Sage Journals: Discover world-class research

Abstract

Previous studies of poverty governance have focused on the welfare system, the criminal justice system, and the connections between them. Yet less attention has been paid to a third institution that bridges the gap between these two systems: child support enforcement. Jailing for child support nonpayment is one of many mechanisms of child support enforcement, but little is known about this tactic. Using data from the Fragile Families and Child Wellbeing Study, the author examines the process of nonresident fathers’ (1) acquiring a formal support order, (2) accruing child support debt, and (3) being jailed for this debt. The author proposes two pathways into jail for child support nonpayment—public assistance and relationship context—and demonstrates how each pathway affects the risk for jail. Overall, 14 percent of debtors spend time in jail for child support by the time their children are nine years old.

Keywords

child support family policy welfare criminal justice

Previous studies of poverty governance (Beckett and Western 2001; Sugie 2012) have focused on how the service provision and punishment functions of the state—conceptualized, respectively, as the state’s “left” and “right” hands—work together in simultaneous welfare state retrenchment and criminal justice expansion (Wacquant 2009). Yet less attention has been paid to a third institution that bridges the gap between the welfare and criminal justice systems, shading into both at each end: child support enforcement. Child support enforcement is connected to the welfare system through the assignment of child support payments to the state when parents receive public assistance (Coven 1997). Child support enforcement also feeds into the criminal justice system, as noncustodial parents (NCPs) who fail to make their court-ordered child support payments can be found in contempt of court and incarcerated for their failure to pay (U.S. Department of Health and Human Services, Administration for Children and Families 2002). Researchers have likened jailing for child support nonpayment to a modern form of debtor’s prison, as many who go to jail for this reason may be unable to afford to pay their child support (Patterson 2008). Jailing for child support nonpayment is just one of many mechanisms of child support enforcement, but little is known about how frequently this tactic is used or against whom.

In this spirit, in this study I investigate who goes to jail for child support nonpayment. Using longitudinal data from four waves of the Fragile Families and Child Wellbeing Study (FFCW), I examines how NCPs find themselves at risk for being sent to jail for their child support debt. Child support debt exceeded $114 billion in 2014 (Office of Child Support Enforcement 2014a). About a quarter of all child support debt is owed to the state to reimburse welfare payments, while the remaining 75 percent is owed to children’s custodial parents (Office of Child Support Enforcement 2014b). With this in mind, I identify two conceptual pathways into jail for child support nonpayment: one instigated by the state against NCPs whose children have received public assistance and the other instigated by the children’s custodial parents. I examine how each pathway—public assistance involvement and relationship context—influences a nonresident parent’s risk for being jailed for child support nonpayment.

This study has theoretical, empirical, and policy implications. Theoretically, this investigation of jailing for child support nonpayment bridges the gap between studies of welfare state retrenchment and the criminal justice system, illustrating how this third institution contributes to the governance of social marginality for an increasing portion of the population. The child support enforcement system also constitutes part of the “shadow carceral state,” as parents jailed for the civil offense of child support nonpayment do not receive the same due process protections afforded to those accused of criminal offenses (Beckett and Murakawa 2012; Patterson 2008). Empirically, I provide national estimates of how many people are incarcerated for their child support debt and what factors shape the risk for jail. To my knowledge, the only existing estimates of jailing for child support nonpayment are at the state or county level (Chambers 1979; Cook 2015; May and Roulet 2005). With regard to policy, this study has ramifications for child well-being. Parental criminal justice involvement is associated with negative outcomes for children (Turney 2014a; Wakefield and Wildeman 2014; Wildeman 2010). One of the aims of the child support enforcement system is to improve child well-being by increasing the support of nonresident parents (Josephson 1997), something that could be compromised when jailing is the chosen tactic for child support enforcement.

Background

Founded with the intention of welfare cost recovery and carrying a sanction of incarceration for those that do not pay, the child support enforcement system exercises both the left and the right hands of the state (Wacquant 2009) at different moments. This study contributes to larger discussions of poverty governance by investigating the child support enforcement system’s connection to both the welfare and criminal justice systems.

Poverty Governance: Incarceration, Welfare, and Child Support Enforcement

Studies of prisoners and welfare recipients have noted the similarities between these two populations, which are disproportionately low income, less educated, and nonwhite (Soss 1999; Western and Wildeman 2009). Moreover, research suggests that a disproportionate share of public assistance recipients have incarcerated partners (Ovwigho, Saunders, and Born 2005) and that parental incarceration is associated with the receipt of some, but not all, public assistance programs (Sugie 2012). In the past 40 years, theorists have identified an increasingly punitive and paternalist approach to poverty governance (Beckett and Western 2001; Soss, Fording, and Shram 2008), which explains the simultaneous decline of welfare rolls and expansion of the criminal justice system. Child support systems modernized during the same time period, becoming more effective at establishing orders and withholding payments (Bartfeld and Meyer 2003; Garfinkel and Klawitter 1990). Rising criminal justice populations, declining welfare rolls, and modernized child support processes all arose during the same general era, and each is connected to the others.

The overlap between the institutions of child support enforcement and the criminal justice system occurs at two points: first, many inmates who have committed criminal offenses accrue child support debt during their incarceration, and second, incarceration is one possible sanction for failure to pay child support. Parents with incarceration history make up a disproportionate share of those owing child support debt (Ovwigho et al. 2005). Recent research has examined the legal debt that accompanies incarceration, finding that these legal financial obligations (LFOs) reproduce social inequality by charging inmates with the cost of their own incarceration, an even more extreme demonstration of neoliberal welfare state retrenchment and expanded punishment apparatus (Beckett and Murakawa 2012; Harris, Evans, and Beckett 2010). Incarcerating people for unpaid fines and fees is an important part of poverty governance in the expansive criminal justice system. Additionally, LFOs reproduce inequality by reducing family income, limiting access to resources, and ensuring continued surveillance by the legal system (Harris et al. 2010). Yet less attention has been paid to the role of child support debt for former inmates, although child support debt has the potential to reproduce inequality in similar ways. Like LFOs, child support debts can lead to incarceration for failure to pay and begin a cycle of debt accumulation and reincarceration among those who cannot afford to pay their debts (Harris 2016; Patterson 2008). Child support debt is estimated to be the highest debt that former inmates owe upon release (Reynolds et al. 2009).

Parents who fail to make their court-ordered child support payments can end up in jail through three possible mechanisms: criminal nonsupport, willful interstate evasion of child support,¹ and civil contempt of court for failure to pay court-ordered child support. Research suggests that civil contempt of court is the most common mechanism (Cook 2015). Contempt of court is a charge that ostensibly aims to coerce compliance rather than punish behavior, as parents can be released upon payment (Cook and Noyes 2011; Patterson 2008). This offense is classified as civil, requiring fewer safeguards than criminal offenses (U.S. Department of Health and Human Services, Administration for Children and Families 2002). Jailing for contempt of court constitutes part of the “shadow carceral state,” an extension of state power to detain beyond the traditional criminal context (Beckett and Murakawa 2012). Under the shadow carceral state, institutions that have not been considered part of the traditional criminal justice system, such as child support enforcement courts, acquire the capacity to detain. Few prior studies have empirically sought to determine how many people are sent to jail for child support nonpayment, and most have estimated this phenomenon on a state or local level (Chambers 1979; Cook 2015; May and Roulet 2005). Therefore, the first aim of this study is to show how many child support debtors spend time in jail for their failure to pay.

The child support enforcement system is also intertwined with public assistance provision. Eligibility for welfare can be revoked if recipients do not conform their work and family behavior to welfare policy prescriptions (Collins and Mayer 2010; Hays 2004). Among these is the requirement that welfare recipients comply with child support enforcement efforts (Josephson 1997). When a parent applies for Temporary Assistance for Needy Families (TANF) (or Medicaid, in some states), the state automatically opens a child support order on his or her behalf so that child support collections can reimburse the state for the cost of its welfare provision (Coven 1997; Roberts 2001). This policy is an example of welfare state retrenchment, a transference of costs from the public in terms of government-provided assistance to private cash transfers from the parent who lives apart from the child (Alexander 2005; Josephson 1997). When parents fall behind on their child support payments and accumulate debt, those with family public assistance histories find that the state owns some portion of their child support debt (Solomon-Fears 2012).

Family public assistance history, therefore, shapes the child support enforcement process. When child support debt is owed to the state, the state can proceed in efforts to collect on this debt even if the custodial parent does not agree (U.S. Department of Health and Human Services, Administration for Children and Families 2002). About half of the child support caseload consists of families with histories of current (10 percent) or prior (40 percent) welfare receipt (Office of Child Support Enforcement 2014a).² Child support debt is huge in magnitude, totaling more than $114 billion in 2014, and about 70 percent of the 15 million open child support cases owe debt (Office of Child Support Enforcement 2014a). About a quarter of all child support debt is owed to the state, and the other 75 percent is owed to the children’s custodial parents, usually the mother (Office of Child Support Enforcement 2014b). Many custodial parents are owed child support debt, but they may not all pursue this debt. A mother might seek to enforce child support against an NCP when relationship quality is low or when either parent has moved on with a new partner or new children. With this in mind, I identify two conceptual pathways into jail for child support nonpayment, one instigated by the state against NCPs whose children have received public assistance and the other instigated by the children’s custodial parents.

