Child poverty and child support policy: A comparative analysis of Colombia and the United States

Child poverty and child support policy: A comparative analysis of Colombia and the United States

Children and Youth Services Review 93 (2018) 143–153 Contents lists available at ScienceDirect Children and Youth Services Review journal homepage: ...

289KB Sizes 0 Downloads 42 Views

Children and Youth Services Review 93 (2018) 143–153

Contents lists available at ScienceDirect

Children and Youth Services Review journal homepage: www.elsevier.com/locate/childyouth

Child poverty and child support policy: A comparative analysis of Colombia and the United States

T



Laura Cuestaa, , Daniel R. Meyerb a b

School of Social Work, Rutgers, The State University of New Jersey, 536 George St., Room 205A, New Brunswick, NJ 08901, United States School of Social Work, University of Wisconsin-Madison, 1350 University Avenue, Madison, WI 53706, United States

A R T I C LE I N FO

A B S T R A C T

Keywords: Child poverty Child support Cross-national research

Child support may merely reshuffle poverty, reducing child poverty among families who receive it at the expense of the economic well-being of children living with a nonresident parent. Our study examines child support's effects on child poverty, considering those who pay child support and those who receive child support, and doing so in Colombia and the United States (U.S.). We use data from the Colombian Longitudinal Survey (N = 13,036) and the U.S. Current Population Survey (N = 53,480). Our findings show that the antipoverty effectiveness of child support among resident parent families is larger in Colombia than in the U.S. Child support payments do increase child poverty among children living in payer families in both countries, but the effects are fairly small. In our base models, 6%–9% of children in nonresident parent families are falling into poverty after child support payments are transferred to other families. Overall, child support receipts decrease poverty to a greater extent than child support payments increase it among children.

1. Introduction Child support from a nonresident parent (NRP) is an important policy area given the high poverty experienced by children living with only one of their parents (the resident parent, or RP) and limitations and cutbacks on governmental support for families. A substantial amount of child support research has been conducted (Huang & Han, 2012; Pirog & Ziol-Guest, 2006); we have begun to understand some of the child support policy schemes in use across countries (e.g., Meyer & Skinner, 2016; Meyer, Skinner, & Davidson, 2011; Skinner & Davidson, 2009), the importance of child support for economic well-being among resident-parent families (e.g., Cuesta & Meyer, 2014; Hakovirta, 2011; OECD, 2011; Skinner & Meyer, 2006; Sorensen, 2016), and whether child support receipt is associated with lower poverty among children living with RPs (e.g., Bartfeld, 2000; Cuesta & Meyer, 2014; Hakovirta, 2011; Meyer & Hu, 1999; Nichols-Casebolt, 1986; Skinner, Cook, & Sinclair, 2017a). The research examining the relationship between child support receipt and child poverty follows the traditional method for examining the antipoverty effectiveness of an income source by comparing poverty rates when child support is or is not included in income. But the method is better suited to an examination of government transfers, not child support. Child support is not money from the government, but money from another parent (the NRP). This means that child support may ⁎

actually increase poverty among those who pay, even as it decreases poverty among those who receive. Those paying may have children living with them, so a complete accounting of the relationship between child support and child poverty needs to examine the economic wellbeing of those in families that pay as well as those in families that receive. The focus on children in RP families is understandable and stems from the fact that historically most children have stayed with their mothers after their parents' separate or divorce (Buehler & Gerard, 1995) and that women are more likely than men to experience poverty or income declines following marital dissolution (e.g., Bartfeld, 2000; Bianchi, Subaiya, & Kahn, 1999; Devaus, Gray, Qu, & Stanton, 2015). A recent analysis for countries in the Organization for Economic Co-operation and Development (OECD) shows that 85% of sole-parent households were headed by women in 2010 (OECD, 2010). Yet, several countries are experiencing significant transformations in union formation and union dissolution, re-partnering, and childbearing across partnerships (e.g., Thomson, Lappegard, Carlson, Evans, & Gray, 2014). These demographic changes have increased the number of parents rearing biological and step-children across multiple families and, as a result, a large number of children are sharing parents' time and money with step-siblings and or half-siblings in a number of countries (Cancian, Meyer, & Cook, 2011a; Carlson & Furstenberg, 2006; Thomson et al., 2014). Estimates from the United States (U.S.) are that

