thomas and
Sawhill assume no behavioral effects of marriage on male labor supply, which sug-
gests they underestimate the effect of marriage on poverty reduction. For these two
reasons, our use of Thomas and Sawhill’s research should not lead to an overesti-
mate of the taxpayer cost of family fragmentation.
4. The main assumption of this study—the percentage of the costs of government
programs due to family fragmentation is proportional to the percent of poverty
due to family fragmentation—overestimates the taxpayer costs.
To the contrary, the following thought experiment suggests that this assumption
likely leads to an underestimate of the taxpayer cost of family fragmentation.
Suppose there was an antipoverty program that cost taxpayers a total of $100 bil-
lion. Also suppose this program provided $5,000 per year to 10 million married
households and $5,000 per year to 10 million single-mother households. In addi-
tion, suppose that 20 million married households were eligible for the program but
only 10 million used it, while all 10 million single-mother households eligible for
the program used it.
Page 25
If each of the 10 million single-mother households in this thought experiment were
instead married households, consider two questions:
• How would the methodology used in this study estimate the taxpayer cost
of family fragmentation?
• What would be the “true” taxpayer cost of family fragmentation?
Using the methodology of this study, 6 million of the single-mother households (at
60 percent) using this program would no longer use it, which means the cost of
family fragmentation would be 6 million multiplied by $5,000, which equals $30 bil-
lion. Also, the analysis would assume that the remaining 4 million single-mother
households that are now married households would still use the program.
Would $30 billion likely be the “true” costs? We suspect not, because married
couples use benefits for which they are income-eligible at a much lower rate
than single-parent households: Only 50 percent of initially married households
eligible for the program use it. As shown in tables 4–6, single-mother house-
holds are far more likely at any given income level to choose to use govern-
ment benefits:
• Single-mother households with incomes less than 200 percent of the
poverty line are 2.6 times more likely to receive Food Stamps than married
households earning less than 200 percent of the poverty line.
• Single-mother households with incomes less than 200 percent of the
poverty line are 2.9 times more likely to receive cash assistance than mar-
ried households earning less than 200 percent of the poverty line.
• Single-mother households with incomes less than 200 percent of the
poverty line are 1.56 times more likely to receive Medicaid than married
households earning less than 200 percent of the poverty line.
If we assume that currently single mothers had instead married and that they
would use government benefits for which they are eligible at the same rate as
other married households (50 percent), then the “true” taxpayer costs of family
fragmentation would be $40 billion ($30 billion + 0.5 x 4 million x $5,000). Thus,
using the methodology of this study would understate the “true” costs by 33 per-
cent using the assumption that married households and single-mother households
receive the same average benefit ($5,000 per household in this example), and that
single-mother households take up the antipoverty program at a rate twice as large
as married households.
The main assumption of this study seems to be a reasonable simplifying assump-
tion, because of the much higher take-up rate of antipoverty programs of single-
parent households relative to similarly situated married households; this assumption
perhaps leads to an underestimate of the taxpayer cost of family fragmentation.
Page 26
To sum up, any differences in unobserved levels of average motivation between
single and married mothers complicate using much of the existing empirical litera-
ture to estimate the taxpayer cost of family fragmentation. The assumptions that
underlie this analysis, however, are extremely cautious in an attempt not to over-
state the taxpayer costs. Specifically, by assuming no beneficial behavioral effects
of marriage on adults or children, we are likely underestimating taxpayer costs.
Is $112 Billion Too Low?
In this section, we consider four arguments that suggest that the $112 billion esti-
mate is too low:
1. Ignoring the EITC, public education, and other government programs
underestimates the true taxpayer costs of family fragmentation.
2. Ignoring the direct impact of family fragmentation on crime (independent
of poverty) underestimates the taxpayer costs.
3. Ignoring any impact of marriage on single fathers understates the tax-
payer cost of family fragmentation.
4. Ignoring the fact that, given income-eligibility, single-mother households
are much more likely than married households to take up subsidies from
transfer programs underestimates the likely taxpayer costs of family frag-
mentation.
1. Ignoring the EITC, public education, and other government programs underesti-
mates the true taxpayer costs of family fragmentation.
We ignore EITC expenditures largely because of the lack of empirical information
needed to make reasonable assumptions about how marriage will affect usage of
the EITC and related programs in our complex tax code. But ignoring potential tax-
payer savings produced by marriage on EITC expenditures means ignoring a very
expensive government program that is almost certainly affected by marriage rates.
Taxpayers spend approximately $40 billion on cash assistance to the working poor
under the EITC. As shown in table A.2, using the assumptions in this study, family
fragmentation would lead to about $12.68 billion in higher taxpayer costs on the
EITC. Adjusting this estimate based on the results of Acs and Maag, as discussed
under the first argument in the previous subsection, would reduce that amount by
about $0.5 billion, leaving a net taxpayer cost of about $12.18 billion.
We have chosen to ignore the EITC expenditures (including potential savings of an
additional $12.18 billion each year) because the consequences of marriage for the
EITC are complex and would involve multiple assumptions of how marriage would
affect men’s and women’s earnings.
In addition to the EITC, this analysis does not assume any costs of family fragmen-
tation to the public school system, which is almost certainly not true. Considerable
research suggests that children raised outside of intact marriages are more likely to
Page 27
be held back a grade, to be in special education, and to qualify for remedial serv-
ices, although we do not have hard data on how much of these effects are due to
unobserved selection bias and how much are “caused” by lack of marriage.
If marriage were to reduce the percentage of children receiving special or remedial
services, then family fragmentation would create significant taxpayer costs for
public education, as federal and state funding formulas tend to provide large
amounts of extra funding for children receiving these services. (These costs may
be offset, however, by more teens dropping out of school as a result of family
fragmentation, which reduces the direct taxpayer costs of public education.
49
)
The lack of evidence of exogenous changes in family structure on the likelihood of
receiving special education or remedial services or staying in school, and the lack of
comparable cost data on remedial and special education services across states,
makes it impossible to estimate these costs with confidence. But the lack of data
does not mean that family fragmentation has no impact on educational expenditures.
Finally, we exclude the approximately 71 percent of Medicaid expenditures devoted
to the disabled and the elderly from the analysis, thereby making the cautious
assumption that family fragmentation has no impact on these expenditures. Most
people do not think of elderly unmarried adults or middle-aged disabled singles as
belonging to “fragmented families.” Nonetheless, there is considerable evidence that
older adults who are unmarried are more likely to become disabled, to manage
chronic diseases less successfully, and to need nursing home care as they age.
50
Excluding these large public costs thus likely significantly underestimates the
actual costs to taxpayers from the decline in marriage.
2. Ignoring the direct impact of family fragmentation on crime (independent of
poverty) underestimates the taxpayer costs.