Going to jail for child support nonpayment is a multistep process (see Figure 1). For a nonresident parent to be at risk for going to jail for child support nonpayment, he³ must live apart from the child’s mother, have a formal child support order, and owe child support debt. The second aim of this study is to determine how these different pathways shape the risk for jail for child support nonpayment. I analyze how the public assistance and relationship context pathways operate to move an NCP through each progressive step of the process: having a formal order, accumulating child support debt, and finally going to jail for this debt.

Figure 1.

Jail for child support nonpayment as a multistep process.

Pathways: Public Assistance and Relationship Context

I argue that public assistance history and relationship context will shape a parent’s progress into becoming at risk for jail for child support nonpayment. Below, I review the literature on how each pathway should operate at each step of the process.

Formal Child Support

As described above, the state automatically opens a child support order against the nonresident parent when a custodial parent applies for TANF (Roberts 2001).⁴ With regard to relationship quality, establishment of a child support order may signal a breakdown in the parental relationship and an associated decline in informal and in-kind support (Kane, Nelson, and Edin 2015; Meyer and Cancian 2012; Nepomnyaschy and Garfinkel 2010; Waller and Plotnick 2001). Evidence from qualitative studies suggests that lower relationship quality and the loss of informal support can lead custodial parents to pursue formal support as a “last resort” (Edin 2000; Waller 2002; Waller and Plotnick 2001).⁵ Beginning a new relationship and having children with a new partner could also lead to a loss of support to the first family, compelling the first mother to open a formal child support order against the father (Edin and Nelson 2013). Thus, public assistance history and relationship quality may shape the likelihood that a family becomes involved with the formal child support enforcement system.

Child Support Debt

A parent must be in the formal child support system in order for him to accrue child support debt. Child support debt consists of any child support or medical support that has been ordered but has not been paid. In Texas, for instance, unpaid child support becomes classified as a debt if it is at least one month past due (Texas Family Code § 157.266). Some states have policies for ordering retroactive child support, whereby child support orders that are established several years after parents break up can be retroactively determined to start on the date the parents broke up, or even the date of the child’s birth, rather than the date the order was established (Office of Inspector General 2000). This can lead to parents’ owing a debt balance right from the moment of establishment. Furthermore, states vary as to the value of interest that they charge on child support debts. Some states, like New Jersey, do not charge any interest, whereas other states, like Illinois and Ohio, charge up to 10 percent annually (Sorenson, Sousa, and Schaner 2007). Retroactive debts, unpaid child support, and interest all make up child support debt.

Public assistance involvement could increase the likelihood of child support debt by affecting a nonresident father’s willingness and ability to pay. While a custodial parent is receiving TANF, she doesn’t get to keep any of her child support payments, aside from a small pass-through in some states. Seeing that she does not receive his child support payments could make a father less willing to pay his child support (Waller and Plotnick 2001), perhaps because he sees the situation as unfair (Lin 2000). Similarly, mothers with public assistance involvement may have children with fathers who have low earnings potential. Studies of assortative mating using the FFCW have shown that most mothers have more or equal educational attainment compared with fathers (Goldstein and Harknett 2006). When mothers have low educational attainment and earnings potential and turn to public assistance to make ends meet, this could signify that the fathers of their children are similarly low income. Ability to pay is a strong negative predictor of child support debt (Huang, Mincy, and Garfinkel 2005). Furthermore, in some states, a mother’s Medicaid receipt can increase the total sum of monthly support a noncustodial father must pay, by adding some amount of medical reimbursement to the existing monthly child support order (Solomon-Fears 2012). In some states, when mothers are on Medicaid, hospital birthing costs are billed to the NCP, creating a medical support debt from the first moments of his child’s life (Bartfeld and Meyer 2003). Having a larger order may increase the likelihood of debt, especially for low-income parents.

Similarly, relationship factors may also shape an NCP’s willingness and ability to pay child support and, by extension, drive the likelihood of child support debt accumulation. Relationship conflict is associated with child support debt (Turner and Waller 2017), and conflict and mistrust between parents is one reason fathers give for their hesitation to pay child support (Cozzolino and Williams 2017). Family complexity also increases the likelihood of child support debt (Meyer, Cancian, and Cook 2005) and could plausibly affect both a nonresident parent’s willingness and his ability to pay. When father has a new partner and children, he often “starts over” with this new family, which results in a lower investment in the first partner and children (Edin and Nelson 2013). A father’s family complexity may also affect his ability to pay, as new residential children must compete for resources with his other nonresidential children. Furthermore, the magnitude of a child support order generally increases as it is spread across multiple households. For instance, a father with three children making $20,000 a year in Wisconsin will pay a child support order of $5,800 if all three children are in one household but $8,564 if these three children are spread across three households (Meyer et al. 2005). A mother’s family complexity may also affect the father’s willingness to stay current on his child support payments, as he becomes unable to monitor how she spends her child support payments (Craigie 2015) and cannot be certain that these payments are not being spent on another man’s child (Cozzolino and Williams 2017). Thus, public assistance history and relationship quality may drive whether a custodial parent accumulates child support debt.

Jail for Child Support Nonpayment

A parent must owe child support debt in order to be jailed for child support nonpayment. Failing to pay child support is a necessary prerequisite to any of the three routes into jail (interstate evasion, criminal nonsupport, and civil contempt). In addition to their contribution to the likelihood of a nonresident parent’s (1) entering the formal child support system and (2) accumulating child support debt, public assistance history and relationship quality may also directly affect the likelihood of a nonresident parent’s being sent to jail for his child support debt. For families with public assistance history, the state has a financial incentive to pursue child support debt, because the state is permitted to keep some portion of what is recovered (Solomon-Fears 2012). In cases in which debt recovery through more routine enforcement mechanisms is difficult (Bartfeld and Meyer 2003), the state may be more willing to turn to punitive actions such as contempt of court. Some research suggests that fathers whose families have public assistance histories are more likely to be jailed than fathers without this involvement (Chambers 1979).

Relationship factors may also drive the likelihood of a parent’s being jailed for child support nonpayment, when this debt is owed mostly to the child’s mother. Just as parents with better relationships are more likely to avoid the formal child support system (Edin 2000; Waller and Plotnick 2001), mothers with good relationships are less likely to cooperate with child support enforcement efforts (Hamer 2001; Rich, Garfinkel, and Gao 2007; Waller 2002). Thus, mothers with worse relationships may be more willing to see their children’s fathers go to jail if they believe this will help them to recover the debt they are owed. Previous research about men on the run from the law has shown that romantic partners are sometimes willing to use the threat of jail in their relationship negotiations (Goffman 2014). Indeed, the judicial child support enforcement process can become a source of power in gendered conflicts between parents (Elmore 2010). Finally, family complexity may also affect the risk for going to jail for child support nonpayment. If a father has children with more than one mother, this could increase his exposure to the child support enforcement system. The more women with whom a noncustodial father has had children, the greater his likelihood that at least one of them will become disenchanted with him and pursue child support debt aggressively (Meyer et al. 2005).

On the basis of the preceding discussion, I hypothesize that public assistance history and relationship context will each increase the odds of a nonresident parent’s (1) having a formal child support order, (2) accumulating child support debt, and (3) being jailed for child support noncompliance.

Data and Method

Data Source

Data come from four waves of the FFCW, when focal children were ages one, three, five, and nine years. FFCW is a longitudinal birth cohort sample of nearly 5,000 families from 20 U.S. cities, of which three quarters of parents were unmarried at birth. When weighted with national sampling weights, the FFCW is representative of births occurring in cities with populations larger than 200,000 between 1998 and 2000. Because of the oversample of unmarried parents, these data are ideal for studying topics related to child support. This survey also collects data on involvement with the criminal justice system. To my knowledge, the FFCW is the only national data set that measures whether an NCP has gone to jail for child support nonpayment.

Analytic Strategy

On the basis of the multistep process of jail for child support nonpayment outlined above, this analysis follows a similar logic. Within a discrete-time event history framework, I conduct a series of three logistic regressions:

Among families with nonresident fathers, what predicts having a formal child support order?

Among fathers with a formal child support order, what predicts having child support debt?

Among fathers with child support debt, what predicts jail for child support nonpayment?

I combine all waves of data and reshape the data set so that each observation represents one person-wave. Each regression is clustered by respondent ID to control for including repeated measures of the same individuals over time.

This series of regressions allows me to parse out what factors affect progression through each step of the process of becoming at risk for jail for child support nonpayment. To adjust for time ordering, I conduct these regressions within a discrete-time event history framework. Results represent how each variable affects the likelihood of becoming eligible for the next stage of analysis by the next wave. At each stage, data were set for discrete-time event history analysis by identifying the dependent variable as a marker of failure and nesting repeated observations within a respondent’s ID, using the st set suite of commands in Stata 14. For each regression, I measure the impact of public assistance and relationship context. The last logistic regression is the key outcome of interest, who out of the eligible population of child support debtors is sent to jail for nonpayment.

