Research Proposal
The Effects of Paternal Imprisonment on Children’s Economic Well-Being
Yiyoon Chung University of Wisconsin–Whitewater
The effect of fathers’ incarceration on the well-being of children is an important concern, especially in the United States, a nation with uniquely high incarceration rates as well as a relatively weak and shrinking safety net. This study uses matched, longitudinal, administra- tive data from Wisconsin to estimate the effects of paternal imprisonment on child support and food stamp receipt by families with nonmarital children. The results illustrate the com- plex interactions among public policies. Paternal imprisonment reduces child support re- ceipt and thereby undermines policies designed to improve child support collection. At the same time, increases in food stamp benefits fill a portion of the resulting income gap, pro- viding a safety net for families but increasing welfare expenditures.
In the context of mass imprisonment, particularly the mass imprisonment of low-income males, a subject of great policy concern is the effect of fa- thers’ imprisonment on the child support and welfare benefits received by the families these fathers leave behind. Policies enacted after the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (US Public Law 104-193) attempt to strengthen child support en- forcement and to restrict the welfare benefits available to single-mother families by imposing time limits on benefit receipt, tightening eligibility constraints, and increasing work requirements. An important goal behind these reforms is to increase the role of noncustodial fathers in supporting their children. Unfortunately, widespread paternal imprisonment may di- minish the success of policies designed to improvechildsupport collection and reduce welfare expenditures. It may also create a context in which welfare reform increases the economic vulnerability of many disadvan- taged children. Understanding the ways in which the criminal justice sys- tem interacts with child support and welfare policies thus is critical.
Social Service Review (September 2012). © 2012 by The University of Chicago. All rights reserved. 0037-7961/2012/8603-0004$10.00
This study explores the effects of paternal imprisonment on child sup- port and welfare outcomes. It focuses on children born outside of mar- riage (hereafter, nonmarital children) because of the societal and public policy concern over this group. Nonmarital children accounted for nearly 41 percent of all children born in the United States during 2010 (Ham- ilton, Martin, and Ventura 2011). These children face particularly nota- ble economic hardships (Grall 2009; Magnuson and Votruba-Drzal 2009; McLanahan 2009) and an especially high risk of paternal imprison- ment (Western, Lopoo, and McLanahan 2004; Waller and Swisher 2006). A recent study reports that paternal imprisonment touches approxi- mately 15 percent of the first-born children of unwed mothers in Wiscon- sin by age 5 (Chung 2011). Further, the majority of children with incarcer- ated fathers are nonmarital children (Mumola 2000). Two additional factors heighten the importance of focusing on non-
marital children. First, policy changes in welfare and child support dis- proportionately affect this group of children; second, these children are the focus of the political debates that drive changes in key social policies. Such changes eventually affect a broader group of children and their families (Sparks 2003). This article uses unique data from Wisconsin to estimate the effects of
fathers’ imprisonment on the formal child support and food stamp ben- efits received by the families of 5,550 nonmarital children. Administrative data from the state’s child support enforcement system are matched to date-specific imprisonment records for both children’s biological fathers and the fathers of the children’s half-siblings. This study makes several unique contributions to the literature. First, it
uses detailed records on imprisonment, child support, and welfare to ex- pand on recent work concerning the role of incarceration in reducing fathers’ financial contributions to their families (Cancian, Noyes, et al. 2009; Geller, Garfinkel, and Western 2011). It addresses the interactions among paternal imprisonment, reduced paternal support, and increased welfare receipt by families. Second, the research makes a contribution in suggesting that paternal imprisonment is an important omitted variable in much of the existing literature on the determinants of welfare receipt. Third, this study contributes to the literature on paternal imprisonment by considering the imprisonment of the fathers of focal children’s half- siblings (other male partners of the mothers). This is the first study in the field to conduct this type of analysis, and the innovation is important be- cause a high proportion of unwed mothers have children with more than one partner (Carlson and Furstenberg 2006; Cancian, Meyer, and Cook 2011). Due to reduced support and family resources, the imprisonment of subsequent partners can have negative effects on all the mother’s chil- dren (Berger et al. 2008; Berger and Langton 2011). Finally, the study tes- tifies to the importance of a cross-systems approach. It considers welfare
456 Social Service Review
as a collateral cost of imprisonment and provides empirical evidence of the relationship between paternal imprisonment and welfare. The study also highlights prevalent paternal imprisonment as a key context that can determine the consequences of child support and welfare policies for children and families and that therefore should be considered in the de- sign and planning of these policies.
Literature Review
The current study explores child support and welfare benefits as key com- ponents of the mechanisms through which fathers’ imprisonment affects the economic resources available to unwed mothers and their children. It views welfare participation as a sensible response by low-income mothers to the potential economic shock resulting from the imprisonment of their partners. Although the federal Food Stamp Program is the specific welfare program considered in this study, the theoretical focus falls on the broader welfare system. The discussion of welfare in this section thus is not restricted to food stamps. Reduced child support is a potentially important mechanism through
which high rates of paternal imprisonment may negatively affect many nonmarital children and their families (Sorensen and Zibman 2001; Ha et al. 2008; Cancian, Noyes, et al. 2009; Grall 2009; Nepomnyaschy and Garfinkel 2010). Amid the decline in public support available to unem- ployed single mothers and growing emphasis on child support enforce- ment, child support has become an increasingly important source of in- come for the families of nonmarital children (Cancian and Meyer 2006). If a father would provide child support but imprisonment prevents him from doing so, his incarceration may force a reduction in formal support that places his nonmarital children, already socially and economically dis- advantaged, at an even higher risk of economic hardship.1
However, the extent to which fathers’ imprisonment influences the child support received by nonmarital children’s families is an empirical issue that remains largely unexamined. Paternal imprisonment may have a minimal economic effect on nonmarital children if the incarcerated fathers of these children, fathers who tend to have low economic capac- ities, would have provided negligible child support even without experi- encing imprisonment. Although previous empirical studies show high rates of incarceration among nonpayers (Sorensen and Zibman 2001; Ha et al. 2008), few studies provide information on the trajectories of fathers’
1. That risk may also have negative effects on their short- and long-term emotional, psy- chological, and developmental well-being ( Johnston 1995; Brooks-Gunn and Duncan 1997; Argys et al. 1998; Hagan and Dinovitzer 1999; Travis and Waul 2003; Swisher and Waller 2008; Geller et al. 2009; Poehlmann 2009; Western and Wildeman 2009; Wakefield and Ug- gen 2010; Geller et al. 2011).
Paternal Imprisonment 457
child support payments prior to, during, and following incarceration, or otherwise estimate the effects of fathers’ incarceration on child support. Expanding upon several recent studies that begin to explore analytically the connection between incarceration and fathers’ financial contribu- tions (Cancian, Noyes, et al. 2009; Geller et al. 2011), the current study fo- cuses on the formal child support received by families with incarcerated fathers and analyzes the implications of reduced paternal support for wel- fare receipt. Paternal imprisonment’s influence on the welfare benefits received by
families is of interest from at least two different, possibly even incompat- ible, viewpoints. If welfare is intended to serve as a vital safety net, an im- portant question is whether children impoverished because of paternal imprisonment are then served by the welfare system. Conversely, if welfare participation is associated with long-term harm to children, or if the focus is on a negative incentive effect (i.e., that welfare discourages mothers’ work), increased welfare reliance may be an important, although uninten- tional, negative consequence of the father’s imprisonment. Regardless of which perspective is adopted, it is important for scholars and policy mak- ers to consider the possible role of paternal imprisonment in increasing welfare expenditures. The current study also may aid criminal justice researchers who seek
to assess the external costs of incarceration. Cost-benefit analyses of incar- ceration usually do not consider the relationship between incarceration and welfare (Lengyel 2006; Ziebert 2006; Vining and Weimer 2010). If incarceration increases welfare reliance among the families left behind, then the external costs of incarceration for taxpayers may be underestimated. Little empirical research relates paternal incarceration to families’ wel-
fare use, although many studies trace welfare participation to such factors as individual-level socioeconomic and demographic characteristics, alco- hol and drug abuse, national and local economic opportunities, welfare pol- icy features, and child support system attributes (Plotnick 1983; McLanahan 1988; Moffitt 1992; Meyer 1993; Blank and Ruggles 1996; Gottschalk 1996; Miller and Sanders 1997; Hoynes 2000; Klawitter, Plotnick, and Ed- wards 2000; Schmidt and McCarty 2000; Gundersen and Oliveira 2001; Huang, Kunz, and Garfinkel 2002; Currie 2003; Grogger 2004; Grogger and Karoly 2005; Teitler, Reichman, and Nepomnyaschy 2007; Ribar, Edelhoch, and Liu 2008). If the causal variables of interest are associated with fathers’ imprisonment, and if imprisonment influences welfare re- ceipt, then estimates of the causal effects on welfare likely suffer from omitted variable bias. Methodological difficulties impede research into the effects of pater-
nal imprisonment on child support and food stamp receipt by nonmar- ital children’s families. For example, representative data on nonmarital
458 Social Service Review
children’s fathers are scarce, and most surveys provide inaccurate or in- sufficient information on fathers’ incarceration ( Johnston 1995; John- son and Waldfogel 2002). Much research focuses on the use of child sup- port enforcement as a way to avoid welfare expenditures (Barnow et al. 2000; Huang, Garfinkel, and Waldfogel 2004). However, scholarship rarely discusses the ways in which imprisonment can simultaneously affect child support and welfare benefits; these issues merit further research.