Estimates of the potential impact of family structure on crime, even those that do
not control for selection bias, are large and arguably should not be ignored. As dis-
cussed in the section on methodology on page 12, it appears that family fragmen-
tation has large effects on crime, both in terms of increasing the likelihood that a
child raised outside of marriage will commit crimes
51
and the likelihood that adult
men will leave criminal activity after they are married.
52
While Harper and
McLanahan use a large number of control variables to help isolate the effect of fam-
ily structure on youth crime, they do not control for unobserved selection effects.
Nonetheless, their estimated effects of family structure on crime are extremely
large—typically children reared in single-mother households are more than twice as
likely to engage in criminal activities as children reared in a married household. For
example, they report that children living with a single mother are 2.168 times more
likely to be incarcerated than children living with both parents, all else being
equal.
53
Suppose we had assumed that over half their result was due to selection
bias—that the single mothers in their sample possessed such poor parenting skills
that even if they got married most of the estimated effect Harper and McLanahan
reported was due to selection bias. Specifically, suppose that children reared with
Page 28
a single mother are only 50 percent more likely to engage in criminal activity than
children raised with both parents, all else being equal. As shown in table A.2, using
this more aggressive approach yields an estimate that family fragmentation is
responsible for about $29 billion in costs to the justice system as opposed to the
$19.3 billion estimate used to generate the main result of this study.
The $29 billion estimated cost of family fragmentation to the justice system is
either too high or too low depending on the true magnitude of any exogenous
effects of marriage on criminal activity. The $19.3 billion figure represents about
8.7 percent of all costs to the justice system ($19.3 billion / $222.8 billion = 0.087),
while the $29 billion figure represents 13 percent of all costs to the justice system
($29 billion / $222.8 billion = 0.13).
To put these two estimates in context, note that, according to the Bureau of Justice
Statistics, in 2002 only 43.6 percent of inmates report that they lived with both par-
ents “most of the time” while growing up.
54
While the majority of inmates did not
live with both parents most of the time while growing up, the figure used to gen-
erate the main estimate of this study suggests that only 8.7 percent of the costs of
the justice system can be attributed to family fragmentation. As stated previously,
Sampson and his colleagues endeavor to control for selection effects and find that
former juvenile offenders commit fewer crimes as adults when married. Because
our estimates here ignore these potential taxpayer savings from marriage, our
method is more likely to underestimate than overestimate the taxpayer costs of fam-
ily fragmentation to the justice system.
3. Ignoring any impact of marriage on single fathers understates the taxpayer cost
of family fragmentation.
Research suggests that married men become more committed workers at least in part
as a result of marriage. Therefore, if single fathers were to marry, it is likely that their
labor supply would increase leading to increased tax payments. Further, tables 4–6
show that single-father households have higher take-up rates of antipoverty pro-
grams than married households with similar incomes.
Adding a second wage earner would render single-father households less likely to
receive government assistance via increased income and economies of scale.
Economies of scale via marriage—essentially “savings from size”—imply that by liv-
ing together, two adults are better able to share expenses and escape poverty. Ribar
provides an example of how marriage leads to economies of scale:
Consider the outcomes for a couple with a 9th–11th grade education and one
child in 2001. The median annual income for a woman with this level of edu-
cation was $10,330, while the median annual income for a similarly educated
man was $19,434. If the mother and child lived apart from the father, their
income would have been below the two-person poverty threshold of $12,207;
however, if the family lived together, their combined income would have
Page 29
exceeded the three-person threshold of $14,255. The mother and child would
have also met the gross income requirement for food stamps if they lived apart
from the father but would [sic] been ineligible if they lived with him. Even if
the mother had no income and the family just depended on the father’s
resources, they would have been above the poverty line and ineligible for food
stamps if they all lived together.
55
4. Ignoring the fact that, given income-eligibility, single-mother households are
much more likely than married households to take up subsidies from transfer
programs underestimates the taxpayer costs.
As shown in tables 4–6, single-mother households have higher take-up rates of gov-
ernment antipoverty programs than married households with similar incomes. Thus,
even if single-mother households that were instead married households were to
remain eligible for transfer programs, it appears they would be less likely to use
them. The methods used to estimate the taxpayer cost of family fragmentation at
$112 billion ignore this likelihood, and suggest this estimate is too low.
To sum up, this study is likely underestimating the taxpayer cost of family fragmen-
tation because (1) there likely would be net savings of EITC expenditures due to
any increase in marriage rates of non-cohabitating single parents and savings from
other programs not considered here; (2) the estimated costs to the justice system
are too low if there is a direct effect of marriage on reducing crime—which seems
likely given the research done to date; (3) there are taxpayer costs of single-father
households that are ignored here; and (4) the take-up rate of antipoverty programs
would likely decline if single-parent households were instead married households
that remained eligible for these programs.
Page 30
Appendix B:
Explaining the Methodology for State-Specific Costs
This appendix describes the methodology used to estimate state-specific tax-
payer costs of family fragmentation. These estimates include costs to state and
local taxpayers.
The methods used to create the state-specific estimates are similar to the methods
employed to create the national estimate described in the body of this report. For
the state-specific estimates, we used the 2006 Current Population Survey to estimate
the state-specific reductions in total poverty and child poverty that would result
from marriage. These estimates are shown in the last columns of tables A.3 and A.4
and are based on assumptions 1–3 described on page 13. These tables include the
underlying data used as well as other information that reveal how total and child
poverty fall disproportionately on unmarried households, and on households
headed by single females in particular.
Table A.5 shows the components and the total state and local taxpayer costs of fam-
ily fragmentation for each state. These taxpayer costs include foregone state and
local tax revenue and costs to the justice system, TANF, Medicaid, SCHIP, and child
welfare programs. State-specific data for the overall costs of these programs come
from the sources listed in the “Notes to Table A.1” on page 33.
State-specific cost estimates, however, were not available for costs to the justice
system, and foregone earnings are not estimated at the state level. To make state-
specific estimates for these two line items, we assume that the proportion of tax-
payer costs that accrues to a given state is equal to the proportion of poverty
caused by family fragmentation in the state. For example, using the information
in table A.3, we calculate that 10.2 percent of all childhood poverty in the U.S.
that is due to family fragmentation occurs in the state of California. Thus, 10.2 per-
cent of the increase in national income that comes from reducing childhood
poverty via marriage is assigned to California. Correspondingly, 10.2 percent of
the reduction in state and local justice costs that results from marriage are also
assigned to California.