For a father to be eligible to be sent to jail for child support nonpayment, he must (1) live apart from the mother, (2) have a formal child support order, and (3) owe child support debt (see Figure 1). Because of this, the sample for each equation is limited to those who are eligible. The first logistic regression predicting who has a formal child support order is limited to nonresident fathers, the second regression predicting child support debt is limited to those with formal child support orders, and the third regression predicting jail for child support nonpayment is limited to those with child support debt. Because my analyses focus on particular subsamples of the full FFCW, no appropriate weights exist for my samples (Turney 2011).⁶ The tables I present are unweighted (Sugie 2012; Turney 2014a, 2014b; Wildeman 2010).

Because the odds of proceeding through to the next regression depend on a respondent’s making it through the previous step, I control for the predicted probability of having made it through the prior step.⁷ The regression predicting child support debt controls for the predicted probability of having a formal child support order, and the regression predicting jail controls for the predicted probability of owing child support debt.⁸

There is a large amount of missing data in the fathers’ reports. To minimize this, I combine fathers’ and mothers’ reports for most measures, coding outcomes as true if either mother or father reported it. Missing data are accounted for using multiple imputation in Stata 14 with 50 imputed data sets (Allison 2009).

Finally, because child support policies vary widely by state (in terms of order amounts, interest on debt, and ability to collect on debt; Roberts 2001), I make use of the geocoded FFCW data to include dummies to control for the state where the respondent resides.⁹

Measures

Dependent Variables

Three dependent variables of interest are used in this analysis, describing whether a father has a formal order, owes debt, or goes to jail for child support nonpayment during any of the survey waves. First is whether a given nonresident father has a formal child support order. This is a binary variable coded 1 = formal child support order and 0 = no formal child support order. Fathers who live with their children at the current wave are excluded from the first logistic regression.

The second outcome of interest is whether a father owes child support debt. This is a binary variable coded 1 = owes child support debt and 0 = no child support debt. To minimize missing values, fathers who are reported as having been to jail for child support nonpayment, having faced other child support enforcement, or owing child support debt for children other than the focal child are also coded 1. Fathers without formal child support orders are excluded from the second logistic regression.

The third outcome of interest is whether a father has gone to jail for child support nonpayment. This is a binary variable coded 1 = has gone to jail for child support nonpayment and 0 = has never gone to jail for child support nonpayment. Jail for child support nonpayment is measured in two questions. The first is measured during the series of questions about the charges of fathers with criminal justice histories. The second is measured during the series of questions asking about child support enforcement. A positive response to either question is coded 1. Fathers without child support debt are excluded from the third logistic regression.

These variables are based on a combination of mothers’ and fathers’ reports for each of the four survey waves. If either a mother or a father reports that the father has a formal order, owes debt, or has been to jail for child support nonpayment, this is coded as true for this survey wave.

Independent Variables

Key independent variables measure the two pathways into jail for child support nonpayment: public assistance involvement and relationship context.

Public Assistance

I measure public assistance receipt as whether the focal child’s mother has received TANF or Medicaid. TANF history is measured as a binary variable indicating whether mother has ever received TANF by the current wave, coded 1 = has received TANF and 0 = has never received TANF. This measure combines two questions on TANF receipt. Mothers are asked if they have (1) ever received TANF as of that year or (2) received TANF at any time during the interview year. A positive response to either question is coded 1 for that wave. Medicaid history is coded as a binary variable measuring whether mother has ever received Medicaid by the current wave. Because the nine-year follow-up does not differentiate between Medicaid and other forms of health insurance, Medicaid receipt is measured only up until the time the child is age five.

Finally, the third logistic regression includes a measure of the amount of child support debt each father owes at each wave. This is an ordered categorical variable ranging from 1 to 8, with values of 1 = owes debt between $1 and $499, 2 = owes debt between $500 and $1,000, 3 = owes debt between $1,001 to $2,000, 4 = owes debt between $2,001 and $3,000, 5 = owes debt between $3,001 and $4,000, 6 = owes debt between $4,001 and $5,000, 7 = owes debt between $5,001 and $10,000, and 8 = owes debt above $10,000. Owing less than $500 in debt is the reference category.

Relational Variables

Four variables measure the relationship between parents. First is the mother’s report of the quality of the parental relationship at each wave. This is a scale with values of 0 = I never see him, 1 = poor quality, 2 = fair quality, 3 = good quality, 4 = very good quality, and 5 = excellent quality. Next is whether either mother or father is in a relationship with someone new at each wave, on the basis of a combination of mother’s and father’s reports, with each parent reporting on himself or herself. This is a binary variable coded 1 = either parent is in a new relationship and 0 = neither parent is in a new relationship. The final two variables measure multipartner fertility (MPF): whether either parent has children by more than one partner at each wave. The measure of mother’s MPF is coded as a binary variable on the basis of mother’s report as 1 = MPF and 0 = no MPF. Father’s MPF is coded the same way, but combines mother’s and father’s reports to minimize missing.

Controls

The first control is the parents’ relationship at baseline. This is a categorical variable with values of 0 = not married or cohabiting, 1 = married at birth, and 2 = cohabiting at birth. Next is whether father was incarcerated at baseline. This is based on mother’s report of whether father was in jail or prison at either of their baseline interviews, with a value of 1 = father is incarcerated. Father’s educational attainment at baseline is measured as a categorical variable with values of 1 = less than high school, 2 = high school degree, 3 = some college, and 4 = college degree or higher, with less than high school as the reference category. Next is father’s age at the baseline survey, which is coded as a continuous variable. Father’s race is included as a series of dummy variables for black, Hispanic, and other (race that is not white, black or Hispanic). White is the reference category. This measure was constructed by the survey developers and measured at the baseline survey.

In addition to these baseline control variables, I account for several time-varying controls. First is a measure of father’s household income, which is measured as a continuous variable at each survey wave. This measure is constructed by the survey developers. I include father’s employment status at each wave as a binary variable reporting whether father was working last week, with values of 1 = working and 0 = not working. Both mother’s and father’s reports are combined. Next is whether mother reports any domestic violence from father. This combines two measures of domestic violence that are reported at years 1, 3, 5, and 9 (see Boynton-Jarrett et al. 2010). Respondents were coded 1 if mother reported that father ever slapped or kicked her or if father ever seriously injured her.

Finally, state of residence is measured as where father resides at each survey wave. This measure is included as a set of fixed effects (not shown but available on request) measuring whether each individual father lived in each state at each wave, with Texas as the reference category. Father’s report of state is prioritized, but to minimize missing, I fill in with mother’s report of state if father’s is missing.

Results

Descriptive Results

Table 1 provides descriptive statistics for the full FFCW sample and each limited sample. A little more than half of FFCW observations report a nonresident father at any survey wave. Of families with nonresident fathers, about half have formal child support orders. Of families with formal child support orders, 60 percent owe debt on these orders. And of these child support debtors, about 14 percent go to jail for child support debt. Figure 2 graphs selected characteristics of these subsamples to demonstrate differences in public assistance and relationship context among these different subsamples. Compared with the analytic subsamples, a smaller share of those in the full FFCW have received public assistance or have relationship complexity (a new partner or new children). As the sample is refined, generally higher percentages of respondents have public assistance histories and relationship complexity. For interested readers, Table A4 presents the percentager of FFCW debtors who go to jail from each city (and aggregated by state).¹⁰ Figures A1 to A3 and Table A5 present the results of two-tailed t tests differentiating between those in and out of each subsample. These additional tables and figures can help contextualize how my samples compare with one another and with the full FFCW.

Table 1.

Descriptive Statistics by Subsample.