Theoretical Framework
Paternal Imprisonment and Child Support Received by Families
Previous research suggests that paternal imprisonment’s effects on child support payments can be modeled using a framework that includes three mechanisms: (a) fathers’ ability to pay, (b) fathers’ willingness to pay, and (c) child support enforcement (Weiss and Willis 1985; Beller and Gra- ham 1993; Graham and Beller 2002; Bartfeld and Meyer 2003; Cancian, Meyer, and Caspar 2008; Meyer, Ha, and Hu 2008). As the first mecha- nism suggests, fathers’ imprisonment may negatively affect child support payments because most incarcerated fathers have little or no earnings and generally cannot provide financial support for their children. Also, employment levels, earnings, and earnings growth after release tend to be lower than before incarceration because incarceration reduces hu- man and social capital, stigmatizes the individual, and prevents access to career-type jobs (Nagin and Waldfogel 1998; Western 2002, 2006; Pettit and Western 2004). The second mechanism, fathers’ willingness to pay support, may also
play a role in paternal imprisonment’s relationship with child support; imprisonment may reduce child support by reducing fathers’ willingness to pay. For example, imprisonment may weaken family bonds, limiting fathers’ involvement in their children’s lives during imprisonment and also after release (Western et al. 2004; Lopoo and Western 2005; Geller et al. 2011). Less directly, fathers tend to accrue large child support debts during imprisonment, and such debts may reduce their willingness to make child support payments if they have little hope of paying off the debts (Noyes 2006; Levingston and Turetsky 2007; Cancian, Heinrich, and Chung 2009; Harris, Evans, and Beckett 2010). The third mechanism, child support enforcement, may affect impris-
onment’s relationship with support if incarceration initiates or deep- ens the involvement of the formal child support enforcement system in fathers’ provision of child support to their children. Several factors suggest that this mechanism may cause fathers’ imprisonment to have positive effects on their child support payments. In many states, the criminal justice system cooperates with the child support enforcement
Paternal Imprisonment 459
system. Although no estimates are available, anecdotal evidence sug- gests that imprisonment sometimes leads to a new child support order if no prior order exists and to the enforcement of the support order following the father’s release (Noyes 2006).2 Additionally, supervision by the criminal justice system (e.g., monitoring child support compli- ance as part of parole arrangements) may lead some fathers to work in the formal labor market, and child support is automatically deducted from paychecks. If such supervision forces a father into the formal la- bor market, that participation can be expected to increase his child sup- port payments. Moreover, many states have criminalized the nonpayment of child support over the last few decades and made nonpayment a pa- role violation. The threat of incarceration (or reincarceration) for child support nonpayment can extract hidden money from fathers and thus may also increase some fathers’ child support payments (Swisher and Waller 2008). However, if punitive child support enforcement tools, such as revoca-
tion of driving and professional licenses and seizure of property, are uti- lized for incarcerated or recently released fathers and make it hard for these fathers to work, this third mechanism may result in a reduction in their child support payments. In sum, although paternal imprisonment leads in some cases to the initiation or reinforcement of child support en- forcement and the threat of imprisonment for nonpayment of child sup- port can increase fathers’ child support payments, most evidence on the three mechanisms suggests that imprisonment decreases child support.
Paternal Imprisonment and Food Stamp Benefits Received by Families
In estimating the effects of fathers’ imprisonment on welfare benefits, the current analysis focuses specifically on the Food Stamp Program, which distributes means-tested benefits that are almost as helpful as cash for low-income families.3 The program stands in this study as an emblematic example of welfare, and food stamp benefits serve as the focal welfare var- iable here because food stamp benefits are more accessible than Tempo- rary Assistance for Needy Families (TANF) cash benefits. So, too, food stamp benefits are more stable over time and across states, and they are less influenced by local political contexts that may also affect incarcera-
2. In certain cases, an increase in child support payments does not represent a change in the overall economic contributions of a father to his children. There may be cases in which imprisonment ends the relationship between cohabiting parents; formal child support pay- ments therefore increase, but overall economic support remains the same or declines. 3. In 2008, the Food, Conservation, and Energy Act (122 Stat. 923) renamed the Food
Stamp Program as the Supplemental Nutrition Assistance Program. The current study uses the program’s former name because the study’s data come from the period prior to the name change.
460 Social Service Review
tion rates. Food stamps account for a higher proportion of the total aid provided to low-income families (Beckett and Western 2001; Wolkwitz 2007; Scholz, Moffitt, and Cowan 2009). The Food Stamp Program is a federally funded effort to help low-
income households obtain a nutritious diet by increasing their food pur- chasing power (US Department of Agriculture 2010). The program is means-tested. To be eligible, an applicant is required to have a gross in- come below 130 percent of the federal poverty line and a net income be- low 100 percent of that amount. An applicant’s countable assets must also be below a specified amount. Because the program imposes few non- financial eligibility criteria and has much more lenient work requirements than TANF, it is considered the only income assistance program that po- tentially serves all financially vulnerable households in the United States (Issar 2010). Economic theory suggests that fathers’ imprisonment and the subse-
quent reductions in the formal support to unwed mothers will increase the food stamp benefits received by mothers (Becker and Barro 1988; Becker 1991; Currie 2003; Kang, Huffman, and Jensen 2004). If impris- onment reduces the economic support fathers provide to their families, the reduction is a negative shock to mothers’ overall income. This shock may be associated with an increase in mothers’ earnings, although the in- crease may not completely compensate for the income loss. The resulting total income reduction could change mothers’ food stamp eligibility and benefit levels. Some mothers who already receive food stamp benefits could become eligible for higher levels of benefits; others whose newly reduced income now falls below the cutoff for eligibility may pursue food stamp participation. Although welfare may be irrelevant for mothers with high wages, empirical studies show that it is not irrelevant for a majority of unwed mothers (McLanahan 2009).
Data and Methods
The underlying hypothesis in this study is that paternal imprisonment reduces formal child support payments, and, in turn, this reduction and other economic stress associated with paternal imprisonment increase the amount of food stamps received by custodial mothers. To test this hypothesis, the analysis follows the families of a sample of first-born non- marital children (focal children). The sample is limited to children whose fathers were not in prison during the first 3 years of the child’s life. By im- plementing this restriction and controlling for some economic outcomes during the prenatal period and in the first 3 years of the focal children’s lives, the analysis estimates the effects of paternal imprisonment in the fourth year after the focal child’s birth (or year 4) on subsequent economic outcomes (i.e., those in year 5).
Paternal Imprisonment 461
Data
The study draws on a unique, matched set of longitudinal administrative data from Wisconsin. The data include detailed information about pater- nity and multiple-partner fertility from the administrative records of the state’s child support enforcement system (the Kids Information Data Sys- tem, or KIDS). Also included is such demographic information as the child’s race and date of birth. The data are matched with date-specific im- prisonment records from the Wisconsin Department of Corrections for 1997–2007. The resulting data set is used to determine the imprisonment status of the nonmarital children’s biological fathers and of the fathers of the focal children’s half-siblings (i.e., those who share mothers with focal children). The data are also matched with data from the Client Assistance for Re-employment and Economic Support system to identify mothers’ par- ticipation in the Food Stamp Program and TANF between 1996 and 2008. Finally, data from the Unemployment Insurance system are matched with the data set to obtain information about mothers’ and fathers’ formal earn- ings between 1996 and 2008. The matched administrative data offer some important advantages
for studying the effects of paternal imprisonment on the economic out- comes of nonmarital children’s families. First, the data contain fairly complete information on fathers of nonmarital children, including in- formation about fathers’ imprisonment statuses (Cancian et al. 2011). The high representation of nonmarital births in KIDS is achieved in part because KIDS contains information about families that contact the child support agency for help establishing paternity, locating a non- custodial parent, establishing a child support order, or collecting support. But further, the high representation is achieved because the system also includes information about families that apply for welfare (because their welfare cases must be referred to the child support agency). Build- ing on research that compares nonmarital cases included in KIDS with birth records (Brown and Cook 2008) and on other related studies (Cancian et al. 2011; Chung 2011), the current analysis estimates that the KIDS data include information on approximately 77 percent of the biological fathers partnered to Wisconsin mothers whose first-born non- marital child was born between 1998 and 2002. The coverage of non- marital children, although not perfect, is quite good compared with the coverage in social surveys. Moreover, the administrative data provide relatively accurate information about fathers’ imprisonment status, and there is relatively little attrition over time. A final advantage of the data is the ability to identify the imprisonment of all fathers tied to a mother who has children with more than one partner. By contrast, the Fragile Families and Child Wellbeing Study, for example, provides information about multiple-partner fertility and the imprisonment of the biologi- cal fathers of the children being studied but lacks information about the imprisonment of fathers of the children’s half-siblings.
462 Social Service Review
Samples
This study begins with a stratified sample of 8,720 mothers who were not married when they gave birth to their first child in Wisconsin between 1998 and 2002. The sample is stratified to overrepresent mothers of chil- dren with incarcerated fathers; when appropriately weighted, the sample is representative of mothers who were not married at the time of their first birth, which occurred during the time considered. The final sample is further restricted to 5,550 mothers of focal children whose biological father did not experience imprisonment before the child reached age 3 and whose half-siblings’ father had no experience of it before the focal child reached that age. All cases meet the following selection criteria: the mother was the cus-
todial parent for 5 years following the focal child’s birth, the child’s date of birth is known, the biological father is known, and the records estab- lish the focal child’s paternity within 5 years of the child’s birth. Unless otherwise noted, all results are weighted according to the sampling scheme. The final base sample consists of 5,550 mother–focal child pairs that are associated with 5,232 biological fathers of focal children; among these fathers, 292 (5.5 percent) are the biological fathers of more than one focal child in the sample (standard errors are adjusted accordingly). The base analysis includes 46 fathers (when the sample is weighted, it
includes 6.5 percent of those imprisoned in year 4) whose imprisonment is associated with the nonpayment of child support (in half of these cases, child support nonpayment is the only indicated offense). Because impris- onment for nonpayment of child support may be endogenous to the fac- tors that affect fathers’ prior financial contributions and the varying char- acteristics of child support enforcement policies, these 46 cases impose a methodological challenge in estimating the effect of imprisonment on child support. In an alternative analysis using a sample that excludes these 46 cases, the point estimates of key coefficients are quite similar to the base estimates (see table A1 in the appendix, test 8).