Corresponding author. E-mail address: [email protected] (L. Cuesta).

https://doi.org/10.1016/j.childyouth.2018.07.013 Received 2 March 2018; Received in revised form 13 July 2018; Accepted 13 July 2018 Available online 17 July 2018 0190-7409/ © 2018 Elsevier Ltd. All rights reserved.

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

2.2. Poverty trends in Colombia and the U.S.

33–40% of non-resident fathers are living with resident children, primarily their biological children, although step-children are also common (Garfinkel, McLanahan, & Hanson, 1998a; Manning & Smock, 2000). When NRPs have co-resident children as well as nonresident children, this highlights the possibility that child support merely reshuffles poverty, reducing child poverty among families who receive it at the expense of the economic well-being of children living in NRP families. There are two gaps in the extant literature that we address in this study. First, we know little about the relationship between child support and the poverty of all children, including those living with a NRP as well as those living with an RP. A second gap is that most of the previous literature is focused on a few rich countries (e.g., Hakovirta, 2011; OECD, 2011), with relatively little work on developing countries. We examine child support's effects on child poverty, considering those who pay child support and those who receive child support, and doing so in Colombia and the U.S. The specific aims of the study are: (1) to examine the extent to which families with children rely on earnings, child support, and other income sources, and whether child support payments represent a significant nondiscretionary expenditure for these families; (2) to estimate the antipoverty effectiveness of child support after considering both child support receipt and child support payments; and (3) to examine the extent to which those already poor have their poverty gap increase by paying child support. To answer these questions we use standard methods adapted to include paying support. We answer each question in Colombia and the U.S. separately, comparing the responses.

Colombians have experienced a significant improvement in their economic well-being over the past two decades. The national poverty rate dropped from 49.7% in 2002 to 28% in 2016 (DANE, 2016). This decline has been linked to a positive economic environment (Barrientos, Ramírez, & Tabares, 2015; UNDP, 2016) and the expansion of the social safety net (Fiszbein & Schady, 2009). However, differences in poverty rates by location persist and Colombians living in rural areas are more likely to be poor (38.6%) than those residing in urban districts (24.6%) (DANE, 2016). Differences by household composition and family structure are also significant. In 2016, the poverty rate among families with three or more children under 12 years old was 66.7%, and among families with two children was 40.7% (DANE, 2016). A recent study that estimated the impact of family change on income poverty also finds that almost half of single-mother families (49.5%) were poor in 2012 while the poverty rate among married-couple families was 30% (Cuesta, Rios-Salas, & Meyer, 2017). While significantly lower than the rates observed in Colombia during the same period of time, the official poverty rates in the U.S. rose from 11.7% in 2001 to 15% in 2011. This increase was one of the various consequences of the severe recession that Americans experienced during the late 2000s. However, this upward trend was followed by a recovery; in 2016, the poverty rate was 12.7% (Semega, Fontenot, & Kollar, 2017). Poverty in the U.S. is still higher than pre-recession levels and is higher than in most OECD countries. Based on a poverty threshold set at 50% of the median income, the OECD estimates that 17.2% of all Americans were living in poverty in 2013 (OECD, 2017). The same analysis also shows that American children are faring worse than children in other developed countries: one in every five American children were living in poverty in 2013 (OECD, 2017). Differences by race and family structure are also particularly significant in the U.S. In 2015, the poverty rate among African Americans was more than twice (22.0%) that of non-Hispanic Whites (8.8%) and the percentage of single-mother families in poverty was more than twice the rate observed in the general population (Semega et al., 2017). Given the high poverty rate experienced by single-parent families in Colombia and the United States, child support policy has the potential to improve the economic well-being of children living in single-parent families in both countries. However, neither of these countries has an explicit antipoverty agenda in their child support policy schemes. An exception that we discuss further in the next section of the paper is that RPs receiving public assistance in the U.S. are required to cooperate with the child support system. While this could lead to a focus on poverty reduction, instead the focus is generally on recovering public costs (Skinner, Meyer, Cook, & Fletcher, 2017b).