	Full FFCW	Nonresident Father = 1	Formal CS Order = 1	CS Debt = 1	Jail for CS Debt = 1
Nonresident father	0.529 (0.499)	1.000 (0.000)	1.000 (0.000)	1.000 (0.000)	1.000 (0.000)
Formal CS order	0.265 (0.441)	0.515 (0.500)	1.000 (0.000)	1.000 (0.000)	1.000 (0.000)
CS debt	0.149 (0.356)	0.297 (0.457)	0.604 (0.489)	1.000 (0.000)	1.000 (0.000)
Jail for CS nonpayment	0.019 (0.138)	0.039 (0.194)	0.082 (0.274)	0.140 (0.347)	1.000 (0.000)
Amount of CS debt (categorical)	4.896 (2.394)	4.979 (2.385)	4.979 (2.385)	4.979 (2.385)	5.974 (2.246)
Public assistance history
Mother has ever received TANF	0.475 (0.499)	0.627 (0.484)	0.641 (0.480)	0.693 (0.461)	0.855 (0.353)
Mother has received Medicaid by wave 3	0.579 (0.494)	0.719 (0.450)	0.730 (0.444)	0.768 (0.422)	0.699 (0.459)
Relational variables
Relationship quality	3.074 (1.523)	2.151 (1.420)	2.051 (1.363)	1.752 (1.249)	1.700 (1.284)
Either parent is in a new relationship	0.266 (0.442)	0.500 (0.500)	0.551 (0.497)	0.598 (0.490)	0.592 (0.492)
Mother has MPF	0.421 (0.494)	0.554 (0.497)	0.580 (0.494)	0.597 (0.491)	0.636 (0.482)
Father has MPF	0.533 (0.499)	0.627 (0.484)	0.719 (0.450)	0.726 (0.446)	0.819 (0.386)
Controls
Father worked last week	0.725 (0.446)	0.577 (0.494)	0.628 (0.483)	0.537 (0.499)	0.508 (0.501)
Relationship at baseline (categorical)	0.968 (0.870)	0.722 (0.922)	0.706 (0.920)	0.685 (0.921)	0.618 (0.908)
Father in jail at baseline	0.039 (0.193)	0.061 (0.239)	0.058 (0.233)	0.065 (0.246)	0.044 (0.206)
Any domestic violence	0.089 (0.284)	0.155 (0.362)	0.163 (0.369)	0.198 (0.399)	0.187 (0.391)
Father’s education (categorical)	2.288 (1.060)	2.173 (1.017)	2.168 (0.979)	2.108 (0.977)	2.145 (1.089)
Father age at baseline	32.127 (7.718)	31.235 (7.719)	31.719 (7.274)	31.457 (6.928)	31.536 (6.412)
Father’s household income	$47,436.18 ($54,954.52)	$35,179.63 ($41,537.10)	$35,842.95 ($39,781.27)	($39,934.43)	$28,443.12 ($27,285.89)
Father is white	0.188 (0.391)	0.107 (0.310)	0.110 (0.313)	0.111 (0.314)	0.122 (0.327)
Father is black	0.504 (0.500)	0.631 (0.483)	0.654 (0.476)	0.661 (0.473)	0.678 (0.468)
Father is Hispanic	0.264 (0.441)	0.219 (0.414)	0.199 (0.399)	0.191 (0.393)	0.155 (0.362)
Father is other race	0.040 (0.196)	0.034 (0.182)	0.032 (0.177)	0.032 (0.177)	0.046 (0.210)
State of residence (categorical)	32.897 (15.841)	33.523 (15.457)	35.277 (15.210)	35.979 (15.154)	38.363 (13.628)
Baseline state of residence (categorical)	33.144 (15.770)	33.742 (15.435)	35.512 (15.145)	36.154 (15.123)	38.678 (13.731)
n (person-waves)	16,665	8,814	4,284	2,359	304

Note: Values are mean (SD). CS = child support; MPF = multipartner fertility.

Figure 2.

Selected characteristics of FFCW subsamples.

I now turn to the results of the multivariate models. For Tables 2 to 4, model 1 includes public assistance measures, model 2 includes relationship context, and model 3 includes both. For Table 4, model 4 incorporates information on the amount of child support debt. All models include controls for state of residence as well as the sociodemographic controls described above. Because state and local practices related to Medicaid vary widely, I did a sensitivity analysis omitting the Medicaid receipt variable from models 1 and 3 for each outcome with substantively similar results (available upon request).

Table 2.

Results from Logistic Regressions Predicting Formal Child Support Order among Families with Nonresident Fathers.

	Odds Ratio (SE)
	Model 1	Model 2	Model 3
Wave	1.131***	1.098***	1.114***
Wave	(0.009)	(0.009)	(0.009)
Public assistance history
Mother has ever received TANF	1.153*		1.127⁺
Mother has ever received TANF	(0.075)		(0.075)
Mother has received Medicaid by wave 3	1.075		1.040
Mother has received Medicaid by wave 3	(0.071)		(0.070)
Relational variables
Relationship quality		1.008	0.947*
Relationship quality		(0.021)	(0.020)
Either parent has ever been in a new relationship		1.226***	1.186**
Either parent has ever been in a new relationship		(0.065)	(0.063)
Mother has MPF		0.973	0.973
Mother has MPF		(0.063)	(0.063)
Father has MPF		2.089***	2.159***
Father has MPF		(0.140)	(0.145)
Constant	0.296***	0.408***	0.293***
Constant	(0.059)	(0.083)	(0.062)
Observations	8,164	8,164	8,164

Note: Robust seeform in parentheses. All models control for state of residence, relationship at baseline, whether father was in jail at baseline, father’s baseline age and education, whether father worked last week, father’s household income, and father’s race. MPF = multipartner fertility; TANF = Temporary Assistance for Needy Families.

p < .10. *p < .05. **p < .01. ***p < .001.

Table 3.

Results from Logistic Regressions Predicting Child Support Debt among Families with Formal Child Support Orders.

	Odds Ratio (SE)
	Model 1	Model 2	Model 3
Wave	0.912	1.109⁺	1.027
Wave	(0.069)	(0.061)	(0.053)
Public assistance history
Mother has ever received TANF	1.167		1.384**
Mother has ever received TANF	(0.142)		(0.147)
Mother has received Medicaid by wave 3	1.103		1.172+
Mother has received Medicaid by wave 3	(0.111)		(0.113)
Relational variables
Relationship quality		0.742***	0.786***
Relationship quality		(0.023)	(0.031)
Either parent has ever been in a new relationship		1.276+	1.163
Either parent has ever been in a new relationship		(0.177)	(0.128)
Mother has MPF		1.099	1.021
Mother has MPF		(0.094)	(0.088)
Father has MPF		1.288	0.731
Father has MPF		(0.562)	(0.273)
Constant	0.294⁺	3.244	1.292
Constant	(0)	(2.640)	(0.738)
Observations	3,849	3,849	3,849

Note: Robust seeform in parentheses. All models control for the predicted probability of having a formal child support order, state of residence, relationship at baseline, whether father was in jail at baseline, father’s baseline age and education, whether father worked last week, father’s household income, and father’s race. MPF = multipartner fertility; TANF = Temporary Assistance for Needy Families.

p < .10. **p < .01. ***p < .001.

Table 4.

Results from Logistic Regressions Predicting Jail for Child Support Nonpayment among Families with Child Support Debt.

	Odds Ratio (SE)
	Model 1	Model 2	Model 3	Model 4
Wave	1.087⁺	1.038	1.051	1.003
Wave	(0.055)	(0.057)	(0.050)	(0.050)
Public assistance history
Mother has ever received TANF	0.932		0.758	0.733
Mother has ever received TANF	(0.232)		(0.206)	(0.198)
Mother has received Medicaid by wave 3	1.406		1.312	1.305
Mother has received Medicaid by wave 3	(0.300)		(0.282)	(0.285)
Relationship factors
Relationship quality		1.326	1.278	1.356⁺
Relationship quality		(0.265)	(0.211)	(0.226)
Either parent has ever been in a new relationship		0.764	0.759	0.741⁺
Either parent has ever been in a new relationship		(0.131)	(0.132)	(0.129)
Mother has MPF		0.991	1.035	1.045
Mother has MPF		(0.165)	(0.165)	(0.169)
Father has MPF		1.649*	1.658**	1.619*
Father has MPF		(0.324)	(0.319)	(0.315)
Amount of CS debt (reference: between $1 and $499)
$500–$1,000				1.462
$500–$1,000				(0.734)
$1,001–$2,000				1.677
$1,001–$2,000				(0.808)
$2,001–$3,000				1.536
$2,001–$3,000				(0.787)
$3,001–$4,000				1.600
$3,001–$4,000				(0.906)
$4,001–$5,000				1.728
$4,001–$5,000				(0.931)
$5,001–$10,000				2.256⁺
$5,001–$10,000				(1.084)
>$10,000				3.849**
>$10,000				(1.838)
Constant	0.054*	0.011*	0.008*	0.004**
Constant	(0.076)	(0.023)	(0.016)	(0.008)
Observations	2,046	2,046	2,046	2,046

Note: Robust seeform in parentheses. All models control for the predicted probability of owing child support debt, state of residence, relationship at baseline, whether father was in jail at baseline, father’s baseline age and education, whether father worked last week, father’s household income, and father’s race. CS = child support; MPF = multipartner fertility; TANF = Temporary Assistance for Needy Families.

p < .10. *p < .05. **p < .01.

Formal Child Support Order

Table 2 displays the multivariate results from a logistic regression predicting a formal child support order among nonresident fathers. In the first model, fathers whose children’s mothers have ever received TANF by the focal survey wave have 15 percent higher odds of a formal child support order, all else held constant. Model 2 incorporates relational variables. In model 2, relationship quality is negatively associated with the having a formal child support order, as every one-point increase in the relationship quality scale is associated with a 5 percent decline in the odds of having a formal order. Either parent having been in a new relationship is associated with an 18 percent increase in the odds of obtaining a formal child support order. Father’s MPF, but not mother’s MPF, is associated with 2.16 times higher odds of obtaining a formal child support order. In the full model, the effect of TANF drops to marginal significance, but the rest of the associations are the same.

Child Support Debt

Table 3 displays the multivariate results from a logistic regression predicting child support debt among fathers with formal child support orders. Compared with the previous model, this model has fewer significant associations, possibly because much of the difference between those with and without child support debt are absorbed by the differences between those with and without a formal child support order. This model controls for the predicted probability of having a formal child support order.