56
Page 31
Page 32
Calculation
1 Justice System (federal, state, & local)
Expenditures on Corrections, Police, and Courts $222,802,421,001
Amount of Crime Attributed to Childhood Poverty 0.24
Reduction in Childhood Poverty via Marriage 0.361
Cost of Family Fragmentation $19,303,601,755 = $22,802,422,001*.24*.361
2 TANF (federal & state)
Expenditures on TANF $16,100,000,000
Cost of Family Fragmentation $5,103,700,000 = $16,100,000,000*.317
3 Food Stamps
Persons Receiving Food Stamps 26,672,000
Reduction in Food Stamp Receipt via Marriage @31.7% 8,455,024 = 26,672,000*.317
Mean Food Stamp Benefit per person per year $1,131.24
Cost of Family Fragmentation $9,564,661,350 = $1,131.24*8,445,024
4 Housing Assistance
HUD Expenditures on Housing Assistance $23,019,000,000
Cost of Family Fragmentation $7,297,023,000 = $24,019,000,000*.317
5 Medicaid (federal & state)
Expenditures on Medicaid $303,222,842,723
Cost of Family Fragmentation $27,875,275,932 = $303,222,842,723*.317*.29
6 SCHIP (federal & state)
Expenditures on SCHIP $7,884,328,870
Cost of Family Fragmentation $2,846,242,722 = $2,846,242,722*.361
7 Child Welfare Services (federal, state, & local)
Expenditures on Child Welfare Services $25,465,943,844
Cost of Family Fragmentation $9,193,205,728 = $25,465,943,844*.361
8 WIC
Expenditures on WIC $4,997,309,299
Cost of Family Fragmentation $1,584,147,048 = $4,997,309,299*.317
9 LIHEAP
Expenditures on LIHEAP $2,181,384,985
Cost of Family Fragmentation $691,499,040 = $691,499,040*.317
10 Head Start
Expenditures on Head Start $7,470,990,545
Cost of Family Fragmentation $2,697,027,587 = $7,470,990,545*.361
11 School Breakfast and Lunch
Expenditures on Subsidized School Breakfast and Lunch Programs $9,638,590,455
Cost of Family Fragmentation $3,479,531,154 = $9,638,590,455*.361
12 Forgone Tax Receipts
Loss in National Income from Childhood Poverty (Holzer, et al., 2007) $170,000,000,000
Reduction in Childhood Poverty via Marriage 36.1%
Loss in National Income from Family Fragmentation $61,370,000,000 = $170,000,000,000*.361
Forgone Federal Taxes @ 10% Tax Rate $6,137,000,000 = $61,370,000,000*.10
Forgone FICA @ 15.3% Tax Rate $9,389,610,000 = $61,370,000,000*.153
Forgone State & Local Taxes @ 11% Average Tax Rate $6,750,700,000 = $61,370,000,000*.11
Table A.1: Sub-Calculations for Estimates in Table 7
Page 33
Notes to Table A.1
The numbers at the beginning of each paragraph correspond to the sub-calculations
listed in table A.1.
1. This calculation is adapted from a similar calculation by Harry Holzer and his colleagues (see
endnote 37). Jens Ludwig estimates that federal, state, and local taxpayers spent $200 billion on the
justice system—prisons, police, courts, etc.—in 2003 (see “The Costs of Crime” testimony to the U.S.
Senate Committee on the Judiciary on September 19, 2006, http://judiciary.senate.gov/
testimony.cfm?id-2068&wit_id-5749). If taxpayer expenditures on the justice system increased at the
rate of inflation indicated by the CPI-U, then taxpayers spent $222.8 billion on the justice system in
2007. Holzer and his colleagues use what they believe to be a conservative estimate that 24 percent
of crime is due to childhood poverty. Combined with the estimate that 36.1 percent of childhood
poverty is caused by family fragmentation, then the taxpayer cost of family fragmentation to the jus-
tice system is $222.8 billion times 0.24 times 0.361, which equals approximately $19 billion. (In the
tables, an asterisk denotes the multiplication function.)
2. We use FY 2005 data on TANF cash assistance expenditures that comes from the National
Association of State Budget Officers FY 2005 State Expenditure Report (Washington, DC: NASBO,
2006). We did not inflate the expenditure data for inflation because TANF expenditures seem to be
leveling off in recent years.
3. The expenditure on Food Stamps was retrieved from http://www.fns.usda.gov/pd
/fssummar.htm and excludes $2.7 billion in administrative costs. Thus, we assume that administrative
costs would not decrease due to a caseload decline.
4. The FY 2006 expenditure on federal housing assistance was retrieved from
http://www.gpoaccess.gov/usbudget/fy06/pdf/budget/hud.pdf. In FY 2006, HUD spent about $23 bil-
lion on homeless programs, rental assistance, and public housing. Thus, the calculation excludes
expenditures on other housing programs such as the Low Income Housing Tax Credit. We did not
inflate FY 2006 expenditure data for inflation because expenditures on housing programs have oscil-
lated in the past few years.
5. FY 2005 Medicaid expenditure data comes from NASBO FY 2005 State Expenditure Report and
includes federal and state expenditures. The FY 2005 expenditure for Medicaid was inflated using the
CPI-U to make an estimate of FY 2007 expenditures. Since Medicaid expenditures tend to grow faster
than the rate of inflation, this FY 2007 estimate is cautious. According to the Kaiser Family Foundation
in 2008 (see http://www.statehealthfacts.org/medicaid.jsp), 40 percent of Medicaid expenditures are for
the disabled, while another 26 percent are for the elderly. Non-elderly adults and children receive 29
percent of Medicaid expenditures. (The uses of the remaining Medicaid expenditures are reportedly
unknown.) For this analysis, we use the cautious assumption that family fragmentation leads to no
Medicaid costs for the elderly or the disabled. Thus, only 29 percent of total Medicaid expenditures are
potentially impacted by family fragmentation under this assumption.
6. FY 2006 federal and state expenditures on SCHIP were retrieved from the Kaiser Family
Foundation website (see http://www.statehealthfacts.org/comparetable.jsp?ind-235&cat-4). Given the
uncertainty surrounding SCHIP reauthorization, we did not inflate expenditures to 2007 dollars.
7. An estimate of FY 2004 federal and state expenditures on the child welfare system came from
C. A. Scarcella et al., The Cost of Protecting Vulnerable Children (Washington, DC: Urban Institute,
2006). We inflated their FY 2004 estimate for 2007 dollars using the CPI-U. Scarcella and colleagues
label the following government programs as “child welfare” programs: services for children and fam-
ilies to prevent abuse and neglect, family preservation services, child protective services, in-home serv-
ices, out-of-home placements such as foster care, and adoption services.
8. Information on FY 2003 federal WIC expenditures came from the U.S. House Committee on
Ways & Means Green Book (see http://www.gpoaccess.gov/wmprints/green/index.html) and was esti-
mated for FY 2007 using the CPI-U.
9. Information on FY 2003 federal LIHEAP expenditures came from the U.S. House Committee on
Ways & Means Green Book (http://www.gpoaccess.gov/wmprints/green/index.html) and was estimated
for FY 2007 using the CPI-U.