Key Measures
In the analysis, the baseline is the birth of the focal children, and out- comes are measured against this baseline. A binary variable is created to measure fathers’ imprisonment in year 4; the variable equals one if any of the partners with whom the mother had children was imprisoned any time during the fourth year following the focal child’s birth, and zero oth- erwise. The focus is on paternal imprisonment in year 4, rather than im- prisonment in earlier years, because child support is one of the key de- pendent variables, and establishing child support orders takes time. The choice to focus on fathers’ imprisonment in year 4 reinforces the im- portance of considering the imprisonment of the fathers of focal chil-
Paternal Imprisonment 463
dren’s half-siblings because the number of focal children who share a mother with a half-sibling increases over time (Cancian et al. 2011). Monthly variables are constructed to measure annual child support
and food stamp benefits; the calendar month of the focal child’s birth is counted as the first month. Then, annual outcomes are constructed by summing figures for the pertinent 12 months. Information about formal earnings comes from the Unemployment Insurance system but is avail- able only as quarterly data. To create figures for the time shortly after the focal child’s birth, monthly earnings are calculated by dividing quarterly earnings by three, and then annual variables are calculated using these monthly figures. Because any child support received by the mother af- fects the well-being of the mother and all her children, and because the primary analysis considers the imprisonment of both biological fathers and fathers of the focal children’s half-siblings, the analysis includes all formal child support received by mothers. This support includes payments from all partners with whom the mothers had children. All financial data are adjusted for inflation to December 2008 dollars using the Consumer Price Index for All Urban Consumers. Information about the county in which each child resides comes from
KIDS. Following the strategy used by Maria Cancian and colleagues (2011), this study analyzes the counties of residence in three categories: Milwaukee County, other urban counties (the 24 counties that are part of Metropolitan Statistical Areas), and rural counties (all other counties). Milwaukee is categorized as a separate urban county because the risk of paternal imprisonment is particularly high there (Chung 2011) and be- cause of its unique demographic characteristics. In 2010, blacks consti- tuted only 6.3 percent of the total state population but nearly 27 percent of Milwaukee residents (US Census Bureau 2011). About one-third of nonmarital births in Wisconsin occur in Milwaukee County (Cancian et al. 2011).
Analytic Strategies
Identifying the causal effects of paternal imprisonment on families’ eco- nomic outcomes is challenging. In part, this is because characteristics of imprisoned fathers are correlated with the characteristics of their chil- dren’s mothers. Mothers do not randomly select partners with differing propensities for imprisonment. For example, mothers with less human capital may receive less child support, collect higher welfare benefits, and be more likely to be poor; they may also be more likely to have children with a father who becomes imprisoned or to have a child with one who faces a longer sentence. Therefore, observed differences in economic out- comes across groups of mothers with different experiences of a partner’s imprisonment may not be due solely to the partner’s imprisonment. Iden- tifying causal effects also is challenging because certain patterns emerge
464 Social Service Review
over time in families’ economic outcomes. For example, average child sup- port payments and food stamp benefits tend to increase over time, at least in the first few years after the child’s birth. Thus, one cannot presume that all differences in families’ economic outcomes before and after fathers’ imprisonment are attributable to fathers’ imprisonment. Therefore, to estimate the effects of fathers’ imprisonment on their
families’ economic outcomes, this study employs various empirical strat- egies. The robustness of the results is tested by examining different as- sumptions and alternative specifications. The study first completes bivar- iate analysis and then employs difference-in-difference (DD) methods. Fixed effects models are utilized to test the sensitivity of results. Primary multivariate analysis: DD methods.—The primary analysis em-
ploys a DD method in which outcomes are compared for two groups in two time periods. Specifically, the difference between year-2 and year-5 economic outcomes for children with fathers imprisoned in year 4 is compared to the difference for those with fathers who were not impris- oned in year 4. Economic outcomes are measured for years 2 and 5, in- stead of years 3 and 5, because imprisonment in year 4 may be associated with such year-3 events as arrest, detention, and trial. Economic out- comes in years 2 and 5 represent pre- and postimprisonment economic outcomes. This method helps to remove biases (in comparisons of eco- nomic outcomes) that could be a result of permanent differences be- tween the two groups as well as biases stemming from economic outcome comparisons over time among families that experienced paternal impris- onment in year 4 (Wooldridge 2007). Separate models are run for the annual amounts of child support and food stamps received by families. The primary analysis includes 11,100 observations (economic outcomes for 5,550 mothers observed in two time periods). A generalized estimat- ing equation is employed to conduct the analysis while adjusting the standard errors to account for cases in which a father was the biological father of more than one focal child in the sample (i.e., cases in which the outcome variable of more than one observation is potentially corre- lated); to implement the generalized estimating equation, SAS statisti- cal software’s “procgenmod” procedure is used. Equation (1) shows the underlying DD models for child support and
food stamps:
y j it 5a 1 a1dt 1 a2d
j 1 a3d j t 1 X
j it′d 1 ε
j it ; ð1Þ
where y j it is the annual amount of child support or food stamps received
by mother i (from 1 to 5,550) in group j in period t; j is 1 if the family ex- perienced paternal imprisonment in year 4 and 0 otherwise; t is 0 for year 2 (the first time period considered) and 1 for year 5 (the second time pe- riod considered); dt is 1 if t equals 1 and 0 otherwise; d
j is 1 if j equals 1 and 0 otherwise; dj
t is 1 if j equals 1 and t equals 1, and 0 otherwise; X
j it′ is a vec-
Paternal Imprisonment 465
tor of control variables for mother i, group j, and period t′ (detailed ex- planations below, including specified t′ for variables); d is the DD esti- mate for the effect of paternal imprisonment on child support or food stamps (the coefficient of key interest in the study); a; a1; a2, and a3 are other coefficient vectors; and ε
j it is a random error term.
The variables represented by function X deserve further discussion. The analysis controls for a range of variables concerning the focal chil- dren’s mothers and biological fathers (biological fathers’ characteristics are used as proxies for the characteristics of the mothers’ partners) to re- duce the remaining bias and to improve the efficiency of estimates by re- ducing the residual variance (Meyer 1995; Wooldridge 2007). The child support and food stamp models use similar sets of control variables.4 The time-constant control variables used across models include the annual earnings of both the mother and the focal father in year 1; the races of the mother and the focal father; the ages of the mother and the focal fa- ther; the county of the focal child’s residence; the year of the focal child’s birth; whether the mother received any TANF cash benefits in year 1; and whether the mother received any food stamp benefits from 1 year before the focal child’s birth until 1 year after it. In addition, the child support models include control variables indicating the establishment of the fo- cal father’s paternity, his child support orders, and his child support pay- ments in year 1. The food stamp models do not include such controls (see table A1, test 3, for an alternative food stamp analysis that also in- cludes these control variables; results remain robust). Time-varying controls used across models include the total number of
children and whether the mother had any other child with any other part- ner. These variables are measured at the end of year 1 and year 3 (points that represent the beginning and end of the preimprisonment period). The controls correspond to those used for the first and second time pe- riods. The study also controls for variables related to mothers’ fertility. These are measured at the end of year 1 and year 3, instead of year 1 and year 4, because these outcomes may be influenced by fathers’ imprison- ment during year 4. The primary analytic strategy, the DD method, has limitations. If time-
varying events affect those whose partners experienced imprisonment differently than they affect those whose partners did not, DD estimates may be biased. Therefore, an alternative DD model is conducted that in- cludes time-varying measures of formal earnings for mothers and fathers. To the extent that mothers’ or fathers’ earnings are also associated with these events, the measures should account for some of the influence of these time-varying events (see table A1, test 7). In addition, fixed effects
4. In reduced-form models, particularly those using administrative data when a relatively limited number of variables are available and most of the variables are utilized (as in this study), the approach of using a similar set of control variables in models for child support and food stamp benefits follows the tradition of the literature (Cancian et al. 2008).
466 Social Service Review
methods (discussed further below) are used to test the sensitivity of re- sults. Multivariate sensitivity analysis: fixed effects methods.—The advantage of
fixed effects methods lies in their ability to control for all time-constant characteristics of individuals, whether the characteristics are observed or not (Allison 2005). The model serves to examine the robustness of the DD results, although a fixed effects analysis cannot account for unobserved changes that occur in family circumstances over time and that affect both fathers’ imprisonment and changes in the family’s economic outcomes (e.g., job loss or shifts in relationship status). Specifically, this alternative analysis employs the change score method (Allison 2005, 253; Morgan and Winship 2007); the dependent variable used in this model captures differences in economic outcomes between year 2 and year 5. Difference scores are used for all time-varying control variables. Consistent with the base DD models, the effects of paternal imprisonment are modeled sep- arately for annual child support and food stamp benefits.
Characteristics of Wisconsin Cases and Implications for Generalizability
Because the data used here come from a single state, knowing the char- acteristics of Wisconsin cases is essential for understanding the implica- tions of these results for other regions. Some of the state’s demographic and economic characteristics are similar to characteristics in other re- gions; however, the population is somewhat less urban and less racially and ethnically diverse than the populations of other states (Meyer et al. 2008; US Census Bureau 2010). Although information is available on the risk of paternal imprisonment among nonmarital children in Wisconsin (Chung 2011), the author knows of no national estimate of the paternal imprisonment risk among nonmarital children. Although imprisonment rates in Wisconsin are generally lower than in national estimates, the im- plications of this comparison for nonmarital children’s fathers are un- clear. In part, this is because Wisconsin’s racial disparity in imprisonment is notable (compared with national estimates, the state’s imprisonment rates are higher among blacks and lower among whites; Oliver and Yo- com 2002) and because nonmarital children are more likely to have fathers in racial minority groups than are marital children (West and Sa- bol 2009). Evidence suggests that Wisconsin’s child support enforcement system is more effective than systems in many other states (Sorensen and Zibman 2001; Council of State Governments 2009). In states where child support enforcement is less effective, fathers’ imprisonment may cause a smaller reduction in child support payments; for example, fathers in such states may pay less child support prior to imprisonment than they would in states with effective enforcement. The full pass-through policy in Wisconsin’s child support system de-
serves consideration. To cover the cost of providing welfare benefits, most
Paternal Imprisonment 467
states keep all child support paid on behalf of children on welfare. Wis- consin, in contrast, has a full pass-through policy; all formal support col- lected for Wisconsin children on welfare passes through the system to the child’s caregiver. Wisconsin implemented the policy between 1997 and 2006, so most of the mothers in this analysis received all child support pay- ments made on behalf of their children during the study period, regard- less of whether or not they received welfare (Cancian et al. 2008). This dif- ference in child support policies has implications for studies in states without pass-through policies. First, under a no-pass-through policy, mothers who received food stamp benefits prior to fathers’ imprisonment have no child support to lose if a noncustodial father is imprisoned, be- cause the state retained all child support paid by the fathers prior to im- prisonment. Second, if a father pays support prior to imprisonment, the family has less incentive to participate in welfare than it would under a policy by which the state retains none of the support paid, because under a no-pass-through policy, obtaining welfare benefits means losing child support; therefore, studies of states without a pass-through policy may find greater increases in food stamp benefits following the fathers’ impris- onment than does the current study. Overall, Wisconsin’s specific characteristics have complex implica-
tions for the results, and it is not possible to determine the ultimate di- rection of any bias. Although the data used in this study are limited to Wisconsin, and thus there are limits on the extent to which these results can be generalized to nonmarital children’s families in other states, a case study of Wisconsin may be particularly relevant for policy making because the state is an important experimental site for national public policies.