In model 1, neither TANF nor Medicaid is significantly associated with the accrual of child support debt.¹¹ In model 2, relationship quality is negatively associated with the accrual of child support debt, as every one-point increase in the relationship quality scale is associated with a 24 percent decline in father’s odds of acquiring child support debt. In model 3, the effect of TANF becomes significant, with fathers whose children’s mothers have ever received TANF by the current wave having 38 percent higher odds of accruing child support debt, compared with those families without public assistance involvement. In this model, relationship quality remains significantly negatively associated with accruing child support debt, at a similar magnitude as in model 2.

Jail for Child Support Nonpayment

Table 4 displays the multivariate results from a logistic regression predicting a jail for child support nonpayment among fathers with child support debt. Again, this model has few significant associations, suggesting that much of the difference between those who do and do not go to jail for child support debt is absorbed through differences between those with and without debt, and those with and without a formal order. This model controls for the predicted probability of having child support debt.

In model 1, neither TANF nor Medicaid is significantly associated with jail for child support nonpayment among child support debtors. In model 2, father’s MPF emerges as a significant predictor of jail for child support nonpayment among child support debtors, increasing the odds of jail by 65 percent. In model 3, father’s MPF remains the only significant predictor of jail for child support debt, with a similar magnitude. Finally, model 4 incorporates information on the amount of child support debt. Compared with parents owing less than $500 in child support debt, parents who owe more than $10,000 are almost four times as likely to go to jail for this debt. In model 4, father’s MPF remains a significant positive predictor of jail, even after controlling for the amount of debt.

Discussion and Conclusion

Taken together, these models provide moderate support for the hypothesized effects of public assistance involvement and relationship context on pathways into jail for child support nonpayment. Medicaid receipt is never significantly associated with jail for child support debt, perhaps because state and local referral practices vary widely. TANF receipt, on the other hand, is marginally positively associated with the acquisition of a child support order (odds ratio = 1.127, p < .10) and significantly positively associated with the accrual of child support debt (odds ratio = 1.384, p < 0.01). These results are as hypothesized. However, public assistance involvement is not significantly associated with jail for child support nonpayment among debtors. This could be because public assistance increases the odds of a father’s making it into the risk pool of child support debtors—by (marginally) increasing the odds of obtaining a formal order and by increasing the odds of accruing child support debt—but, once he makes it into this risk pool, TANF does not significantly increase his odds of going to jail for this debt.

Yet the amount of child support debt is an extremely strong predictor of jail for child support nonpayment, with fathers who owe above $10,000 almost four times as likely to go to jail as fathers who owe less than $500. Because TANF is associated with the accrual of debt, it is possible that TANF is operating indirectly through the amount of child support debt to affect the risk for jail. However, more analyses are necessary to determine whether this is the case.

These results provide consistent support for the relationship context pathway into jail for child support nonpayment. At the first stage, relationship quality, either parent having a new relationship, and father’s MPF, but not mother’s, are all associated with obtaining a formal child support order. In line with prior qualitative research (e.g., Edin and Nelson 2013; Waller and Plotnick 2001), parents with lower relationship quality and parents with new partners and children are more likely to have a formal child support order. To my knowledge, this is one of the first studies to find this same effect in quantitative data. At the second stage, relationship quality is significantly negatively associated with child support debt accrual, as a one-point increase in the relationship quality scale is associated with about a 20 percent decrease in the odds of accruing child support debt, in line with the hypotheses. Finally, in the third stage, father’s MPF (but not mother’s MPF; cf. Craigie 2015) is positively associated with jail for child support nonpayment among child support debtors, as hypothesized.

There are several limitations to these findings. First, I want to hedge the findings about which of these pathways is stronger. Because the public assistance pathway is measured by only two binary variables and the relationship context pathway is measured through four variables (three binary and one categorical), the measures of public assistance receipt have less variance than the measures of relationship context. Because of this, the generally stronger support for the relationship context pathway could be a statistical artifact. Although both the public assistance and relationship context pathways affect the process of becoming at risk for jail for child support, these differences in variation make it hard to say whether either pathway is stronger.

Second, my sample comes from a particularly low-income and nonwhite subsample of the FFCW (see Table A5 and Figures A1 –A3 for more on how my samples differ from the full FFCW). Because I am analyzing a subsample, these estimates are not weighted with national sampling weights. Therefore, the estimates from this article should not be generalized to the entire U.S. population. However, we know that incarceration is not equally distributed across the population (Alexander 2012; Wakefield and Wildeman 2014). Because of the structure of the child support enforcement system, as well as demographic predictors of single parenthood and child support debt (Bogenschneider 2000; Sorenson et al 2007), my analytic sample is likely representative of those most at risk for being jailed for child support debt, and my results can still be instructive for those involved in studying the child support, welfare, and criminal justice systems.

It is perhaps because of this sample refinement that I do not find significant effects for race (see full models in Tables A6 –A8). Compared with those in the full FFCW, the fathers in my formal, debt, and jail samples are all significantly less likely to be white and significantly more likely to be black. Perhaps the effect of differential selection by race is absorbed through the predicted probabilities I include that move from each prior regression to the next.¹² Research in criminal contexts shows that race matters for judicial discretion (e.g., Demuth 2003; Demuth and Steffensmeier 2004; Steffensmeier and Demuth 2001). Future studies should look more closely at the effect of race on jailing for child support debt.

Furthermore, because child support policy operates on a state level, I control for state of residence but do not analyze the effects of any particular state on the risk for jail for child support nonpayment. Future research should leverage state-level variation to examine how different child support policy levers affect the likelihood of jail for child support debt.¹³

Finally, there are two reasons a father may not pay his court-ordered child support: because he cannot afford to and because he is unwilling (Mincy and Sorenson 1998; Sorenson and Zibman 2001). The hypothesized pathways here of public assistance involvement and relationship context would function the same way regardless of the reason for child support nonpayment. Public assistance involvement could signify a low earnings potential for fathers or it could discourage fathers from paying child support because the money that they pay does not go to their children (Waller and Plotnick 2001), or some combination of both. Similarly, relationship context could affect both a father’s ability to pay (as new relationships and new children create greater competition for his finite resources; Meyer et al. 2005) and his willingness to pay (through personal animus and/or a loss of connection to former partners and older children; Edin and Nelson 2013; Turner and Waller 2017).

In the models, I control for father’s household income, educational attainment, and employment, using these measures as proxies for ability to pay (see Tables A6 –A8 for full models). I find that father’s household income is not significantly associated with any of the outcomes but that father’s work is positively associated with having a formal order and negatively associated with owing debt. Likewise, higher levels of education are positively associated with having a formal order and negatively associated with debt. Yet work and education are positively associated with jail (although only education reaches a conventional level of significance).¹⁴ At the same time, fathers who make it into the jail sample are less educated, less likely to work, and have lower income than fathers in the full FFCW (see Table A5). The effect of ability to pay on jailing is not clear from my data. Future research should investigate more closely how unwillingness and inability to pay child support shape the likelihood of serving jail time for child support nonpayment.

Overall, I find that 14 percent of child support debtors spend time in jail by the time their children are nine years old. This may seem perplexing, given that 60 percent of those in the formal sample owe debt. There are a few reasons why this number may be so low. First, contempt of court can be a lengthy process, necessitating a number of legal motions that can be time-consuming. Little is known about how this process works across the country, but research from select jurisdictions has shown that sheriffs see contempt of court as a low priority, service process can take a long time, and in some states,¹⁵ parents have right to counsel, which can delay a hearing on a contempt motion (Doolittle and Lynn 1998). Even when custodial mothers are seeking child support enforcement, some complain that the state does not act fast enough or does not do enough to try and collect support from fathers (Josephson 1997). Because it is a civil rather than a criminal matter, law enforcement officers may be slower to enforce contempt of court than criminal offenses.

I sought in this study to investigate who, among those with child support debt, goes to jail for this debt. Among child support debtors in the FFCW, about 14 percent go to jail for child support debt in the first nine years of their children’s lives. To determine risk factors, I used three logistic regressions within a state fixed-effects discrete-time event history framework to determine how two pathways, public assistance involvement and relationship context, affect progression through the steps of becoming at risk for jail for child support debt. To be at risk, a parent must live away from his child, have a formal child support order, and owe debt. I find that public assistance involvement and relationship context predict movement into having a formal child support order and accruing child support debt, but only relationship context predicts jail for child support nonpayment. In addition, fathers with $10,000 or more of child support debt have much higher odds of going to jail for this debt than fathers owing less than $500.

To summarize, the findings of this study point to a number of empirical questions for future research. First, researchers could investigate the impact of specific state-level child support policies (e.g., automatic referral to child support from Medicaid, retroactive support, and interest rates) on the risk for jail. Second, researchers could use qualitative or mixed-methods studies to examine where nonpayment of child support is due to unwillingness to pay, inability to pay, or some combination of the two. Finally, researchers should look at the consequences of jailing for child support nonpayment and whether these consequences are similar or different from incarceration for other reasons.