10. Information on FY 2003 federal Head Start expenditures came from the U.S. House Committee
on Ways & Means Green Book (http://www.gpoaccess.gov/wmprints/green/index.html) and was esti-
mated for FY 2007 using the CPI-U.
Page 34
11. Information on FY 2003 federal School Breakfast and Lunch expenditures came from the U.S.
House Committee on Ways & Means Green Book (http://www.gpoaccess.gov/wmprints
/green/index.html) and was estimated for FY 2007 using the CPI-U.
12. Holzer and his colleagues estimate that U.S. national income is $170 billion lower because of
childhood poverty. Using the estimate that 36.1 percent of childhood poverty is due to family frag-
mentation, then the decrease in national income from family fragmentation is approximately $61 bil-
lion. To make these calculations we assume that this extra national income would be taxed at a 10
percent federal marginal income tax rate, a 15.3 percent FICA tax rate, and an 11 percent state plus
local tax rate. The latter figure is the average percent of income spent on state plus local taxation (see
the Tax Foundation website at http://www.taxfoundation.org/taxdata/show/335.html). We use a 15.3
percent FICA tax rate because tax economists generally believe that employees bear the full burden
of the employer share of these payroll taxes via lower wages (see Daniel Hamermesh and Albert Rees,
The Economics of Work and Pay, 5th ed. (New York: HarperCollins College Pub., 1993).
Page 35
EITC
Expenditures on EITC $40,000,000,000
Cost of Family Fragmentation @ 31.7% cost reduction via marriage $12,680,000,000 = $40,000,000,000*.317
Savings from Family Fragmentation Using Estimates from Acs and Maag (2005) $500,000,000
Net Cost of Family Fragmentation $12,180,000,000 = $12,680,000,000 - $500,000,000
Justice System (federal, state, & local)
Expenditures on Corrections, Police Protection, and Courts in U.S. $222,802,421,001
Amount of Crime Attributed to Childhood Poverty 0.24
Amount of Crime Attributed to Childhood Poverty and Family
Fragmentation @ assumed 50% higher likelihood of children reared
by single mothers to engage in crime 0.36 = .24*1.5
R
eduction in Childhood Povertyvia Marriage 0.361
Cost of Family Fragmentation $28,955,402,633 = $222,802,421,001*.361*.36
Table A.2: Sub-Calculations for EITC and Justice System Estimates
Page 36
(Source: 2006 CPS)
Table A.3: Total Poverty and Family Structure by State
Percent Reduction
Number in Poverty Number in Poverty in To tal Poverty
in Unmarried in Unmarried Percent ofTo tal Poverty if Marriage
Number in Poverty Households with Households with Percent of Total Living in Unmarried Reduced Poverty
Tot al Number in in Husband-Wife Male Householder Female Householder Poverty Living in Households with ofFemale-headed
Poverty (thousands) Family (thousands) (thousands) (thousands) Unmarried Household Female Householder Households by 60%
AL 650 196 88 366 69.8% 56.3% 33.8%
AK 58 10 13 3582.8% 60.3% 36.2%
A
Z 902 361 188 353 60.0% 39.1% 23.5%
AR 487 140 67 280 71.3% 57.5% 34.5%
CA 4,427 1,732 760 1,935 60.9% 43.7% 26.2%
CO 466 159 88 220 65.9% 47.2% 28.3%
CT 275 54 49172 80.4% 62.5% 37.5%
DE 80 27 14 3866.3% 47.5% 28.5%
DC 104 11 29 6489.4% 61.5% 36.9%
FL 2,068 580 383 1,106 72.0% 53.5% 32.1%
GA 1,172 300 156 716 74.4% 61.1% 36.7%
HI 116 27 35 5376.7% 45.7% 27.4%
ID 141 41 30 7070.9% 49.6% 29.8%
IL 1,338 274 241 823 79.5% 61.5% 36.9%
IN 674 134 129 411 80.1% 61.0% 36.6%
IA 301 79 67156 73.8% 51.8% 31.1%
KS 349 91 77182 73.9% 52.1% 31.3%
KY 690 234 117 339 66.1% 49.1% 29.5%
LA 713 200 96 417 71.9% 58.5% 35.1%
ME 134 29 28 7778.4% 57.5% 34.5%
MD 469 79 110 279 83.2% 59.5% 35.7%
MA 758 183 166 409 75.9% 54.0% 32.4%
MI 1,323 315 259 749 76.2% 56.6% 34.0%
MN 422 125 111 186 70.4% 44.1% 26.4%
MS 596 150 67 379 74.8% 63.6% 38.2%
MO 659 228 113 318 65.4% 48.3% 29.0%
MT 125 28 23 7577.6% 60.0% 36.0%
NE 180 71 40 6960.6% 38.3% 23.0%
NV 241 66 56120 72.6% 49.8% 29.9%
NH 71 16 20 3577.5% 49.3% 29.6%
NJ 762 241 121 400 68.4% 52.5% 31.5%
NM 328 116 64 148 64.6% 45.1% 27.1%
NY 2,668 666 468 1,534 75.0% 57.5% 34.5%
NC 1,225 348 188 689 71.6% 56.2% 33.7%
ND 70 14 17 3980.0% 55.7% 33.4%
OH 1,371 316 204 851 77.0% 62.1% 37.2%
OK 531 165 103 263 68.9% 49.5% 29.7%
OR 439 158 54 227 64.0% 51.7% 31.0%
PA 1,397 294 295 808 79.0% 57.8% 34.7%
RI 110 27 17 6675.5% 60.0% 36.0%
SC 474 127 51 296 73.2% 62.4% 37.5%
SD 82 18 16 4878.0% 58.5% 35.1%
TN 879 310 129 441 64.7% 50.2% 30.1%
TX 3,816 1,446 625 1,745 62.1% 45.7% 27.4%
UT 235 65 35135 72.3% 57.4% 34.5%
VT 48 12 10 2675.0% 54.2% 32.5%
VA 651 184 112 355 71.7% 54.5% 32.7%
WA 502 132 118 253 73.7% 50.4% 30.2%
WV 277 96 46136 65.3% 49.1% 29.5%
WI 555 117 97 341 78.9% 61.4% 36.9%
WY 51 14 10 2772.5% 52.9% 31.8%
Page 37
(
Source: 2006 CPS)
T
able A.4: Child Poverty and Family Structure by State
Number of Children Number of Children Percent Reduction
in Poverty in Poverty Percent of Total inTotal Child