2. Country context 2.1. Selected demographic trends in Colombia and the U.S. Like most countries in Latin America, cohabiting unions are very common in Colombia (Esteve et al., 2012; Profamilia, 2016). Given the low prevalence of marriage, it is not surprising that in 2010, 84% of children were born to unmarried women (Child Trends, 2015). While divorce rates have remained at a 14% low over the past decade (Profamilia, 2016), a recent study finds the type of family with children that exhibited the highest growth rate between 1997 and 2008 were repartnered-mother families (i.e., families in which there was a mother, a mother's new spouse or partner and at least one child whose father is alive but living somewhere else) (Cuesta & Meyer, 2014). Some of the new partners may have children elsewhere that they are supporting, so increases to the economic well-being of their old family may come at the expense of their new family. Unfortunately there are no national estimates of re-partnering or the extent to which nonresident parents have resident children as well as nonresident children. In the U.S., family change has been driven by dramatic transformations in union formation, union dissolution, and fertility. While marriage is more prevalent in the U.S. than it is in Colombia, the proportion of married-couple households has dropped systematically since 1960 and represented less than half of all U.S. households in 2014 (Jacobsen, Mather, & Dupuis, 2012); this decline in marriage has been accompanied by a significant increase in cohabitation (Kroeger & Smock, 2014). One distinctive feature of cohabiting unions in the U.S. is that they tend to be short-lived (Andersson, Thomson, & Duntava, 2017); because the divorce rate has remained fairly stable over the past three decades, most of the union instability seen in the U.S. in recent decades comes from cohabitation disruption. Another significant demographic trend observed in this country is the dramatic increase in non-marital births: non-marital births rose from 10.7% of births in 1970 to 40.1% in 2015 (Martin, Hamilton, Osterman, Driscoll, & Mathews, 2017). The combination of union instability and re-partnering has ultimately increased childbearing across partnerships (Thomson et al., 2014), and the United States has one of the highest rates of parental repartnering among developed countries (Andersson et al., 2017).

2.3. The child support policy schemes in Colombia and the U.S. The Colombian child support policy scheme has evolved from an adversarial model in which family courts were the main actors of the system to a mediation model that involves three institutions, the judicial system, a government agency called the National Institute of Family Well-being, and local governments. If the parents were married, child support arrangements are generally made through legal divorce proceedings. If the child was born to an unmarried woman, she can initiate action privately or work with the National Institute of Family WellBeing or a family court. Private arrangements are enforceable by the child support system although as we discuss below, Colombia does not have the type of enforcement mechanisms that we observe in developed countries like the U.S. More information on the Colombian child support system can be found in Cuesta and Meyer (2012). Here we provide a brief overview. The child support system does not intervene unless the RP explicitly requests assistance. The National Institute of Family Well-Being provides free services but these are not available in all areas. The child 144

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

payments for most parents; if the NRP does not comply with the order, penalties for noncompliance include reporting the NRP to credit bureaus, interception of windfall income and tax refunds, loss of occupation and driver's license and criminal nonsupport charges (Cuesta & Meyer, 2012). Under the current system, approximately half of all NRPs had either informal (private) or legal agreements in 2015; of the RPs who were due child support, 69% were receiving either full or partial payments (Grall, 2018). Two characteristics of the U.S. child support system limit the extent to which child support decreases child poverty (Skinner et al., 2017). First, RPs participating in public assistance programs do not receive all child support paid on behalf of their children (some or all is retained by the government to offset welfare costs (National Conference of State Legislators, 2017) and second, child support income is counted for determination of benefit eligibility in other programs, limiting the extent to which child support paid actually increases the incomes of RPs.