This study contributes to the discussion on poverty governance by examining how a third, relatively understudied institution, the child support enforcement system, is connected to the welfare and criminal justice systems. Receiving public assistance requires families to interact with the formal child support system, and parents who owe child support can be sent to jail for this debt. Contempt of court for child support nonpayment is a civil, not a criminal, offense; therefore the child support enforcement system constitutes part of the shadow carceral state, extending the capacity of the state to deprive individuals of liberty, even if they have not committed criminal offenses (Beckett and Murakawa 2012). The child support enforcement system is a potentially large dragnet into the criminal justice system. Although there are almost 7 million individuals under correctional supervision today (Kaeble and Glaze 2016), there are more than 11 million individuals who owe child support debt (Office of Child Support Enforcement 2014a). Because about half of all children spend time living without one parent (Nepomnyaschy and Garfinkel 2010), child support enforcement is an institution of poverty governance with a potentially large reach that deserves greater study. This is one of the first studies to investigate what factors affect one’s likelihood of spending time in jail for child support debt.

Finally, this study has wide implications. I find that about one in seven fathers who owe child support spends time in jail for it. Because there are more than 11 million child support debtors, this could constitute a significant cost to taxpayers. Furthermore, there is a large literature on the collateral consequences of incarceration, showing negative effects of incarceration on earnings, employment, relationships, and child health and well-being (Pager 2003; Turney and Wildeman 2013; Wakefield and Wildeman 2014). Child support enforcement aims to increase child well-being by ensuring that noncustodial fathers contribute to children’s material well-being. Yet owing child support debt puts nonresident fathers at risk for going to jail, triggering potentially negative collateral consequences. Understanding more about jail for child support nonpayment, therefore, is important for child well-being.

Footnotes

Appendix

Table A8.

Full Regression Table Predicting Jail.

	Odds Ratio (SE)
	Model 1	Model 2	Model 3	Model 4
Wave	1.087⁺	1.038	1.051	1.003
Wave	(0.055)	(0.057)	(0.050)	(0.050)
Public assistance history
Mother has ever received TANF	0.932		0.758	0.733
Mother has ever received TANF	(0.232)		(0.206)	(0.198)
Mother has received Medicaid by wave 3	1.406		1.312	1.305
Mother has received Medicaid by wave 3	(0.300)		(0.282)	(0.285)
Amount of CS debt (reference: between $1 and $499)
$500–$1,000				1.462
$500–$1,000				(0.734)
$1,001–$2,000				1.677
$1,001–$2,000				(0.808)
$2,001–$3,000				1.536
$2,001–$3,000				(0.787)
$3,001–$4,000				1.600
$3,001–$4,000				(0.906)
$4,001–$5,000				1.728
$4,001–$5,000				(0.931)
$5,001–$10,000				2.256⁺
$5,001–$10,000				(1.084)
More than $10,000				3.849**
More than $10,000				(1.838)
Relationship factors
Relationship quality		1.329	1.278	1.356⁺
Relationship quality		(0.267)	(0.211)	(0.226)
Either parent has ever been in a new relationship		0.758	0.759	0.741⁺
Either parent has ever been in a new relationship		(0.131)	(0.132)	(0.129)
Mother has MPF		0.991	1.035	1.045
Mother has MPF		(0.166)	(0.165)	(0.169)
Father has MPF		1.650*	1.658**	1.619*
Father has MPF		(0.325)	(0.319)	(0.315)
Controls
Father worked last week	1.249	1.073	1.539	1.733⁺
Father worked last week	(0.399)	(0.158)	(0.482)	(0.549)
Relationship at baseline (reference: not romantic)
Married	0.703	0.898	0.858	0.806
Married	(0.274)	(0.386)	(0.338)	(0.321)
Cohabiting	0.918	0.929	0.906	0.871
Cohabiting	(0.145)	(0.147)	(0.142)	(0.139)
Any domestic violence (kick, slap, serious injury)	0.898	0.877	0.875	0.888
Any domestic violence (kick, slap, serious injury)	(0.295)	(0.243)	(0.233)	(0.240)
Father in jail at baseline	0.581	0.475⁺	0.523⁺	0.499⁺
Father in jail at baseline	(0.211)	(0.197)	(0.193)	(0.185)
Father’s household income	1.000	1.000	1.000	1.000
Father’s household income	0.000	0.000	0.000	0.000
Father’s education at baseline (reference: LTHS)
High school degree	0.641*	0.758	0.722	0.745
High school degree	(0.129)	(0.196)	(0.159)	(0.166)
Some	0.857	0.940	0.897	0.862
college/trade school	(0.187)	(0.239)	(0.202)	(0.194)
College degree	1.493	1.887⁺	1.805⁺	1.920*
College degree	(0.416)	(0.684)	(0.579)	(0.629)
Father age at baseline	0.999	0.994	0.994	0.994
Father age at baseline	(0.013)	(0.015)	(0.014)	(0.014)
Father’s race (reference: white)
Father is black	1.142	1.078	1.135	1.168
Father is black	(0.293)	(0.257)	(0.289)	(0.304)
Father is Hispanic	0.860	0.902	0.897	0.903
Father is Hispanic	(0.270)	(0.291)	(0.292)	(0.291)
Father is other race	1.878	1.806	1.860	1.980⁺
Father is other race	(0.758)	(0.729)	(0.767)	(0.816)
Constant	0.054*	0.010*	0.008*	0.004**
Constant	(0.076)	(0.021)	(0.016)	(0.008)
Observations	2,046	2,046	2,046	2,046

Note: Robust seeform in parentheses. All models control for the predicted probability of owing child support debt, state of residence, relationship at baseline, whether father was in jail at baseline, father’s baseline age and education, whether father worked last week, father’s household income, and father’s race. CS = child support; LTHS = less than high school; MPF = multipartner fertility; TANF = Temporary Assistance for Needy Families.

p < .10. *p < .05. **p < .01.

Acknowledgements

Thank you to Jennifer Glass for invaluable mentorship and guidance on this article and the larger dissertation of which it is a part. Thanks also to Frank Edwards, Ken Hou-Lin, Carmen Gutierrez, and Chelsea Smith for comments and modeling advice. I also thank the Family Demography Working Group in the University of Texas Population Research Center for their feedback on this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research has been supported by the National Science Foundation (NSF DDRIG 1628128) and the National Institutes of Health (National Institute of Child Health and Human Development R24 grant R24HD042849 made to the University of Texas Population Research Center and National Institute of Child Health and Human Development training grant T32HD007081.)

1

See 18 U.S.C. § 228 ().

2

Since Medicaid is a much bigger programs in terms of receipt than TANF (Sugie 2012), the number of child support cases with Medicaid involvement is likely much higher, but, caseload by Medicaid status is not reported by the .

3

About four of every five NCPs are fathers (), so I use masculine pronouns in this article for ease of understanding for the reader. Furthermore, the pool of female NCPs in the FFCW is too small for analysis. However, female NCPs are likewise at risk for incarceration for child support nonpayment and may face unique vulnerabilities.

4

In some states, individuals are referred to child support enforcement for their Medicaid involvement too.

5

Other analyses of the FFCW show that initial levels of informal support are higher than formal support and that over time, formal support provided does not reach the initial levels of informal support (). Thus, the breakdown of the parental relationship that goes along with the establishment of a formal support order may also signal a decline in the value of support provided to children.

6

I find significant differences on all key variables of interest on the basis of whether a respondent is part of the national FFCW sample. Table A1 presents the results of t tests demonstrating the differences on key variables between respondents who are and are not missing on a FFCW national probability weight and presents weighted means for the descriptive statistics (Bzostek, McLanahan, and Carlson 2012; ). Because of these systematic differences, I present unweighted estimates.

7

Because my outcome variables are related, it is possible that the error terms of these three regressions might be correlated. To address any disturbances across the three regressions, I redid the models as seemingly unrelated regressions (using the Stata suest command for binary outcomes) as an additional sensitivity check. This allowed me to estimate all three of the dependent variables simultaneously, to allow for the error terms to be correlated across the three outcomes. The results were largely the same (see ).

8

To further account for the selection of parents into my sample, I also used coarsened exact matching to account for the selection of parents into these groups. Parents were matched on age, race, education, and whether they made it through the prior step (e.g., for the debt regression, parents were matched on having a formal child support order). Results of these models are substantively the same, so I do not show them (available upon request).

9

The FFCW is based on a sample of large cities. As a sensitivity analysis, I ran the models with city instead of state fixed effects. Results are largely the same (see ). Because child support policy is made at the state level, I report results from the state fixed effects models in this article.

10

In their description of the FFCW sampling procedure, Reichman et al. (2001) described how cities were sampled on the basis of the stringency of their welfare and child support regimes. Comparing with the original classification shows that most of the cities identified by survey designers as punitive or lenient are also punitive or lenient when it comes to jailing, with a few exceptions (Austin incarcerates more debtors, and Boston and Detroit each incarcerate fewer debtors than would be expected on the basis of their 2001 child support stringency ratings).

11

A quick note on causality: It is plausible that father’s child support debt (or his child support nonpayment that leads to this debt) could impel mother to apply for public assistance, instead of mother’s public assistance receipt increasing the likelihood of father’s child support debt (). As a sensitivity analysis, therefore, I also ran models that lagged mother’s TANF receipt by one wave (available on request). The results were robust, so I have confidence that causality is operating in the hypothesized direction.