Number of Children inUnmarried in Unmarried Percent of Total Child Poverty Poverty if Marriage
Total Number of in Poverty in Households with Households with Child Poverty Living in Unmarried Reduced Poverty
Children in Poverty Husband-Wife Family Male Householder Female Householder Living in Households with of Female-headed
(thousands) (thousands) (thousands) (thousands) Unmarried Household Female Householder Households by 60%
AL 209 64 12133 63.6% 69.4% 38.2%
AK 21 3 315 71.4% 85.7% 42.9%
AZ 329 142 29 159 48.3% 56.8% 29.0%
AR 183 56 6121 66.1% 69.4% 39.7%
CA 1,724 788 149 787 45.6% 54.3% 27.4%
CO 161 69 11 81 50.3% 57.1% 30.2%
CT 84 18 759 70.2% 78.6% 42.1%
DE 25 10 213 52.0% 60.0% 31.2%
DC 37 4 4 29 78.4% 89.2% 47.0%
FL 590 177 18 395 66.9% 70.0% 40.2%
GA 499 142 13 344 68.9% 71.5% 41.4%
HI 31 10 516 51.6% 67.7% 31.0%
ID 53 19 429 54.7% 64.2% 32.8%
IL 472 130 23 319 67.6% 72.5% 40.6%
IN 226 33 18175 77.4% 85.4% 46.5%
IA 101 34 958 57.4% 66.3% 34.5%
KS 137 38 14 8562.0% 72.3% 37.2%
KY 236 64 31141 59.7% 72.9% 35.8%
LA 252 60 20 173 68.7% 76.2% 41.2%
ME 37 9 42567.6% 75.7% 40.5%
MD 151 29 21100 66.2% 80.8% 39.7%
MA 199 51 15133 66.8% 74.4% 40.1%
MI 469 104 48 316 67.4% 77.8% 40.4%
MN 139 56 974 53.2% 59.7% 31.9%
MS 221 40 8173 78.3% 81.9% 47.0%
MO 247 100 20 127 51.4% 59.5% 30.9%
MT 36 9 1 26 72.2% 75.0% 43.3%
NE 58 30 423 39.7% 48.3% 23.8%
NV 81 29 646 56.8% 64.2% 34.1%
NH 17 5 2 9 52.9% 70.6% 31.8%
NJ 260 111 13 136 52.3% 57.3% 31.4%
NM 120 52 12 5545.8% 56.7% 27.5%
NY 858 205 57 597 69.6% 76.1% 41.7%
NC 449 131 33 285 63.5% 70.8% 38.1%
ND 23 6 2 15 65.2% 73.9% 39.1%
OH 515 109 32 373 72.4% 78.8% 43.5%
OK 188 63 25101 53.7% 66.5% 32.2%
OR 143 67 868 47.6% 53.1% 28.5%
PA 469 111 46 312 66.5% 76.3% 39.9%
RI 37 12 323 62.2% 67.6% 37.3%
SC 163 44 8110 67.5% 73.0% 40.5%
SD 26 7 2 16 61.5% 73.1% 36.9%
TN 306 125 21 159 52.0% 59.2% 31.2%
TX 1,436 593 86 756 52.6% 58.7% 31.6%
UT 101 29 765 64.4% 71.3% 38.6%
VT 12 3 1 758.3% 75.0% 35.0%
VA 241 72 12157 65.1% 70.1% 39.1%
WA 159 48 10101 63.5% 69.8% 38.1%
WV 83 26 948 57.8% 68.7% 34.7%
WI 199 56 5138 69.3% 71.9% 41.6%
WY 17 5 21058.8% 70.6% 35.3%
Page 38
Table A.5: Estimates of State and Local Taxpayer Costs of Family Fragmentation (in millions)
State & Local Foregone
State Tax Burden Tax Revenue Justice System TANF Medicaid SCHIP Child Welfare Total
C
alifornia 11.5% $717 $1,621 $515 $1,083 $153 $739 $4,829
New York 13.8% $657 $1,230 $202 $1,184 $73 $322 $3,668
T
exas 9.3% $559 $1,557 $64 $635 $46 $96 $2,957
Ohio 12.4% $368 $768 $48 $1,271 $32 $251 $2,739
Pennsylvania 10.8% $267 $643 $109 $839 $36 $421 $2,315
Florida 10.0% $313 $814 $57 $546 $43 $181 $1,953
Illinois 10.8% $275 $657 $19 $650 $98 $250 $1,949
Michigan 11.2% $281 $651 $90 $374 $30 $135 $1,562
Georgia 10.3% $281 $709 $59 $290 $49 $72 $1,460
N
orth Carolina 11.0% $250 $587 $59 $339 $21 $73 $1,329
New Jersey 11.6% $126 $280 $16 $419 $47 $115 $1,003
Massachusetts 10.6% $112 $274 $3 $335 $46 $175 $945
Indiana 10.7% $150 $361 $5 $158 $15 $150 $839
Virginia 10.2% $127 $323 $17 $234 $23 $51 $776
Tennessee 8.5% $107 $328 $12 $220 $0 $89 $757
Wisconsin 12.3% $135 $284 $11 $198 $15 $95 $737
Maryland 10.8% $86 $206 $9$266 $30 $127 $724
Connecticut 12.2% $57 $122 $45 $438 $3 $48 $712
Washington 11.1% $89 $208 $52 $266 $1$95 $711
Louisiana 11.0% $152 $356 $2 $109 $13 $37 $670
Missouri 10.1% $102 $262 $5 $211 $9 $75 $664
Arizona 10.3% $131 $328 $13 $134 $10 $38 $654
Kentucky 10.9% $122 $290 $21 $119 $8 $92 $654
Mississippi 10.5% $144 $356 $0 $84 $10 $10 $605
Minnesota 11.5% $68 $152 $9 $213 $11 $121 $574
Alabama 8.8% $93 $274 $3$124 $10 $44 $548
Arkansas 11.3% $109 $249 $4 $86 $6 $16 $471
South Carolina 10.7% $94 $227 $4 $132 $3 $8$469
Colorado 10.4% $67 $167 $0$111 $11 $99 $454
Oklahoma 9.0% $72 $208 $23 $92 $9 $26 $430
Kansas 11.2% $76 $175 $9 $78 $6 $45 $389
Oregon 10.0% $54 $140 $16 $111 $7$33 $361
Iowa 11.0% $51 $119 $17 $104 $6$61 $359
Utah 10.7% $55 $134 $8 $50 $4 $25 $276
West Virginia 10.9% $42 $99 $8 $50 $3$29 $231
New Mexico 9.8% $43 $113 $3 $57 $1 $12 $230
Maine 14.0% $28 $52 $37 $87 $4 $7$214
Rhode Island 12.7% $23 $47 $6 $78 $9 $43 $206
Nevada 10.1% $37 $95 $6 $44 $5 $13 $199
District of Columbia 12.5% $29 $60 NA NA $1 $72 $162
Nebraska 11.9% $22 $47 $6$41 $3 $23 $142
Idaho 10.1% $23 $60 $0$34 $2 $8$127
Alaska 6.6% $8 $31 $7 $42 $3 $22 $114
Montana 9.7% $20 $54 $5 $22 $2 $10 $113
Hawaii 12.4% $16 $33 $10 $32 $2$18 $112
New Hampshire 8.0% $6 $19 $8 $48 $1 $18 $99
Delaware 8.8% $9 $27 $1 $39 $1 $11 $88
South Dakota 9.0% $11 $33 $3 $22 $1 $7$77
Vermont 14.1% $8 $14 $5 $35 $0 $12 $74
North Dakota 9.9% $12 $31 $3 $17 $1 $5 $69
Wyoming 9.5% $8 $21 $12 $13 $1 $7$61
Page 39
Endnotes
1. Of course, the death of one’s spouse is another reason why an adult may not be married.
2. U.S. Census Bureau, 2005 American Community Survey.
3. Joyce A. Martin et al., “Births: Final Data for 2004,” National Vital Statistics Reports 55, no. 1
(September 29, 2006): 3.