support system is generally based on negotiation: the parents (sometimes with the assistance of the National Institute of Family Well-Being or the local government) work together to establish the amount, regularity, and type of support that the NRP is going to provide, although if parents cannot reach an agreement, child support may be ordered through an administrative action. There are no specific guidelines for the amount of support ordered except that it cannot be higher than 50% of the NRP's monthly wage. Child support amounts tend to be based on what children “need” and thus are relatively higher for children in lower-income families, who are seen as needing more (Cuesta & Meyer, 2012; Cuesta & Meyer, 2014). The mechanism for collection of child support is also established by the parents. Wage withholding only occurs if the RP pursues child support through a family court and the judge makes this determination. When wages are withheld, the court forwards the payment to the judicial system and the judicial system issues a check to the RP. If the NRP does not pay, no enforcement actions occur unless the RP requests the intervention of the system. If so, the NRP can face a variety of penalties including prison sentences and loss of parental rights. The few studies that have examined the Colombian child support system show that the vast majority of RPs do not receive child support and the amount of support is small. In 2008, only 28% of resident-mother families eligible for this transfer received any support. However, for those families in poverty receiving child support, this transfer represents approximately two-thirds of their income (Cuesta & Meyer, 2014). Like Colombia, the current U.S. child support policy scheme is a hybrid system that includes the courts, the Federal Office of Child Support Enforcement (OCSE), and the State Child Support Agencies. The OCSE was created in 1975, when the U.S. child support system started moving from a reactive scheme, where government intervention only occurred after problems have arisen (e.g., noncompliance of child support payments), to a pro-active system in which the government is involved to try to prevent problems from occurring (Huang & Han, 2012; Meyer, 2012). One distinctive characteristic of the U.S. system is that even though OCSE has a policymaking and funding role, states have discretion in the implementation of the policies directed by the federal government (Cuesta & Meyer, 2012). In the U.S., the procedure that takes place when parents are married is very similar to the one described for Colombia. Essentially, the child support arrangement is established as part of the divorce proceedings. If the parents were not married, the RP can voluntarily request the intervention of the child support system or make a private arrangement with the NRP. However, if the RP is receiving public assistance, the RP must cooperate with the agency in its efforts to establish a child support order (Cuesta & Meyer, 2012). Order amounts in divorce cases or nonmarital cases are set based on each state's guidelines. Guidelines in most states are based on a principle of income sharing, rather than need; the order amount should not reflect what a child needs but that children can expect a parent to share their income with them (Minow, 1998). This means that amounts ordered are generally higher for those NRPs with higher incomes, who tend to have ex-partners with higher incomes as well (Meyer, 1998). One aspect of the guidelines that has implications for our child poverty analyses is that in most states there are special procedures for low-income nonresident parents; generally these parents are assessed a lower percentage of their income in child support (Cancian, Meyer, & Han, 2011b). However, generally no adjustments are made when the NRP has new family responsibilities or the RP's living situation has changed (Meyer, 2012; Meyer & Skinner, 2016). If child support orders remain unchanged when the NRP has a new partner, a new biological child, or takes care of step-children, child support policy may reduce poverty among RP families at the expense of the economic well-being of children in the NRP family. Collection and distribution of payments is substantially more developed in the U.S. than it is in Colombia. In the U.S., payments are routinely withheld from wages and forwarded to the child support agency, which then transfers the payment to the RP. The child support agency monitors