12

I do not find any main effects for race (see Tables A6 –). To further investigate this question, I conducted a series of interaction tests to see if the effect of relationship context or public assistance varied by race. I interacted each of the focal public assistance and relationship context variables by the dummy variables for Hispanic, black, and white. None of the Hispanic interactions was significant. Turning to the black interactions, Relationship Quality × Black was significant and negative for both the debt and formal regressions. This suggests that an increase in relationship quality for blacks is especially protective against entrance into the formal system or debt accumulation, compared with fathers who are not black. For the white interactions, Being in a New Relationship × White was significant and positive for the jail regression. Yet only 3 percent of the sample fit into this category, suggesting that this significant effect could be because of small cell sizes.

13

Although there is variation in the percentage of debtors incarcerated for child support by city and state (see Table A4), the role of judicial discretion makes it unlikely that state or city completely determines outcomes. Because the FFCW is a sample of large cities, many of the IV-D courts in these cities are likely presided over by more than one judge. One foundational study of child support establishment found great disparities in outcomes across individual judges who presided over the same jurisdiction in Denver (). Because contempt of court is so dependent on judicial discretion, and because the FFCW cities are so large, I am confident that there remains room for judicial discretion within cities and within states, even if we see some variation across cities and states.

14

If more advantaged fathers are more likely to go to jail, this finding could provide some insight into judicial discretion. Judges who see employed or educated fathers may be more likely to interpret their nonpayment as willful and jail them.

15

Some states, such as Maryland, provide a lawyer throughout the enforcement process. In others, such as Texas, qualifying low-income NCPs are not informed of their right to counsel until incarceration becomes a possible outcome. Still other states never provide a court-appointed attorney, because NCPs should not be held in contempt unless they have the ability to pay. If they have the ability to pay their child support debt, they also must have the ability to pay for their own counsel (U.S. Department of Health and Human Services, Administration for Children and Families 2002).

Author Biography

Elizabeth Cozzolino is a PhD candidate in sociology and a Population Research Center trainee at the University of Texas at Austin. Her dissertation is a multimethod study of incarceration for child support nonpayment. More about her and her work is available at .

References

Alexander

M. Jacqui

. 2005. Pedagogies of Crossing: Mediations on Feminism, Sexual Politics, Memory, and the Sacred. Durham, NC: Duke University Press.

Alexander

Michelle

. 2012. The New Jim Crow: Mass Incarceration in the Age of Colorblindness. New York: New Press.

Allison

Paul

. 2009. “Missing Data.” Pp. 72–89 in The SAGE Handbook of Quantitative Methods in Psychology, edited by Millsap

R. E.

Maydeu-Olivares

. Thousand Oaks, CA: Sage.

Bartfeld

Judi

Meyer

Daniel R.

. 2003. “Child Support Compliance among Discretionary and Nondiscretionary Obligors.” Social Service Review 77(3):347–72.

Beckett

Katherine

Murakawa

Naomi

. 2012. “Mapping the Shadow Carceral State: Toward an Institutionally Capacious Approach to Punishment.” Theoretical Criminology 16(2):221–44.

Beckett

Katherine

Western

Bruce

. 2001. “Governing Social Marginality: Welfare, Incarceration, and the Transformation of State Policy.” Punishment & Society 3(1):43–59.

Bogenschneider

Karen

. 2000. “Has Family Policy Come of Age? A Decade Review of the State of US Family Policy in the 1990s.” Journal of Marriage and the Family 62(4):1136–59.

Boynton-Jarrett

Renee

Fargnoli

Jessica

Suglia

Shakira Franco

Zuckerman

Barry

Wright

Rosalind J.

. 2010. “Association between Maternal Intimate Partner Violence and Incident Obesity in Preschool-aged Children.” Archives of Pediatrics and Adolescent Medicine 164(6):540–6.

Bzostek

Sharon H.

McLanahan

Sara S.

Carlson

Marcia J.

. 2012. “Mothers’ Repartnering after a Nonmarital Birth.” Social Forces 90(3):817–41.

10.

Chambers

David L.

1979. Making Fathers Pay: The Enforcement of Child Support. Chicago: University of Chicago Press.

11.

Collins

Jane L.

Mayer

Victoria

. 2010. Both Hands Tied: Welfare Reform and the Race to the Bottom in the Low-wage Labor Market. Chicago: University of Chicago Press.

12.

Cook

Steven

. 2015. “Child Support Enforcement Use of Contempt and Criminal Nonsupport Charges in Wisconsin.” Institute for Research on Poverty. Retrieved January 30, 2018 (https://www.irp.wisc.edu/research/childsup/cspolicy/pdfs/2009-11/CS-2009-11-Task7B-Cook.pdf).

13.

Cook

Steven

Noyes

Jennifer L.

. 2011. The Use of Civil Contempt and Criminal Nonsupport as Child Support Enforcement Tools: A Report on Local Perspectives and the Availability of Data. Madison, WI: Institute for Research on Poverty.

14.

Coven

Mark S.

1997. “Welfare Reform, Contempt and Child Support Enforcement.” Suffolk University Law Review 30:1067–95.

15.

Cozzolino

Elizabeth

Williams

Christine L.

. 2017. “Child Support Queens and Disappointing Dads: Gender and Child Support Compliance.” Social Currents 4(3): 228–45.

16.

Craigie

Terry-Ann L.

2015. “Multipartner Fertility and Child Support.” Eastern Economic Journal 41(4):571–91.

17.

Demuth

Stephen

. 2003. “Racial and Ethnic Differences in Pretrial Release Decisions and Outcomes: A Comparison of Hispanic, Black, and White Felony Arrestees.” Criminology 41(3):873–908.

18.

Demuth

Stephen

Steffensmeier

Darrell

. 2004. “The Impact of Gender and Race-ethnicity in the Pretrial Release Process.” Social Problems 51(2):222–42.

19.

Doolittle

Fred

Lynn

Suzanne

. 1998. Working with Low Income Cases: Lessons for the Child Support Enforcement System from Parents’ Fair Share. New York: Manpower Demonstration Research Corporation.

20.

Edin

Kathryn

. 2000. “What Do Low-income Single Mothers Say about Marriage?” Social Problems 47(1):112–33.

21.

Edin

Kathryn

Nelson

Timothy J.

. 2013. Doing the Best I Can: Fatherhood in the Inner City. Berkeley: University of California Press.

22.

Elmore

Cindy

. 2010. “On and On, Over and Over: The Gender War in Child Support Enforcement Court.” Feminist Studies 36(2):397–403.

23.

Garfinkel

Irwin

Klawitter

Marieka M.

. 1990. “The Effect of Routine Income Withholding of Child Support Collections.” Journal of Policy Analysis and Management 9(2):155–77.

24.

Goffman

Alice

. 2014. On the Run: Fugitive Life in an American City. Chicago: University of Chicago Press.

25.

Goldstein

Joshua R.

Harknett

Kristin

. 2006. “Parenting across Racial and Class Lines: Assortative Mating Patterns of New Parents Who Are Married, Cohabiting, Dating, or no Longer Romantically Involved.” Social Forces 85(1):121–43.

26.

Hamer

Jennifer

. 2001. What It Means to Be Daddy: Fatherhood for Black Men Living Away from Their Children. New York: Columbia University Press.

27.

Harris

Alexes

. 2016. A Pound of Flesh: Monetary Sanctions as Punishment for the Poor. New York: Russel Sage.

28.

Harris

Alexes

Evans

Heather

Beckett

Katherine

. 2010. “Drawing Blood from Stones: Legal Debt and Social Inequality in the Contemporary United States.” American Journal of Sociology 115(6):1753–99.

29.

Haskins

Anna R.

2016. “Beyond Boys’ Bad Behavior: Paternal Incarceration and Cognitive Development in Middle Childhood.” Social Forces 95(2):861–92.

30.

Hays

Sharon

. 2004. Flat Broke with Children: Women in the Age of Welfare Reform. Oxford, UK: Oxford University Press.

31.

Huang

Mincy

R. B.

Garfinkel

. 2005. “Child Support Obligations and Low-income Fathers.” Journal of Marriage and Family 67(5):1213–25.

32.

Josephson

Jyl J.

1997. Gender, Families and State: Child Support Policy in the United States. Lanham, MD: Rowman & Littlefield.

33.

Kaeble

Danielle

Glaze

Lauren

. 2016. “Correctional Population in the United States, 2015.” Bureau of Justice Statistics. Retrieved (https://www.bjs.gov/content/pub/pdf/cpus15.pdf).

34.

Kane

Jennifer B.

Nelson

Timothy J.

Edin

Kathryn

. 2015. “How Much In-kind Support Do Low-income Nonresident Fathers Provide? A Mixed-method Analysis.” Journal of Marriage and Family 77(3):591–611.

35.

Lin

I-Fen

. 2000. “Perceived Fairness and Compliance with Child Support Obligations.” Journal of Marriage and Family 62(2):388–98.

36.