Results
Sample Characteristics
Table 1 includes summary data on the characteristics of the primary sam- ple. Results suggest that approximately 3 percent of mothers experienced the imprisonment of at least one partner in year 4. Estimates (not shown) indicate that 14 percent of these mothers reportedly experienced the im- prisonment of a subsequent partner (or partners) but not the imprison- ment of the focal child’s biological father; 1 percent reportedly experi- enced the imprisonment of both the focal child’s father and a subsequent partner (or partners). Without this study’s unique data on the imprisonment of the fathers of children’s half-siblings, the 14 percent of mothers who experienced a partner’s (but not focal partner’s) impris- onment in year 4 would not be identified as having experienced a part- ner’s imprisonment. Nearly 24 percent of the mothers in the sample had additional children by the time the focal child turned 4; about 14 percent
468 Social Service Review
reportedly had additional children only by the same father; over 9 per- cent had one or more children with one or more additional partners. (In the fifth year, 71 percent of sampled mothers are found to receive some child support and 53 percent are found to receive some food stamp ben- efits; see table A2 in the appendix.) Sampled mothers are generally very young: over 60 percent were youn-
ger than age 20 at the time of the focal child’s birth (table 1). The focal children’s biological fathers are, on average, somewhat older than the mothers, but the age difference between the mother and the focal father is reportedly less than 2 years in about 47 percent of the cases. Of the sam- pled mothers, 62 percent are white, 22 percent are black, and 5 percent are Hispanic; racial information is missing for 9 percent of the sample. Estimates that examine the parents’ races in combination suggest that 44 percent of sampled couples include two white parents and 20 percent include two black parents. As table 1 shows, most mothers (about 83 percent) had formal (re-
corded) earnings in year 1 (the rate remains at over 80 percent in the first 6 years following the birth; see appendix table A2). In year 1, nearly 70 percent of mothers had formal annual earnings under $10,000. Focal fathers’ recorded annual earnings are generally higher than those of mothers; about 55 percent of fathers had formal earnings under $10,000 in year 1. However, the proportion of fathers who had no formal earn- ings in that year is somewhat higher than the proportion of mothers without reported earnings. The data indicate that 47 percent of mothers received food stamp benefits between 1 year prior to the focal child’s birth and 1 year after it; if TANF benefits received in year 1 are included, 51 percent of mothers are estimated to receive benefits from one or both programs.
Bivariate Analysis
Table 2 includes the results of simple bivariate analysis for annual child support and food stamp benefits received by the two groups, which dif- fer by experience with partner imprisonment. The first group includes mothers who experienced the imprisonment of any partner in year 4, and the second includes those who did not (by construction, neither group experienced partner imprisonment between the time of the focal child’s birth and the child’s third year). For families experiencing pa- ternal imprisonment in year 4, the annual amounts of child support received in years 4 and 5 are substantially lower than that received in year 3. In contrast, there is a smooth increase in child support for families that did not experience paternal imprisonment. This pattern is suggestive of the effects of paternal imprisonment on child support. However, the pattern for food stamp benefits is less clear.
Paternal Imprisonment 469
Table 1
DESCRIPTION OF SAMPLE CHARACTERISTICS
Variable t = 1 t = 2 t = 3
Mother with any imprisoned partner in year 4 2.94 Mother’s family characteristics at year t after focal
child’s birth:*
Any subsequent children 1.06 18.01 23.58 Any children by same father; no children by another partner .86 11.97 14.19
With one subsequent partner .20 6.01 9.16 With more than one subsequent partner .00 .03 .23
Focal father’s paternity, child support order, and payment in year t:y
No paternity established by end of year t 31.38 6.94 2.40 Paternity established by year t, no support orders in year t 25.05 11.91 7.97
Paternity established by year t, child support owed; no payment made in year t 10.02 17.31 21.69
Paternity established by year t, child support owed; payments made in year t 33.56 63.84 67.94
Age of mother at focal child’s birth: Under 18 19.13 18–20 42.00 21–23 21.78 24–27 9.03 28–39 7.79 ≥ 40 .27
Age of focal father at focal child’s birth: Under 18 8.36 18–20 29.32 1–23 25.94 24–27 17.67 28–39 16.52 ≥ 40 2.19
Age of focal father in relation to mother’s age: 101 years younger .32 5–9 years younger 1.40 2–4 years younger 4.88 Same age (± 2 years) 46.66 2–4 years older 26.28 5–9 years older 14.63 ≥ 10 years older 5.83
Race of mother: Black 21.94 White 61.65 Hispanic 4.57 Otherz 2.79 Missing 9.05
Race of focal father: Black 28.88 White 50.87 Hispanic 7.23 Other 2.98 Missing 10.04
Multivariate Analysis: DD Analysis
Table 3 shows the results of the DD analysis. These models provide a for- mal test of whether the two groups (mothers with imprisoned partners and those without) differ with respect to the change in economic out- comes over time (from year 2 to year 5), and the models control for other covariates. The annual amounts of formal child support and food stamps are estimated separately.
Table 1 (Continued )
Variable t = 1 t = 2 t = 3
Both black 20.10 Both white 44.44 Both Hispanic 2.70 Mother white, father black 6.47 Mother white, father Hispanic 3.63 All other combinations 6.74 Either unknown 15.92
Mother’s formal earnings in year 1: Zero quarters of earnings 16.96 $1–4,999 34.29 $5,000–9,999 18.74 $10,000–19,999 20.28 ≥ $20,000 9.73
Focal father’s formal earnings in year 1: Zero quarters of earnings 20.73 $1–4,999 21.45 $5,000–9,999 12.94 $10,000–19,999 18.90 ≥ $20,000 25.98
County of focal child’s residence: Milwaukee 28.94 Urban counties 46.66 Rural counties 24.40
Year of focal child’s birth: 1998 19.92 1999 19.99 2000 20.58 2001 20.63 2002 18.88
Any food stamp receipt from 1 year before focal child’s birth to 1 year after 46.88
Any TANF cash benefit receipt in year 1 24.45 Any food stamps or TANF cash benefits during time
considered 50.50
Source.—Wisconsin administrative data. Note.—Results are percentages. N 5 5,550 focal child–mother pairs. Baseline is the birth
of the mother’s first-born nonmarital child; outcomes are measured against this baseline, and weights are applied. All monetary amounts are adjusted to December 2008 dollars.
* t refers to the time elapsed (in years) following the baseline. y The categories of the variable were constructed hierarchically in the order listed. z Other includes Asian Americans, Pacific Islanders, and Native Americans.
Paternal Imprisonment 471
T a b le
2
A N N U A L E C O N O M IC
O U T C O M E S B Y P A T E R N A L IM
P R IS O N M E N T E X P E R IE N C E
IN YE
A R 4
A n n u a l C h il d Su
pp o r t R ec
ei v ed
fr o m
A ll
F a t h er
s A n n u a l F o o d St
a m ps
R ec
ei v ed
A ll M o th e rs
M o th e rs
w it h
A n y IP
in Ye a r 4
M o th e rs
w it h
N o IP
in Ye a r 4
A ll M o th e rs
M o th e rs
w it h
A n y IP
in Ye a r 4
M o th e rs
w it h
N o IP
in Ye ar
4
M e a n $
% o f
A n y
M e a n $
% o f
A n y
M e a n $
% o f
A n y
M e a n $
% o f
A n y
M e a n $
% o f
A n y
M e a n $
% o f
A n y
Ye a r 1
5 0 0
3 4
1 4 1 * *
2 1 * *
5 1 1
3 5
8 2 4
4 4
1 ,5 4 0 * *
6 5 * *
8 0 2
4 3
Ye a r 2
1 ,4 0 4
5 7
4 7 7 * *
4 0 * *
1 ,4 3 2
5 7
8 7 1
4 8
1 ,5 9 8 * *
7 0 * *
8 4 9
4 7
Ye a r 3
1 ,7 1 8
6 5
5 8 6 * *
4 9 * *
1 ,7 5 2
6 6
1 ,0 2 9
5 1
1 ,8 8 0 * *
7 4 * *
1 ,0 0 4
5 0
Ye a r 4
1 ,9 1 6
6 9
4 3 7 * *
4 5 * *
1 ,9 6 1
7 0
1 ,1 5 1
5 3
2 ,1 9 4 * *
7 8 * *
1 ,1 1 9
5 2
Ye a r 5
2 ,0 2 8
7 1
5 3 7 * *
5 1 * *
2 ,0 7 3
7 2
1 ,2 1 7
5 3
2 ,2 7 9 * *
7 8 * *
1 ,1 8 5
5 2
U W N
5 ,5 5 0
7 2 0
4 ,8 3 0
5 ,5 5 0
7 2 0
4 ,8 3 0
W N
5 ,5 5 0
1 6 3
5 ,3 8 7
5 ,5 5 0
1 6 3
5 ,3 8 7
So u r c e. — W is co
n si n a d m in is tr a ti ve
d a ta .
N o t e. — IP
5 im
p ri so n e d p a rt n e r; U W N
5 u n w e ig h te d n u m b e r o f o b se rv a ti o n s; W N
5 w e ig h te d n u m b e r o f o b se rv a ti o n s. N 5
5 ,5 5 0 fo ca l ch
il d – m o th e r
p a ir s. B a se li n e is th e b ir th
o f m o th e r’ s fi rs t- b o rn
n o n m a ri ta l ch
il d ; o u tc o m e s a re
m e a su re d a g a in st th is b a se li n e , a n d w e ig h ts a re
a p p li e d . M o n e ta ry
a m o u n ts
a re
a d ju st e d to
D e ce m b e r 2 0 0 8 d o ll a rs .