4. W. Bradford Wilcox et al., Why Marriage Matters: 26 Conclusions from the Social Sciences (New
York: Institute for American Values, 2005), 10–11, and Institute for American Values, The Marriage
Movement: A Statement of Principles (New York: Institute for American Values, 2000),
http://center.americanvalues.org/?p=19.
5. Institute for American Values, The Marriage Movement, 11.
6. T. Ooms, S. Bouchet, and M. Parke, Beyond Marriage Licenses: Efforts in States to Strengthen
Marriage and Two-Parent Families (Washington, DC: Center for Law and Social Policy, 2004). The wel-
fare reform act of 1996 converted federal welfare funding (now known as TANF or Temporary
Assistance for Needy Families) into block grants to the states, becoming the first federal law explic-
itly to promote marriage. Personal Responsibility and Work Opportunity Reconciliation Act of 1996
(PRWORA), 42 U.S.C. 1305 (P.L. 104–193, Aug. 22, 1996). Three of the four purposes of the welfare
reform law relate to marriage, giving the states broad latitude in the use of the welfare funds: (1) to
provide assistance to needy families so that children may be cared for in their own homes or in the
homes of relatives; (2) to end the dependence of needy parents on government benefits by promot-
ing job preparation, work and marriage; (3) to prevent and reduce the incidence of out-of-wedlock
pregnancies; and (4) to encourage the formation and maintenance of two-parent families.
7. 2007 Texas H.B. 2683, fiscal note, http://www.capitol.state.tx.us/. Data on Texas program also
from personal communication with Bill Coffin, Special Assistant for Marriage Education at Administration
for Children and Families, Department of Health and Human Services.
8. Deficit Reduction Act of 2005, § 7103, P.L. 109–171 (codified at 42 U.S.C. 603(a)(2)). Allowable
marriage activities under the marriage initiative include the following: (1) public advertising campaigns
on the value of marriage and the skills needed to increase marital stability and health; (2) education
in high schools on the value of marriage, relationship skills, and budgeting; (3) marriage education,
marriage skills, and relationship skills programs, that may include parenting skills, financial manage-
ment, conflict resolution, and job and career advancement, for non-married pregnant women and non-
married expectant fathers; (4) premarital education and marriage skills training for engaged couples
and for couples or individuals interested in marriage; (5) marriage enhancement and marriage skills
training programs for married couples; (6) divorce reduction programs that teach relationship skills;
(7) marriage mentoring programs which use married couples as role models and mentors in at-risk
communities; and (8) programs to reduce the disincentives to marriage in means-tested aid programs,
if offered in conjunction with any activity above.
9. To address this issue, we propose the following thought experiment: If all currently unmarried
adult women were instead married (which would also mean all children now living with a single
mother were instead living with two married parents), how much would taxpayers save? The amount
that taxpayers would save if all single women (including mothers) married is the taxpayer cost of fam-
ily fragmentation. (Again, we are not saying that all women should be married, rather that posing such
a scenario helps us to capture the costs of family fragmentation.) Throughout the analysis, individu-
als who are not married or who have experienced a divorce or a nonmarital birth are considered to
be living in a “fragmented” family. As discussed below, we exclude all adults and children living with
a male householder with no spouse present from the analysis only in the interest of creating a very
cautious estimate of the taxpayer cost of family fragmentation. Further, as shown in table 2, single
motherhood occurs much more frequently than single fatherhood. As discussed below, it is likely not
theoretically possible (nor necessarily desirable) for all women to be married, and the analysis in this
study takes this concern into account.
10. Lack of fulltime work seems to be the biggest cause of poverty in America with family frag-
mentation being the second largest cause; see Ron Haskins and Isabel Sawhill, “Work and Marriage:
The Way to End Poverty and Welfare,” Policy Brief, Welfare Reform & Beyond #28 (Washington, DC:
Brookings Institution, 2003). See also Paul Amato, “The Impact of Family Formation Change on the
Cognitive, Social and Emotional Well-Being of the Next Generation,” The Future of Children 15, no. 2
(Fall 2005): 75–96; Wilcox et al., Why Marriage Matters; K. A. Moore et al, “Marriage from a Child’s
Perspective: How Does Family Structure Affect Children and What Can We Do about It?” Child Trends
Research Brief, June 2002; Sara McLanahan and Gary Sandefur, Growing Up with a Single Parent: What
Helps, What Hurts (Cambridge, MA: Harvard University Press, 1994).
Page 40
11. Wilcox et al., Why Marriage Matters; David Ribar, “What Do Social Scientists Know about the
Benefits of Marriage? A Review of Quantitative Methodologies,” IZA Discussion Paper # 998 (Bonn,
Germany: Institute for the Study of Labor, January 2004), http://ftp.iza.org/dp998.pdf.
12. Examples of habits, traits, and disadvantages that may lead to negative life outcomes (which
are costly to taxpayers) and to a lack of marriage, to divorce, and to nonmarital childbearing include
not considering the impact of present actions and choices on the future, proclivity to violence, and a
lack of employment skills.
13. See, for example, P. R. Amato and R. A. Maynard, “Decreasing Nonmarital Births and
Strengthening Marriage to Reduce Poverty,” The Future of Children 17, no. 2 (Fall 2007): 75–96.
14. G. S. Becker, A Treatise on the Family (Cambridge, MA: Harvard University Press, 1981) pro-
vides the seminal economic explanation: specialization and exchange and economies of scale. Ribar,
“What Do Social Scientists Know?” provides an extensive review of the empirical literature on these
effects. For example, Ginther and Zavodny suggest that marriage leads to a causal increase in male
labor supply; see D. Ginther and M. Zavodny, “Is the Male Marriage Premium Due to Selection? The
Effect of Shotgun Weddings on the Return to Marriage,” Journal of Population Economics 14 (2001):
313–328.