May

Rebecca

Roulet

Marguerite

. 2005. “A Look at Arrests of Low-income Fathers for Child Support Nonpayment: Enforcement, Court and Program Practices.” Center for Family Policy and Practice. Retrieved January 30, 2018 (https://cffpp.org/wp-content/uploads/LookAtArrests.pdf).

37.

Meyer

Daniel R.

Cancian

Maria

. 2012. “‘I’m Not Supporting His Kids’: Nonresident Fathers’ Contributions Given Mothers’ New Fertility.” Journal of Marriage and Family 74(1):132–51.

38.

Meyer

Daniel R.

Cancian

Maria

Cook

Steven T.

. 2005. “Multi-partner Fertility: Incidence and Implications for Child Support Policy.” Social Service Review 79(4):577–601.

39.

Mincy

Ronald B.

Sorenson

Elaine J.

. 1998. “Deadbeats and Turnips in Child Support Reform.” Journal of Policy Analysis and Management 17(1):44–51.

40.

Nepomnyaschy

Lenna

Garfinkel

Irwin

. 2010. “Child Support Enforcement and Fathers’ Contributions to Their Nonmarital Children.” Social Service Review 84(3):341–80.

41.

Office of Child Support Enforcement. 2014a. “FY 2014 Preliminary Report.” U.S. Department of Health and Human Services, Administration for Children and Families. Retrieved (http://www.acf.hhs.gov/sites/default/files/programs/css/fy2014_preliminary.pdf).

42.

Office of Child Support Enforcement. 2014b. “Major Change in Who Is Owed Child Support Arrears.” U.S. Department of Health and Human Services, Administration for Children and Families. Retrieved January 30, 2018 (https://www.acf.hhs.gov/sites/default/files/programs/css/changes_in_who_is_owed_arrears.pdf).

43.

Office of Inspector General. 2000. “The Establishment of Child Support Orders for Low Income Non-custodial Parents.” U.S. Department of Health and Human Services. Retrieved (https://oig.hhs.gov/oei/reports/oei-05-99-00390.pdf).

44.

Pager

Devah

. 2003. “The Mark of a Criminal Record.” American Journal of Sociology 108(5):937–75.

45.

Patterson

Elizabeth G.

2008. “Civil Contempt and the Indigent Child Support Obligor: The Silent Return of Debtor’s Prison.” Cornell Journal of Law and Public Policy 18:95–141.

46.

Perkinson

Robert

. 2010. Texas Tough: The Rise of America’s Prison Empire. New York: Picador.

47.

Pew Research Center. 2013. “Asian-Americans Lead All Others in Household Income.” Retrieved (http://www.pewresearch.org/daily-number/asian-americans-lead-all-others-in-household-income/).

48.

Reichman

Nancy

Teitler

Julien O.

Garfinkel

Irwin

McLanahan

Sara S.

. 2001. “Fragile Families: Sample and Design.” Children and Youth Services Review 23(4/5):303–16.

49.

Reynolds

Carl

Cowherd

Mary

Barbee

Andy

Fabelo

Tony

Wood

Ted

Yoon

Jamie

. 2009 “A Framework to Improve How Fines, Fees, Restitution and Child Support Are Assessed and Collected from People Convicted of Crimes: Interim Report.” Lexington, KY: Justice Center: The Council of State Governments.

50.

Rich

Lauren M.

Garfinkel

Irwin

Gao

Qin

. 2007. “Child Support Enforcement Policy and Unmarried Fathers’ Employment in the Underground and Regular Economies.” Journal of Policy Analysis and Management 26(4):791–810.

51.

Roberts

Paula

. 2001. “An Ounce of Prevention and a Pound of cure: Developing State Police on the Payment of Child Support Arrears by Low-income Parents.” Washington, DC: Center for Law and Social Policy.

52.

Solomon-Fears

Carmen

. 2012. “Analysis of Federal-state Financing of the Child Support Enforcement Program.” Congressional Research Service RL334422. Retrieved January 30, 2018 (https://greenbook-waysandmeans.house.gov/sites/greenbook.waysandmeans.house.gov/files/2012/documents/RL33422_gb.pdf).

53.

Sorenson

Elaine

Sousa

Liliana

Schaner

Simon

. 2007. “Assessing Child Support Arrears in Nine Large States and the Nation.” The Urban Institute. Retrieved January 30, 2018 (https://www.urban.org/sites/default/files/publication/29736/1001242-Assessing-Child-Support-Arrears-in-Nine-Large-States-and-the-Nation.PDF).

54.

Sorenson

Elaine

Zibman

Chava

. 2001. “Getting to Know Poor Fathers Who Do Not Pay Child Support.” Social Service Review 75(3):420–35.

55.

Soss

Joe

. 1999. “Lessons of Welfare: Policy Design, Political Learning, and Political Action.” American Political Science Review 93(2):363–80.

56.

Soss

Joe

Fording

Richard C.

Schram

Sanford F.

. 2008. “The Color of Devolution: Race, Federalism, and the Politics of Social Control.” American Journal of Political Science 52(3):536–53.

57.

Steffensmeier

Darrell

Demuth

Stephen

. 2001. “Ethnicity and Judges’ Sentencing Decisions: Hispanic-black-white Comparisons.” Criminology 39(1):145–78.

58.

Sugie

Naomi F.

2012. “Punishment and Welfare: Paternal Incarceration and Families’ Receipt of Public Assistance.” Social Forces 90(4):1403–27.

59.

Turner

Kimberly J.

Waller

Maureen R.

. 2017. “Indebted Relationships: Child Support Arrears and Nonresident Fathers’ Involvement with Children.” Journal of Marriage and Family 79(1):24–43.

60.

Turney

Kristin

. 2011. “Labored Love: Examining the Link between Maternal Depression and Parenting Behaviors.” Social Science Research 40(1):399–415.

61.

Turney

Kristin

. 2014a. “The Consequences of Paternal Incarceration for Maternal Neglect and Harsh Parenting.” Social Forces 92(4):1607–36.

62.

Turney

Kristin

. 2014b. “The Intergenerational Consequences of Mass Incarceration: Implications for Children’s Co-residence and Contact with Grandparents.” Social Forces 93(1):299–327.

63.

Turney

Kristin

Wildeman

Christopher

. 2013. “Redefining Relationships: Explaining the Countervailing Consequences of Paternal Incarceration for Parenting.” American Sociological Review 78(6):949–79.

64.

U.S. Census Bureau. 2012. “Custodial Mothers and Fathers and Their Child Support: 2011 Detailed Tables” (Series P60-246). Retrieved (http://www.census.gov/people/childsupport/data/files/chldsu11.pdf).

65.

U.S. Department of Health and Human Services, Administration for Children and Families. 2002. “Chapter Ten: Enforcement of Support Obligations.” Pp. 207–310 in Essentials for Attorneys in Child Support Enforcement, 3rd ed. Washington, DC: U.S. Department of Health and Human Services.

66.

U.S. Department of Justice. 2012. “Citizen’s Guide to U.S. Federal Law on Child Support Enforcement: 18 U.S.C. §228.” Retrieved January 30, 2018 (https://www.justice.gov/criminal-ceos/citizens-guide-us-federal-law-child-support-enforcement).

67.

Wacquant

Loic J.

2009. Prisons of Poverty (Vol. 23). Minneapolis: University of Minnesota Press.

68.

Wakefield

Sara

Wildeman

Christopher James

. 2014. Children of the Prison Boom: Mass Incarceration and the Future of American Inequality. Oxford, UK: Oxford University Press.

69.

Waller

Maureen R.

2002. My Baby’s Father: Unmarried Parents and Paternal Responsibility. Ithaca, NY: Cornell University Press.

70.

Waller

Maureen R.

Plotnick

Robert

. 2001. “Effective Child Support Policy for Low-income Families: Evidence from Street Level Research.” Journal of Policy Analysis and Management 20(1):89–110.

71.

Western

Bruce

Wildeman

Christopher

. 2009. “The Black Family and Mass Incarceration.” The Annals of The American Academy of Political and Social Science 621(1):221–42.

72.

Wildeman

Christopher

. 2010. “Paternal Incarceration and Children’s Physically Aggressive Behaviors: Evidence from the Fragile Families and Child Wellbeing Study.” Social Forces 89(1):285–309.

73.

Yee

Lucy Marsh

. 1979. “What Really Happens in child Support Cases: An Empirical Study of Establishment and Enforcement of Child Support Orders in the Denver District Court.” Denver Law Journal 57(1):21–68.

Public Assistance,Relationship Context,and Jail for Child Support Debt

Abstract

Keywords

Background

Poverty Governance: Incarceration, Welfare, and Child Support Enforcement

Pathways: Public Assistance and Relationship Context

Formal Child Support

Child Support Debt

Jail for Child Support Nonpayment

Data and Method

Data Source

Analytic Strategy

Measures

Dependent Variables

Independent Variables

Public Assistance

Relational Variables

Controls

Results

Descriptive Results

Formal Child Support Order

Child Support Debt

Jail for Child Support Nonpayment

Discussion and Conclusion

Footnotes

Appendix

Acknowledgements

Funding

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Author Biography

References