* * R e su lt s d if fe r st a ti st ic a ll y si g n ifi ca n tl y fr o m
th o se
fo r m o th e rs
w it h n o im
p ri so n e d p a rt n e rs
(p < .0 1 ).
The terms for the time × imprisonment interaction represent estimates of the effects of fathers’ imprisonment. The models control for the influ- ence of time and preexisting group differences in economic outcomes. The results suggest that paternal imprisonment reduces child support by $701 per year and increases food stamp benefits by $208 per year. The coefficients for time, which covers mothers with and without an impri- soned partner, stem from comparison of economic outcomes in year 5 (the postimprisonment period) with those in year 2 (the preimprison- ment period). The estimates suggest that mothers received $563 more in child support in year 5 and $181 more in food stamps in that year. The im- prisonment group coefficient represents the time-invariant difference in overall means for economic outcomes between families that experienced a father’s imprisonment and those that did not. Compared with families that did not experience a father’s imprisonment in year 4, those that ex- perienced a father’s imprisonment in that year are estimated to receive $221 less in child support and $233 more in food stamps. The estimates for other parameters in the models (table 3) are gen-
erally consistent with prior research. The total child support received by families is estimated to be higher for mothers with multiple partners than for those with only one; it is estimated to be higher for cases in which both the mother and the focal father are white than for cases in which both the mother and the focal father are black or Hispanic. Sup- port is positively correlated with mother’s earnings. Finally, the total child support received by families is positively associated with establish- ment of the focal father’s paternity in year 1, with a formal order to pay support, and with the focal father’s payment of support in that year. A possible explanation for the positive association of mother’s earnings with child support received may be that mothers with higher earnings tend to partner with males who have higher earning capacities and, there- fore, a higher ability to make payments. However, there is mixed evidence concerning the relationship of father’s earnings with the child support received by mothers, and some of the observed relationships are less in- tuitive. For example, compared to families with focal fathers who had no formal earnings in year 1, families with focal fathers who had formal earn- ings of $5,000–$9,999 are estimated to receive less in child support payments. This somewhat unexpected relationship may occur because fathers with no formal earnings are a heterogeneous group in terms of their economic situation; some may have earnings that are not captured in this study’s data. Such earnings may come from self-employment, the informal labor market, and out-of-state jobs. Food stamp benefits are found to be higher for families with more chil-
dren and higher for mothers with lower earnings. Such benefits are esti- mated to be lower for families in which focal fathers had annual earnings above $10,000 in year 1 than for families with focal fathers who have no formal earnings in that year. Also, food stamp benefits are estimated to be
Paternal Imprisonment 473
Table 3
DIFFERENCE-IN-DIFFERENCE MODELS: EFFECTS OF PATERNAL IMPRISONMENT ON CHILD SUPPORT AND FOOD STAMPS
Annual Child Support Received (in US$)
Annual Food Stamp
Benefits (in US$)
Coefficient SE Coefficient SE
Intercept 1,276** 175 205* 85 Time (reference is year 5 versus year 2) 563** 53 181** 29 Imprisonment group
(paternal imprisonment in year 4) 2221** 54 233** 68 Time × imprisonment 2701** 71 208** 80 Mother had any child by another partner
(time-varying variables)a 1,092** 168 257* 130 No. of mother’s children (time-varying
variables; reference is one child): 2 116 117 99** 100 ≥ 3 2191 475 1,876** 491
Mother’s formal earnings in year 1 (reference is no earnings):
$1–$4,999 83 80 9 68 $5,000–$9,999 255** 95 2246** 69 $10,000–$19,999 278** 99 2467** 66 ≥ $20,000 664** 184 2564** 75
Focal father’s formal earnings in year 1 (reference is no earnings):
$1–$4,999 2414** 102 222 67 $5,000–$9,999 2285* 118 276 73 $10,000–$19,999 45 121 2208** 61 ≥ $20,000 956** 142 2278** 54
Age of focal father (in child support model) or age of mother (in food stamps model) at focal child’s birth (reference is 21–23):
Under 18 2203* 102 286** 65 19–20 2112 78 32 44 24–27 2241 123 44 73 ≥ 28 2571 141 113 98
Age of focal father in relation to mother’s age (reference is same age [± 2 years]):
≥ 10 years younger 2262 199 116 75 5–9 years younger 2370** 130 45 57 2–5 years younger 2213** 81 136** 47 2–5 years older 246 272 18 90 5–9 years older 247 353 247 221 ≥ 10 years older 392 576 234 241
Parents’ races combined (reference is both white):
Both black 2556** 101 342 74 Both Hispanic 2414** 142 259 186 Mother white, father black 2248 224 197** 76 Mother white, father Hispanic 2216 153 46 109 All other combinations 2376** 136 127 80 Either unknown 2107 98 20 51
highest among families with mothers who are under the age of 18 at the time of the birth of their first child. Families residing in urban counties (particularly Milwaukee) tend to receive higher food stamp benefits than do families residing in rural counties. Families who have previously re- ceived food stamp or TANF cash benefits are found to receive larger amounts of food stamp benefits compared to those who have not received food stamp or TANF cash benefits previously. Interestingly, even in the analysis that controls for the total number
of children, multiple-partner fertility is found to be positively associated with the level of food stamp benefits. To examine the extent to which multiple-partner fertility functions as an intervening variable in the as- sociation of father’s imprisonment with food stamp benefits, the author conducted sensitivity tests that include control variables for multiple- partner fertility and the number of children (both measured at the end of year 4). Results suggest that multiple-partner fertility and the subse-
Table 3 (Continued )
Annual Child Support Received (in US$)
Annual Food Stamp
Benefits (in US$)
Coefficient SE Coefficient SE
Milwaukee 281 102 322** 65 Urban counties 71 79 115** 41
Year of focal child’s birth (reference is 1998): 1999 239 112 52 51 2000 2122 116 207** 54 2001 2278* 110 234** 59 2002 2191 119 289** 56
Any food stamps from 1 year prior to focal child’s birth until 1 year after 21281 70 855** 45
Any TANF in year 1 2203** 72 124* 57 Focal father’s paternity establishment and child
support status (reference is paternity established, positive child support owed; no payment made in year 1):
No paternity established until year 1 2208* 89 Paternity established but no support orders until year 1 21951 113
Paternity established, positive child support owed; payments made in year 1 1,215** 103
Source.—Wisconsin administrative data. Note.—N 5 5,550 focal child–mother pairs; 11,100 observations. Baseline is the birth of
the mother’s first nonmarital child (focal child); outcomes are measured against this baseline, and weights are applied. Monetary amounts are adjusted to December 2008 dollars.
a Time-varying control variables were measured by the end of year 1 and year 3 (corre- sponding to the first and second time periods, respectively).
1 p < .10. * p <. 05. ** p < .01.
Paternal Imprisonment 475
quent increase in the number of children may account for up to 53 per- cent of paternal imprisonment’s positive relationship with food stamp benefits.
Sensitivity Tests
Table 4 shows key results from fixed effects models, and the results are consistent with estimates from the DD models; fathers’ imprisonment is estimated to reduce the child support received by families and to increase the food stamp benefits they receive. Additional sensitivity tests are con- ducted using alternative model specifications and samples (see table A1 in the appendix). The alternative models include but are not restricted to DD models that consider only the imprisonment of the focal child’s bio- logical father in identifying paternal imprisonment, DD models that con- trol for time-varying measures of formal employment of the mother and the focal father, and DD models that consider formal child support or food stamp benefits in year 6 as the dependent variable. To assess the ex- tent to which base results may be affected by incomplete (continuing) periods of incarceration among men incarcerated in year 4, an alterna- tive analysis estimates models that restrict the treatment group to families that experienced the father’s imprisonment in year 4 as well as in year 5. In sum, key results remain consistent across models.
Table 4
FIXED EFFECTS MODELS: PATERNAL IMPRISONMENT’S EFFECTS ON CHILD SUPPORT AND FOOD STAMPS
Difference in Annual
Child Support Received,
Year 2–Year 5
Difference in Annual
Food Stamp Benefits,
Year 2–Year 5
Coefficient SE Coefficient SE
Intercept 528** 57 237 29 Paternal imprisonment in year 4 2696** 69 253** 80 Mother had children with other partner by end of year 3 versus end of year 1 488** 181 150 159
Mother had more than one child by end of year 1 277 340 2680 489
No. of children increased by one between year 1 and year 3 517** 136 547** 130
No. of children increased by two or more between year 1 and year 3 375 503 1,399** 407
Source.—Wisconsin administrative data. Note.—N 5 5,550 focal child–mother pairs. Results are adjusted to December 2008 US
dollars. Baseline is the birth of the mother’s first-born nonmarital child, and outcomes are measured against this baseline. Weights are applied. ** p < . 01.
476 Social Service Review
In results from logistic regression analysis, families that experienced paternal imprisonment in year 4 are, in year 5, estimated to be .69 times less likely to receive any child support and .56 times more likely to receive food stamp benefits than they are in year 2 (see table A1, test 6). The tests also estimate effects of fathers’ imprisonment on mothers’ formal earn- ings and TANF cash benefits: the estimated effects on both outcomes are statistically nonsignificant (see table A2, tests 13 and 14). These results are robust to alternative model specifications.