15. In a forthcoming article in the Journal of Human Resources titled “The Effect of Marital
Breakup on the Income Distribution of Women with Children,” Elizabeth O. Ananat and Guy Michaels
use exogenous variation in the sex of the firstborn child to estimate the impact of divorce on income.
Prior studies have shown that marriages in which the firstborn child is male are less likely to end in
divorce (see, e.g., K. Bedard and O. Deschenes, “Sex Preferences, Marital Dissolution, and the
Economic Status of Women,” Journal of Human Resources 40, no. 2 [Spring 2005]: 411–434).
16. Ananat and Michaels, “The Effects of Marital Breakup,” table 3.
17. Robert Lerman, “The Impact of the Changing U.S. Family Structure on Child Poverty and
Income Inequality” Economica 63 (1996): S119–S139.
18. In “For Richer or for Poorer: Marriage as an Antipoverty Strategy,” Journal of Policy Analysis
and Management 21, no. 4 (2002): 587–599, Adam Thomas and Isabel Sawhill (like Lerman) did not
match all single mothers with husbands. They report that there were more than enough single white
males to marry all single white mothers in the CPS, but there was a shortage of single black males to
marry all single black mothers. In many low-income communities, women probably outnumber mar-
riageable men, because of higher death and incarceration rates of males, meaning it would not be the-
oretically possible to marry all single mothers. They suggest that the lack of marriageable black males
or the reported undercount of minorities in the CPS could be responsible for the dearth of black males
available in the CPS. In addition to black males, there is likely also a shortage of elderly males of all
races eligible to marry elderly females because of higher death rates at younger ages among males.
19. Ribar, “What Do Social Scientists Know?” 38–46.
20. See Hilary Hoynes, Marianne Page, and Ann Stevens, “Poverty in America: Trends and
Explanation,” Journal of Economic Perspectives 20, no. 1 (Winter 2006): 47–68, published by the
American Economic Association; and R. Blank and D. Card, “Poverty, Income Distribution and
Growth: Are They Still Related?” Brookings Papers on Economic Activity 48, no.2 (1993): 285–340, pub-
lished by The Brookings Institution.
21. Thomas and Sawhill, “For Richer or for Poorer.”
22. Robert M. O’Brien and Jean Stockard, “The Cohort-size Same-size Conundrum: An Empirical
Analysis and Assessment Using Homicide Arrest Data from 1960 to 1999,” Journal of Quantitative
Criminology 19 (2003): 1–32.
23. Cynthia Harper and Sara McLanahan, “Father Absence and Youth Incarceration,” Journal of
Research on Adolescence 14, no. 3 (2004): 369–397. Harper and McLanahan do not attempt to control
for unobserved selection effects, which limits our confidence that all of the large differences in risk of
incarceration they found due to family structure are causally related to parents’ marital status.
However, the large number of control variables in their empirical model and the large magnitudes of
their results make it hard to believe that the impact of family fragmentation of boys’ and young men’s
criminal conduct is zero.
24. Robert Sampson, J. Laub, and C. Wimer, “Does Marriage Reduce Crime? A Counterfactual
Approach to Within-Individual Causal Effects,” Criminology 44, no. 3 (2006): 465–504.
25. Paul R. Amato and Alan Booth, A Generation at Risk: Growing Up in an Era of Family
Upheaval (Cambridge: Harvard University Press, 1997), 220. While the theoretical and empirical case
for marriage having a beneficial impact on men, women, and children may be strong, surely in some
cases spouses and children are better off without one parent in the home. For example, a woman and
children may be better off without the father when the father is violent or when the marriage is high-
conflict. In “Until Death Do You Part: The Effects of Unilateral Divorce on Spousal Homicides,”
Page 41
Economic Inquiry 41 (2003): 163–183, T. S. Dee finds that unilateral divorce laws, which were found
to lead to increases in divorce (see L. Friedberg, “Did Unilateral Divorce Raise Divorce Rates?
Evidence from Panel Data,” American Economic Review 88 [1998]: 608–627), have a negligible effect
on the incidence of husbands murdering wives, but unilateral divorce coupled with laws that
favored husbands in the division of marital property led to a 21 percent increase of lethal spousal
violence against husbands. Contrary to Dee’s results, B. Stevenson and J. Wolfers (see “Bargaining
in the Shadow of the Law: Divorce Laws and Family Distress” Quarterly Journal of Economics 121,
no. 1 [2006]: 267–288) find that no-fault divorce led to a large decline in spousal homicide against
wives and no change in spousal homicide against husbands. Dee replicates the results of a 2000
version of Stevenson and Wolfers’ work and finds that their results are sensitive to a variety of
assumptions and specification (see Dee, pp. 177–178). It is unclear whether any changes in their
analysis that were present in Stevenson and Wolfers’ published version in 2006 render Dee’s con-
cerns in 2003 moot, as Stevenson and Wolfers do not directly address these concerns in 2006.
Therefore, there does not seem to be definitive evidence that unilateral divorce laws decrease
spousal homicide against wives. Nevertheless, Wilcox et al., Why Marriage Matters, 31, summarize
evidence that married women are subject to less violence inside and outside the home relative to
single and cohabitating women. In addition, Ananat and Michaels find that divorce benefits some
women with children financially, as they moved in with relatives with significant incomes and/or
the former husband was not contributing anything or very much to family income (see their forth-
coming article “The Effects of Marital Breakup”). Nevertheless, as discussed in the text, Ananat and
Michaels also find that divorce greatly increases the odds that women with children are in the low-
est income quartile. In some individual cases women and children may be better off without the
children’s father because they avoid violence or poverty, but empirical evidence suggests that mar-
riage tends to have the opposite effect by keeping women and children away from violence and
increasing material resources.
26. There are other taxpayer-funded programs that likely experience larger expenditures due to
family fragmentation such the Earned Income Tax Credit, remedial school programs, and special edu-
cation programs. These programs were excluded because we did not feel comfortable making reason-
able cost estimates given the available empirical literature. The likelihood that these programs would
experience reduced costs if more single-adult households became married families suggests that this
estimate of the taxpayer cost of family fragmentation is an underestimate of the true costs. Some state
funds for TANF programs that benefit children are included in the “child welfare” calculation below,
but other non-cash assistance TANF funds are excluded from this analysis.
27. Some evidence suggests that more children in single-parent families are hospitalized for asthma
or childhood diabetes because single parents can be less able to manage the complex stresses of
chronic illness for themselves and their children, and because they have less access to adequate health
care. See Linda Waite and Maggie Gallagher, The Case for Marriage: Why Married People Are Happier,
Healthier, and Better-Off Financially (New York: Doubleday, 2000).