Conclusion and Policy Implications
To help researchers and policy makers understand and enhance both the economicwell-beingofdisadvantagedchildrenandtheinteractionamong social systems, this research examines the effects of fathers’ imprisonment on child support and welfare outcomes among nonmarital children’s fam- ilies. Although results indicate that paternal imprisonment reduces the child support received by mothers, they also suggest that concurrent in- creases in mothers’ food stamp benefits fill a portion of the resulting gap. An additional analysis suggests that the associations of fathers’ imprison- ment with economic outcomes continue over a longer time frame than that captured in this study (see table A1, test 5). The implications of these findings differ with one’s perspective on the
function of welfare. If one understands welfare as a vital safety net, the results can be interpreted as a limited success story: food stamp benefits work as a safety net when families face an economic crisis due to fathers’ imprisonment. This perspective may give rise to concern for those who, because of a father’s imprisonment, become impoverished and eligible for food stamp benefits but do not use those benefits (this group may be of concern especially because the associations of fathers’ imprisonment with mothers’ earnings and TANF cash benefits are estimated to be statis- tically nonsignificant).Conversely, one whose viewpoint on welfare is framed by concern about welfare dependency and work incentives may see increases in family welfare reliance as a potentially important and negative consequence of paternal imprisonment. A third perspective may find concern in the evidence that a father’s imprisonment increases the family’s welfare usage because growth in welfare usage imposes a sub- stantial economic burden on taxpayers. Several limitations of the data used in this study are noteworthy. First,
the data provide limited information about the quality of fathers’ involve- ment with their children beyond formal child support payments. Also, approximately 23 percent of focal fathers are estimated to be missing from the data because their paternity of focal children is not formally es- tablished. If the effects of these fathers’ imprisonment on economic out- comes differ from those of fathers reported in the data, estimates will be
Paternal Imprisonment 477
biased. The data are likely to miss nonmarital births that shortly preceded either the parents’ marriage (unless the couple later divorced and the mother enrolled the child’s case in the child support enforcement sys- tem) or long-lasting cohabitation without formal establishment of pater- nity in KIDS.5 These missing cases will downwardly bias the estimates of the influence of fathers’ imprisonment on food stamp benefits if fathers in more stable relationships with mothers are more likely to contribute to their children prior to imprisonment. Missing information about some types of fertility affects the selection of the study sample (as explained above) and also limits the study’s ability to identify any subsequent part- ners with whom the mothers have children. This limitation is relevant be- cause the current study considers the imprisonment of these subsequent partners (whether or not they are married to the mother). However, the data constructed for this analysis are unique in providing the opportunity to simultaneously consider the imprisonment of fathers of all children born to the sampled mothers. The analysis also is limited to examining the correlates of incarceration in Wisconsin state prisons. The data do not capture incarceration in such other correctional facilities as local jails or federal penitentiaries. To the extent that the comparison group includes families with fathers incarcerated in other correctional systems, but not in a Wisconsin state prison, missing incarceration information will tend to downwardly bias the absolute sizes of the estimated effects. Moreover, because the effects of incarceration may differ across jurisdic- tions, caution should be taken in the generalization of the results. Finally, the data do not include reliable information on fathers’ impris-
onment histories prior to the focal child’s birth; therefore, the analysis does not control for this factor in the model of imprisonment effects. This limitation should be alleviated to some extent by the construction of the sample, which excludes families that experienced fathers’ imprisonment in the first 3 years following the focal child’s birth. However, this sample restriction limits the generalizability of the results; because this process disproportionately excludes cases in which fathers are unlikely to provide support to their children, a study that includes a broader group of non- marital children might produce smaller estimates of the relationships between imprisonment and economic outcomes (for a relevant analysis, see table A2, test 10). Some additional limitations of the analysis merit discussion and may
prompt innovation in future research. First, the study’s empirical food stamp model does not simultaneously consider mothers’ work and partic- ipation in such other public programs as Medicaid, the Earned Income Tax Credit, and housing and child subsidies. These factors conjointly shape the options available to mothers and simultaneously affect their re-
5. It is also possible that some of the missing fathers are not represented in the data be- cause the state child support enforcement system has difficulty locating them. Such fathers may include those who are incarcerated.
478 Social Service Review
source decisions. Second, the dissimilar effects of paternal imprisonment for different groups (e.g., mothers from different racial groups or those with different levels of human capital) merit further research. Another limitation is that the study does not differentiate the effects of the impris- onment of focal children’s biological fathers from the effects of the im- prisonment of their half-siblings’ fathers. In addition, although the cur- rent study focuses on whether paternal imprisonment occurred in a given period of time, further research could improve understanding of pater- nal imprisonment by investigating the effects of the length of paternal im- prisonment on economic outcomes. Simultaneous consideration of mothers’ imprisonment would also provide a nuanced understanding of the issue. Notwithstanding these limitations, the study makes several important
contributions to the literature on child support, welfare, and parental in- carceration. By considering the imprisonment of the fathers of focal chil- dren’s half-siblings, the study makes conceptual and empirical contribu- tions to the literature on social policy that targets populations in which multiple-partner fertility is prevalent. Comparison of the base and sensi- tivity results suggests that, in studies of paternal imprisonment, consider- ing other partners of the mothers would likely affect the magnitude of es- timated imprisonment effects (although, as in this study, the results may be similar in terms of the direction and statistical significance of key coef- ficients). For example, the magnitudes of imprisonment effects are likely underestimated in empirical studies in which children who experience the imprisonment of their social fathers (not their biological fathers) are categorized as having no experience of paternal imprisonment. The results of this study have implications for the intersection of mul-
tiple public policies. Findings suggest that widespread imprisonment may undermine state governments’ efforts to collect child support and may reduce states’ abilities to meet performance standards. High rates of imprisonment may also undermine the success of recent welfare policy reforms designed to reduce welfare expenditures and promote parental responsibility among low-income fathers. The empirical results of the current study challenge the traditional cost-benefit analysis of incarcera- tion, which ignores the costs of incarceration on the families left behind as well as on other public systems. By identifying empirical evidence of the extensive collateral costs of imprisonment, this study provides further support for the dominant claim that the current US incarceration rate is above the optimal level (Raphael and Stoll 2009). In addition, the study suggests that widespread paternal imprisonment
may, in the context of declining public support, exacerbate the conse- quences of welfare reform for children and should therefore be consid- ered in the design, evaluation, and planning of social policies. Research- ers and advocates increasingly pay attention to the connection between incarceration and child support. Efforts are underway to develop policy options to alleviate the negative consequences of incarceration for both
Paternal Imprisonment 479
children and public systems, including child support systems (Cancian, Noyes, et al. 2009; Comfort et al. 2011). Such efforts should be continued in order to enhance the economic well-being of disadvantaged families and children.
Appendix
Table A1
THE EFFECTS OF FATHERS’ IMPRISONMENT ON ECONOMIC OUTCOMES: ALTERNATIVE SAMPLES AND MODELS
Effects of Imprisonment on:
Child Support Food Stamps
N Coefficient SE Coefficient SE
Results using base specificationsa 5,500 2701** 71 208** 80 Sensitivity test results for child
support and food stamp models: 1. Considered only focal father in identifying paternal imprisonment 5,500 2683** 72 169* 84
2. Considered only focal father in identifying paternal imprisonment and calculating child support payments 5,500 2659** 68 . . .
3. Included child support order establishment and payments from focal father (measured year 1) as control variables in food stamp model 5,500 . . . 208** 80
4. Controlled for second father’s formal earnings in year 1 (if any second father by time considered) 5,500 2677** 70 223** 80
5. Examined economic outcomes in year 6 as dependent variable 5,500 2553** 81 303** 90
6. Employed logistic analysis of any receipt of child support or food stamps in year 5 (vs. year 2) 5,500 2.69** .18 .56* .22
7. Controlled for time-varying measures of formal earnings of mother and focal father 5,500 2647** 72 170* 80
8. Excluded cases in which imprisonment was for nonpayment of child support 5,504 2702** 71 200* 80
9. Included cases in which paternity date was unknown or paternity established more than 5 years after focal child’s birth 5,784 2685** 68 228** 68
10. Included cases previously excluded because any of mother’s partners were imprisoned when focal child was age 0–3 8,624 2499** 54 101* 47
480 Social Service Review
Table A1 (Continued )
Effects of Imprisonment on:
Child Support Food Stamps
N Coefficient SE Coefficient SE
11. Restricted treated group to families that experienced paternal imprisonment in years 4 and 5 5,459 2762 73 225 86
12. Used alternative treated group (test 11); restricted control group to families that experienced fathers’ imprisonment in neither year 4 nor year 5 4,907 2781 74 234 86
Imprisonment effects on other economic outcomes:
13. Considered TANF cash benefits as dependent variable 5,500 6 108
14. Considered formal earnings as dependent variable 5,500 55 421
Source.—Wisconsin administrative data. Note.—N 5 5,550 focal child–mother pairs. TANF 5 Temporary Assistance for Needy
Families program. Baseline is the birth of the mother’s first-born nonmarital child, and outcomes are measured against this baseline. Weights are applied.
a Base specifications are as follows. When identifying paternal imprisonment, a dummy variable is used to indicate the imprisonment of the focal child’s biological father or the fa- ther (or fathers) of the focal child’s half-sibling (or half-siblings). The key economic out- come variables are child support and food stamp benefits in dollars in year 5 (vs. year 2). Control variables include, but are not restricted to, time-varying measures of multiple-partner fertility and number of children, and earnings of mother and focal father in year 1. Base mod- els do not control for the earnings of subsequent partners of the mother. * p < .05. ** p < .01.
Table A2
ANNUAL ECONOMIC OUTCOMES OF MOTHERS FOLLOWING THE BIRTH OF FIRST CHILDREN
Mean ($)
Median ($)
Minimum ($)
Maximum ($)
% of Positive
Child support payments received by mothers from all partners:
Year 1 500 0 0 12,707 34 Year 2 1,404 276 0 122,803 57 Year 3 1,718 815 0 96,718 65 Year 4 1,916 1,084 0 81,514 69 Year 5 2,028 1,281 0 27,539 71 Year 6 2,129 1,341 0 27,135 71
Food stamps received by mothers: Year 1 824 0 0 17,622 44 Year 2 871 0 0 14,846 48 Year 3 1,029 55 0 13,046 51 Year 4 1,151 114 0 20,408 53 Year 5 1,217 115 0 20,966 53 Year 6 1,286 90 0 14,012 52
Mothers’ formal earnings: Year 1 7,613 4,720 0 72,904 83 Year 2 10,148 7,741 0 72,431 85 Year 3 11,015 8,299 0 94,465 84 Year 4 11,548 8,965 0 108,736 83 Year 5 11,801 9,080 0 84,933 81 Year 6 12,147 9,399 0 108,993 79
TANF cash benefits received by mothers:
Year 1 715 0 0 10,333 24 Year 2 562 0 0 10,258 15 Year 3 602 0 0 9,671 16 Year 4 476 0 0 10,179 14 Year 5 383 0 0 9,460 11 Year 6 301 0 0 9,604 10
Source.—Wisconsin administrative data. Note.—N 5 5,550 focal child–mother pairs. Weights are applied. All monetary amounts
are adjusted to December 2008 dollars.
Note
Yiyoon Chung is an assistant professor of social work at the University of Wisconsin–Whitewater. Her research addresses the consequences of public pol- icies for economically disadvantaged families and children and for members of racial and ethnic minority groups. Her specific research interest is the intersec- tion of multiple social systems, including the criminal justice system, child sup- port and welfare policies, and child welfare. She can be reached at chungy@uww .edu. The author is grateful to Maria Cancian, Dan Meyer, Lonnie Berger, Pam Oliver, and Marcy Carlson for invaluable feedback on earlier versions of this ar- ticle and related research. She also thanks Steve Cook for expert advice; Jennifer Noyes for contributions to related projects; Felix Elwert for helpful discussions; Katherine Thornton, Jane Smith, Lynn Wimer, and the rest of the Institute for Research on Poverty (IRP) Datacore staff for the construction of the data files
482 Social Service Review
used in this study; and Jennifer Eggerling-Boeck, Lila Stromer, and Dawn Duren for assistance in preparing this manuscript. The author is also grateful to the anonymous reviewers at Social Service Review for insightful comments on an earlier draft. This project was supported in part by the 2010–11 Graduate Research Fellow Dissertation Grant from the IRP. This article builds on research conducted under a contract between the Wisconsin Department of Children and Families and the University of Wisconsin–Madison IRP.
References Allison, Paul D. 2005. Fixed Effects Regression Methods for Longitudinal Data Using SAS. Cary, NC:
SAS Institute. Argys, Laura M., H. Elizabeth Peters, Jeanne Brooks-Gunn, and Judith R. Smith. 1998. “The
Impact of Child Support on Cognitive Outcomes of Young Children.” Demography 35 (2): 159–73.
Barnow, Burt S., Timothy M. Dall, Mark W. Nowak, and Barbara E. Dannhausen. 2000. “The Potential of the Child Support Enforcement Program to Avoid Costs to Public Pro- grams: A Review and Synthesis of the Literature.” Report, April. Office of Child Support Enforcement, Department of Health and Human Services, Washington, DC.
Bartfeld, Judi, and Daniel R. Meyer. 2003. “Child Support Compliance among Discretionary and Nondiscretionary Obligors.” Social Service Review 77 (3): 347–72.
Becker, Gary S. 1991. A Treatise on the Family. Cambridge, MA: Harvard University Press. Becker, Gary S., and Robert J. Barro. 1988. “A Reformulation of the Economic Theory of
Fertility.” Quarterly Journal of Economics 103 (1): 1–25. Beckett, Katherine, and Bruce Western. 2001. “Governing Social Marginality: Welfare, Incar-
ceration, and the Transformation of State Policy.” Punishment and Society 3 (1): 43–59. Beller, Andrea H., and John W. Graham. 1993. Small Change: The Economics of Child Support.
New Haven, CT: Yale University Press. Berger, Lawrence M., Marcia J. Carlson, Sharon H. Bzostek, and Cynthia Osborne. 2008.
“Parenting Practices of Resident Fathers: The Role of Marital and Biological Ties.” Jour- nal of Marriage and Family 70 (3): 625–39.
Berger, Lawrence M., and Callie E. Langton. 2011. “Young Disadvantaged Men as Fathers.” Annals of the American Academy of Political and Social Science 635 (1): 56–75.
Blank, Rebecca M., and Patricia Ruggles. 1996. “When Do Women Use Aid to Families with Dependent Children and Food Stamps? The Dynamics of Eligibility versus Participa- tion.” Journal of Human Resources 31 (1): 57–89.
Brooks-Gunn, Jeanne, and Greg J. Duncan. 1997. “The Effects of Poverty on Children.” Fu- ture of Children 7 (2): 55–71.
Brown, Patricia R., and Steven T. Cook. 2008. “A Decade of Voluntary Paternity Acknowledg- ment in Wisconsin, 1997–2007.” Report to the Wisconsin Department of Workforce De- velopment. University of Wisconsin, Institute for Research on Poverty, Madison.
Cancian, Maria, Carolyn Heinrich, and Yiyoon Chung. 2009. “Does Debt Discourage Em- ployment and Payment of Child Support? Evidence from a Natural Experiment.” Dis- cussion Paper 1366-09, July. University of Wisconsin, Institute for Research on Poverty, Madison. http://www.ssc.wisc.eduirpweb/publications/dps/pdfs/dp136609.pdf.
Cancian, Maria, and Daniel R. Meyer. 2006. “Child Support and the Economy.” 338–65 in Working and Poor: How Economic and Policy Changes Are Affecting Low-Wage Workers, edited by Rebecca M. Blank, Sheldon H. Danziger, and Robert F. Schoeni. New York: Russell Sage Foundation.
Cancian, Maria, Daniel R. Meyer, and Emma Caspar. 2008. “Welfare and Child Support: Com- plements, Not Substitutes.” Journal of Policy Analysis and Management 27 (2): 354–75.
Cancian, Maria, Daniel R. Meyer, and Steven T. Cook. 2011. “The Evolution of Family Com- plexity from the Perspective of Nonmarital Children.” Demography 48 (3): 957–82.
Cancian, Maria, Jennifer L. Noyes, Yiyoon Chung, and Katherine Thornton. 2009. “Holding Child Support Orders of Incarcerated Payers in Abeyance: Interim Evaluation Report.” Report to the Wisconsin Department of Workforce Development, December. Univer- sity of Wisconsin, Institute for Research on Poverty, Madison.
Paternal Imprisonment 483
Carlson, Marcia J., and Frank F. Furstenberg Jr. 2006. “The Prevalence and Correlates of Mul- tipartnered Fertility among Urban U.S. Parents.” Journal of Marriage and Family 68 (3): 718–32.
Chung, Yiyoon. 2011. “Children’s Exposure to Paternal Imprisonment: Incidence, Evolu- tion, and Correlates among Young Nonmarital Children.” Children and Youth Services Re- view 33 (5): 575–87.
Comfort, Megan, Anne M. Nurse, Tasseli McKay, and Katie Kramer. 2011. “Taking Children into Account.” Criminology and Public Policy 10 (3): 839–50.
Council of State Governments. 2009. “The State Comparative Performance Measurement: Public Assistance.” Report. Council of State Governments, Lexington, KY. http://www .csg.org/programs/policyprograms/documents/CPM_Public_Assistance.pdf.
Currie, Janet. 2003. “U.S. Food and Nutrition Programs.” 199–289 in Means-Tested Transfer Programs in the United States, edited by Robert A. Moffitt. Chicago: University of Chicago Press.
Geller, Amanda, Irwin Garfinkel, Carey E. Cooper, and Ronald B. Mincy. 2009. “Parental In- carceration and Child Well-Being: Implications for Urban Families.” Social Science Quar- terly 90 (5): 1186–1202.
Geller, Amanda, Irwin Garfinkel, and Bruce Western. 2011. “Paternal Incarceration and Sup- port for Children in Fragile Families.” Demography 48 (1): 25–47.
Gottschalk, Peter. 1996. “Is the Correlation in Welfare Participation across Generations Spu- rious?” Journal of Public Economics 63 (1): 1–25.
Graham, John W., and Andrea H. Beller. 2002. “Nonresident Fathers and Their Children: Child Support and Visitation from an Economic Perspective.” 431–53 in Handbook of Father Involvement: Multidisciplinary Perspectives, edited by Catherine S. Tamis-LeMonda and Natasha Cabrera. Mahwah, NJ: Erlbaum.
Grall, Timothy S. 2009. Custodial Mothers and Fathers and Their Child Support: 2007. Current Population Reports no. P60-237, November. Washington, DC: US Census Bureau.
Grogger, Jeffrey. 2004. “Welfare Transitions in the 1990s: The Economy, Welfare Policy, and the EITC.” Journal of Policy Analysis and Management 23 (4): 671–95.
Grogger, Jeffrey, and Lynn A. Karoly. 2005. Welfare Reform: Effects of a Decade of Change. Cam- bridge, MA: Harvard University Press.
Gundersen, Craig, and Victor Oliveira. 2001. “The Food Stamp Program and Food Insuffi- ciency.” American Journal of Agricultural Economics 83 (4): 875–87.
Ha, Yoonsook, Maria Cancian, Daniel R. Meyer, and Eunhee Han. 2008. “Factors Associated with Nonpayment of Child Support.” Report to the Wisconsin Department of Work- force Development, September. University of Wisconsin, Institute for Research on Pov- erty, Madison.
Hagan, John, and Ronit Dinovitzer. 1999. “Collateral Consequences of Imprisonment for Children, Communities, and Prisoners.” Crime and Justice 26:121–62.
Hamilton, Brady E., Joyce A. Martin, and Stephanie J. Ventura. 2011. Births: Preliminary Data for 2010. National Vital Statistics Reports, vol. 60, no. 2, November 17. Hyattsville, MD: National Center for Health Statistics.
Harris, Alexes, Heather Evans, and Katherine Beckett. 2010. “Drawing Blood from Stones: Legal Debt and Social Inequality in the Contemporary United States.” American Journal of Sociology 115 (6): 1753–99.
Hoynes, Hilary Williamson. 2000. “Local Labor Markets and Welfare Spells: Do Demand Conditions Matter?” Review of Economics and Statistics 82 (3): 351–68.
Huang, Chien-Chung, Irwin Garfinkel, and Jane Waldfogel. 2004. “Child Support Enforce- ment and Welfare Caseloads.” Journal of Human Resources 39 (1): 108–34.
Huang, Chien-Chung, James Kunz, and Irwin Garfinkel. 2002. “The Effect of Child Sup- port on Welfare Exits and Re-Entries.” Journal of Policy Analysis and Management 21 (4): 557–76.