28. Assumption 3 does not imply that only households in poverty are eligible for means-tested
programs. Most federal means-tested programs serve significant numbers of households with incomes
above poverty thresholds. As marriage would reduce poverty by increasing household incomes, mar-
riage would also increase the income of households that already had incomes above poverty thresh-
olds but were receiving means-tested transfers. At least some of these households would be rendered
ineligible for these means-tested transfers due to marriage.
29. Ananat and Michaels, “The Effects of Marital Breakup” and Thomas and Sawhill, “For Richer
or for Poorer.”
30. In table 3, female-headed households refers to all households with a female householder and
no spouse present, including households with and without children. In 2006 there were 12,827,000
children in poverty. Among those, 7,715,000 lived with a single mother. If marriage were to lift 60 per-
cent of these 7,715,000 children out of poverty, then 4,629,000 children would escape poverty. Thus,
marriage would lift 4,629,000/12,827,000 or 36.1 percent of these children out of poverty. In 2006 there
were 36,460,000 total persons in poverty. Among those, 19,257,000 lived with a female householder
with no spouse present or were themselves the female householder with no spouse. If marriage were
to lift 60 percent of these 19,257,000 individuals out of poverty, then 11,554,000 people would escape
poverty. Thus, marriage would lift 11,554,000/36,460,000 or 31.7 percent of these individuals out of
poverty.
31. Based on their estimate of the impact of marriage on the poverty status of female-headed
households, Thomas and Sawhill find that the overall 1998 poverty rate would have been 24 percent
lower if the proportion of children living in female-headed households in 1998 was the same as had
existed in 1970 (see Thomas and Sawhill, “For Richer or for Poorer”).
Page 42
32. Based on table 3, assumption 2 suggests that family fragmentation is responsible for 36.1 per-
cent of childhood poverty, and assumption 3 suggests that family fragmentation is responsible for 36.1
percent of taxpayer costs on these programs.
33. For example, under current rules, if increases in marriage moved some children off the Head
Start rolls (because they are no longer poor), then other children who are eligible but do not currently
receive Head Start services would be admitted into newly freed-up Head Start spaces.
34. Harper and McLanahan, “Father Absence and Youth Incarceration.”
35. Sampson, Laub, and Wimer, “Does Marriage Reduce Crime?”
36. Harry Holzer et al., The Economic Costs of Poverty in the United States: Subsequent Effects of
Children on Growing Up Poor (Washington, DC: Center for American Progress, January 24, 2007),
http://www.americanprogress.org/issues/2007/01/pdf/poverty_report.pdf.
37. See, for example, Ribar, “What Do Social Scientists Know?”
38. See, for example, McLanahan and Sandefur, Growing Up with a Single Parent.
39. Holzer et al., The Economic Costs of Poverty.
40. In the estimates for individual states, we use state-specific average tax rates from
TaxFoundation.org, “Tax Data: State and Local Tax Burdens Compared to Other U.S. States,
1970–2007,” April 4, 2007, http://www.taxfoundation.org/taxdata/show/335.html.
41. The specific calculations for each line item and the data sources used are contained in table
A.1 and its notes. These taxpayer costs can be considered as annual recurring costs under the assump-
tion that current rates of single motherhood remain constant into the future.
42. For example, the U.S. spends almost $500 billion per year on public education. See Thomas
D. Snyder, Mini-Digest of Education Statistics, 2007, NCES 2008-023. (Washington, DC: National Center
for Education Statistics, Institute of Educational Sciences, U.S. Department of Education, 2008).
43. See, for example, Waite and Gallagher, The Case for Marriage; and Elizabeth Marquardt, “The
New Alone,” Washington Post, January 27, 2008, B01.
44. As shown in tables 4–6 based on the 2006 Current Population Survey, single-mother house-
holds with income less than 200 percent of the poverty line are 2.6 times as likely to receive Food
Stamps, 2.9 times as likely to receive cash assistance, and 1.56 times as likely to receive Medicaid than
married couples also earning less than 200 percent of the poverty threshold.
45. Using estimates and calculations from Gregory Acs and Elaine Maag, Irreconcilable
Differences? The Conflict between Marriage Promotion Initiatives for Cohabiting Couples with Children
and Marriage Penalties in Tax and Transfer Programs (Washington, DC: Urban Institute, 2005), we
estimate the $0.5 billion amount.
46. Holzer et al., The Economic Costs of Poverty, 6.
47. Ribar, “What Do Social Scientists Know?”
48. Thomas and Sawhill, “For Richer or for Poorer.”
49. Of course, higher dropout rates would lower future earnings, and government spends con-
siderable resources on attempting to prevent high school students from dropping out and providing
services to help dropouts earn a high school diploma or GED.
50. See Waite and Gallagher, The Case for Marriage.
51. Harper and McLanahan, “Father Absence and Youth Incarceration.”
52. Sampson, Laub, and Wimer, “Does Marriage Reduce Crime?”
53. Harper and McLanahan, “Father Absence and Youth Incarceration,” table 2.
54. Doris J. James, The Profile of Jail Inmates, 2002, Special Report NCJ 201932 (Washington, DC: U.S.
Department of Justice, Bureau of Justice Statistics, July 2004), http://www.ojp.usdoj.gov/bjs/abstract/pji02.htm.
55. Ribar, “What Do Social Scientists Know?” 38.
56. Based on data used by Jens Ludwig in “The Costs of Crime” testimony to the U.S. Senate
Committee on the Judiciary” on September 19, 2006, in the entire U.S., state and local taxpayers spent
about $183 billion on the justice system in FY 2007, while the remaining $49 billion was spent by the
federal government (see http://judiciary.senate.gov/testimony.cfm?id=2068&wit_id=5749).
About the Institute for American Values
The Institute for American Values is a nonprofit, nonpartisan organization devoted to
contributing intellectually to the renewal of family life and the sources of competence,
character, and citizenship. For more information, go to www.americanvalues.org or
(212) 246-3942.
About the Institute for Marriage and Public Policy
The Institute for Marriage and Public Policy is a nonprofit, nonpartisan organization
dedicated to high quality research and public education on ways that law and public
policy can strengthen marriage as a social institution. For more information, go to
www.imapp.org or (202) 216-9430.
About Georgia Family Council
Georgia Family Council is a non-profit organization that works to strengthen and
defend the family in Georgia by equipping marriage advocates, shaping laws, prepar-
ing the next generation and influencing culture. For more information, go to
www.georgiafamily.org or (770) 242-0001.
About Families Northwest
Families Northwest builds strong marriages, families, and communities by advocating
cultural change, offering trusted resources, and training Relationship Champions to help
others. For more information, go to www.familiesnorthwest.org or (425) 679-5671.
For additional copies, please contact:
Institute for American Values
1841 Broadway, Suite 211
New York, NY 10023
Tel: (212) 246-3942
Fax: (212) 541-6665
Email: [email protected]
Web: www.americanvalues.org