Issar, Sukriti. 2010. “Multiple Program Participation and Exits from Food Stamps among El- ders.” Social Service Review 84 (3): 437–59.
Johnson, Elizabeth I., and Jane Waldfogel. 2002. “Parental Incarceration: Recent Trends and Implications for Child Welfare.” Social Service Review 76 (3): 460–79.
Johnston, Denise. 1995. “Effects of Parental Incarceration.” 59–88 in Children of Incarcerated Parents, edited by Katherine Gabel and Denise Johnston. New York: Lexington Books.
484 Social Service Review
Kang, GiChoon, Sonya Kostova Huffman, and Helen H. Jensen. 2004. “An Empirical Anal- ysis of Joint Decisions on Labour Supply and Welfare Participation.” Applied Economics Letters 11 (14): 869–72.
Klawitter, Marieka, Robert D. Plotnick, and Mark Evan Edwards. 2000. “Determinants of Ini- tial Entry onto Welfare by Young Women.” Journal of Policy Analysis and Management 19 (4): 527–46.
Lengyel, Thomas E. 2006. “Spreading the Pain: The Social Cost of Incarcerating Parents.” Unpublished manuscript. September. Alliance for Children and Families, Department of Research and Evaluation Services, Milwaukee, WI. http://www.convictcriminology .org/pdf/SpreadingthePain.pdf.
Levingston, Kirsten D., and Vicki Turetsky. 2007. “Debtors’ Prison—Prisoners’ Accumula- tion of Debt as a Barrier to Reentry.” Clearinghouse Review 41, nos. 3–4 ( July–August): 187–97.
Lopoo, Leonard M., and Bruce Western. 2005. “Incarceration and the Formation and Sta- bility of Marital Unions.” Journal of Marriage and Family 67 (3): 721–34.
Magnuson, Katherine, and Elizabeth Votruba-Drzal. 2009. “Enduring Influences of Child- hood Poverty.” 153–79 in Changing Poverty, Changing Policies, edited by Maria Cancian and Sheldon Danziger. New York: Russell Sage Foundation.
McLanahan, Sara S. 1988. “Family Structure and Dependency: Early Transitions to Female Household Headship.” Demography 25 (1): 1–16.
———. 2009. “Fragile Families and the Reproduction of Poverty.” Annals of the American Academy of Political and Social Science 621 (1): 111–31.
Meyer, Bruce D. 1995. “Natural and Quasi-Experiments in Economics.” Journal of Business and Economic Statistics 13 (2): 151–62.
Meyer, Daniel R. 1993. “Child Support and Welfare Dynamics: Evidence from Wisconsin.” Demography 30 (1): 45–62.
Meyer, Daniel R., Yoonsook Ha, and Mei-Chen Hu. 2008. “Do High Child Support Orders Discourage Child Support Payments?” Social Service Review 82 (1): 93–118.
Miller, Robert A., and Seth G. Sanders. 1997. “Human Capital Development and Welfare Participation.” Carnegie-Rochester Conference Series on Public Policy 46:1–43.
Moffitt, Robert A. 1992. “Incentive Effects of the U.S. Welfare System: A Review.” Journal of Economic Literature 30 (1): 1–61.
Morgan, Stephen L., and Christopher Winship. 2007. Counterfactuals and Causal Inference: Methods and Principles for Social Research. New York: Cambridge University Press.
Mumola, Christopher J. 2000. “Incarcerated Parents and Their Children.” Bureau of Justice Statistics Special Report NCJ 182335, August. US Department of Justice, Office of Jus- tice Programs, Washington, DC.
Nagin, Daniel, and Joel Waldfogel. 1998. “The Effect of Conviction on Income through the Life Cycle.” International Review of Law and Economics 18 (1): 25–40.
Nepomnyaschy, Lenna,andIrwin Garfinkel. 2010. “Child Support Enforcement andFathers’ Contributions to Their Nonmarital Children.” Social Service Review 84 (3): 341–80.
Noyes, Jennifer L. 2006. “Review of Child Support Policies for Incarcerated Payers.” Report to the Wisconsin Department of Workforce Development, December. University of Wis- consin, Institute for Research on Poverty, Madison.
Oliver, Pamela, and James Yocom. 2002. “Racial Disparities in Criminal Justice: Madison and Dane County in Context.” Discussion Paper 1257-02, November. University of Wiscon- sin, Institute for Research on Poverty, Madison.
Pettit, Becky, and Bruce Western. 2004. “Mass Imprisonment and the Life Course: Race and Class Inequality in U.S. Incarceration.” American Sociological Review 69 (2): 151–69.
Plotnick, Robert. 1983. “Turnover in the AFDC Population: An Event History Analysis.” Jour- nal of Human Resources 18 (1): 65–81.
Poehlmann, Julie. 2009. “Children of Incarcerated Mothers and Fathers.” Wisconsin Journal of Law, Gender and Society 24 (2): 331–40.
Raphael, Steven, and Michael A. Stoll, eds. 2009. Do Prisons Make Us Safer? The Benefits and Costs of the Prison Boom. New York: Russell Sage Foundation.
Ribar, David C., Marilyn Edelhoch, and Qiduan Liu. 2008. “Watching the Clocks: The Role of Food Stamp Recertification and TANF Time Limits in Caseload Dynamics.” Journal of Human Resources 43 (1): 208–39.
Paternal Imprisonment 485
Schmidt, Laura A., and Dennis McCarty. 2000. “Welfare Reform and the Changing Land- scape of Substance Abuse Services for Low Income Women.” Alcoholism: Clinical and Ex- perimental Research 24 (8): 1298–1311.
Scholz, John Karl, Robert A. Moffitt, and Benjamin Cowan. 2009. “Trends in Income Sup- port.” 203–41 in Changing Poverty, Changing Policies, edited by Maria Cancian and Shel- don Danziger. New York: Russell Sage Foundation.
Sorensen, Elaine, and Chava Zibman. 2001. “Getting to Know Poor Fathers Who Do Not Pay Child Support.” Social Service Review 75 (3): 420–34.
Sparks, Holloway. 2003. “Queens, Teens, and Model Mothers: Race, Gender, and the Dis- course of Welfare Reform.” 171–95 in Race and the Politics of Welfare Reform, edited by San- ford F. Schram, Joe Soss, and Richard C. Fording. Ann Arbor: University of Michigan Press.
Swisher, Raymond R., and Maureen R. Waller. 2008. “Confining Fatherhood: Incarceration and Paternal Involvement among Nonresident White, African American, and Latino Fathers.” Journal of Family Issues 29 (8): 1067–88.
Teitler, Julien O., Nancy E. Reichman, and Lenna Nepomnyaschy. 2007. “Determinants of TANF Participation: A Multilevel Analysis.” Social Service Review 81 (4): 633–56.
Travis, Jeremy, and Michelle Waul, eds. 2003. Prisoners Once Removed: The Impact of Incarcera- tion and Reentry on Children, Families, and Communities. Washington, DC: Urban Institute Press.
US Census Bureau. 2010. Statistical Abstract of the United States, 2011. 130th ed. Washington, DC: US Department of Commerce.
———. 2011. “State and County Quick Facts: Milwaukee County, Wisconsin.” http://quick facts.census.gov/qfd/states/55/55079.html (accessed June 25, 2011).
US Department of Agriculture, Food and Nutrition Service. 2010. “Supplemental Nutrition Assistance Program: A Short History of SNAP.” http://www.fns.usda.gov/snap/rules /Legislation/about.htm (accessed June 25, 2011).
Vining, Aidan, and David L. Weimer. 2010. “An Assessment of Important Issues concerning the Application of Benefit-Cost Analysis to Social Policy.” Journal of Benefit-Cost Analysis 1 (1): 1–38.
Wakefield, Sara, and Christopher Uggen. 2010. “Incarceration and Stratification.” Annual Review of Sociology 36:387–406.
Waller, Maureen R., and Raymond R. Swisher. 2006. “Fathers’ Risk Factors in Fragile Fami- lies: Implications for ‘Healthy’ Relationships and Father Involvement.” Social Problems 53 (3): 392–420.
Weiss, Yoram, and Robert J. Willis. 1985. “Children as Collective Goods and Divorce Settle- ments.” Journal of Labor Economics 3 (3): 268–92.
West, Heather C., and William J. Sabol. 2009. Prison Inmates at Midyear 2008: Statistical Tables. Bureau of Justice Statistics Report NCJ 225619, March. Washington, DC: US Depart- ment of Justice, Office of Justice Programs. http://bjs.ojp.usdoj.gov/content/pub /pdf/pim08st.pdf.
Western, Bruce. 2002. “The Impact of Incarceration on Wage Mobility and Inequality.” Amer- ican Sociological Review 67 (4): 526–46.
———. 2006. Punishment and Inequality in America. New York: Russell Sage Foundation. Western, Bruce, Leonard M. Lopoo, and Sara S. McLanahan. 2004. “Incarceration and the
Bonds between Parents in Fragile Families.” 21–45 in Imprisoning America: The Social Ef- fects of Mass Incarceration, edited by Mary Patillo, David Weiman, and Bruce Western. New York: Russell Sage Foundation.
Western, Bruce, and Christopher Wildeman. 2009. “The Black Family and Mass Incarcera- tion.” Annals of the American Academy of Political and Social Science 621 (1): 221–42.
Wolkwitz, Kari. 2007. “Trends in Food Stamp Program Participation Rates, 1999 to 2005.” Current Perspectives on Food Stamp Program Participation Report, June. US Depart- ment of Agriculture, Food and Nutrition Service, Alexandria, VA.
Wooldridge, Jeffrey M. 2007. “Difference-in-Difference Estimation.” Lecture 11 presented at the University of Wisconsin, Institute for Research on Poverty, Applied Microecono- metrics Workshop, Madison, August 5. http://www.irp.wisc.edu/newsevents/work shops/appliedmicroeconometrics/appliedmicroeconometrics.htm.
Ziebert, Ryan. 2006. “No Easy Answers: The Effects of Parental Incarceration on Children.” Report, September. Alliance for Children and Families, Milwaukee, WI.
486 Social Service Review
Copyright of Social Service Review is the property of University of Chicago Press and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.