3. Literature review The literature on the antipoverty effectiveness of child support has been primarily focused on understanding the extent to which NRP monetary contributions reduce income poverty among RP families in a few rich countries (e.g., Hakovirta, 2011; OECD, 2011) with relatively little work on developing countries. In this section we discuss findings from a number of studies that have looked at the antipoverty effectiveness of child support in developed countries and the only study that has examined this question in the context of a developing country. We use the terms resident-mother and nonresident father instead of the more generic form of resident-parent (RP) and nonresident-parent (NRP) that we generally use in other sections of the paper because the vast majority of studies that have looked at the antipoverty effectiveness of child support are focused on resident-mother families. Child support reduces poverty among resident-mother families in the U.S. (Renwick & Fox, 2016; for historical analysis see Bartfeld, 2000;Meyer & Hu, 1999; Nichols-Casebolt, 1986) and Colombia (Cuesta & Meyer, 2014) but the impact on overall poverty is relatively small in both countries. The percentage of resident-mother families brought out of poverty by child support ranges from 6% (Meyer & Hu, 1999; Nichols-Casebolt, 1986) to 11% (Bartfeld, 2000) in the U.S. The only estimate available for Colombia shows that in 2008, 9% of all residentmother families were lifted out of poverty by child support (Cuesta & Meyer, 2014). The small effect among all resident-mother families is partly explained by the low receipt rate observed in both countries, 28% in Colombia in 2008 (Cuesta & Meyer, 2014), and 32% in the U.S. in 2015 (Grall, 2018). The antipoverty effectiveness of child support is of course significantly larger when the analysis is limited to those families who are receiving child support. This effect has been estimated in 21% in the U.S. (Meyer & Hu, 1999) and 32% in Colombia (Cuesta & Meyer, 2014). In the U.S., the most recent estimate shows that child support reduced the poverty rate among children from 16% to 15%, bringing 6.6% of all children out of poverty (Renwick & Fox, 2016). Child support also reduces the poverty gap among those who remain poor after receiving child support. Among all resident-mother families, child support reduced the poverty gap by 5.3% in the U.S. (Hakovirta, 2011) and 9% in Colombia (Cuesta & Meyer, 2014). Again, the magnitude of the effect is larger among resident-mother families receiving child support, 21% in the U.S. (Hakovirta, 2011) and 39% in Colombia (Cuesta & Meyer, 2014). The magnitude of the effect of child support on the poverty gap versus the magnitude of the effect on the poverty rate suggests that child support is less effective at helping resident-mother families to escape poverty than it is at helping them to move closer to the poverty threshold (Cuesta & Meyer, 2014). Virtually all studies have measured the antipoverty effectiveness of child support using a poverty threshold specific to the country being studied. The effect itself is estimated with a relatively straightforward approach in which the authors calculate poverty rates pre-child support 145

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

Our study makes a number of contributions to the literature. First, to the best of our knowledge, this is the first study that looks at child support's effects among children living in NRP families in a developing country, and the first to compare these findings with a developed country. Because the processes that are driving family complexity are increasingly common in a wide range of nations (Child Trends, 2015), our findings have the potential to inform policymakers dealing with child support policy issues in both developed and developing countries. Second, this is the first study that provides separate estimates of child support's effects on child poverty for children in RP families and children in NRP families. This is a key difference because the few studies that have looked at this question present poverty rates for all children; while still informative, that analytic approach masks the extent to which child support pushes children living with NRPs into poverty.

and post-child support income, and then estimate the percentage point reduction in poverty and the percentage of resident-mother families who are brought out poverty after child support receipt. An analysis that follows this approach but uses a relative poverty measure (i.e., 50% of the household's median income) provides a cross-national perspective on the antipoverty effectiveness of child support from the perspective of children (Hakovirta, 2011; OECD, 2011). The study by Hakovirta (2011) finds that the lowest effects of child support on child poverty occurred in the U.S. (18% children brought out of poverty) and the highest impact was observed in Sweden (half of children brought out of poverty) and Norway (45% children lifted out of poverty). Few studies have provided national estimates of poverty for nonresident parents. An early estimate for the U.S. from Meyer (1998) is that child support payments increase the poverty rate of nonresident fathers from 23.6 to 24.6%, bringing about 1.3% of those who were not poor into poverty. Nonresident fathers who had children living with them have lower poverty rates, but 1.2–2.2% of them are pushed into poverty by paying support (Meyer, 1998). This study used the total amount of support paid to all children. Some information on the effect of paying child support on poverty is within a more recent study by Renwick and Fox (2016) focused on various measures of poverty. They also show that considering all child support paid by nonresident parents does not increase overall poverty rates by very much: by about 0.5% for all people and for children. Because of the focus of the study, it provides no other analyses of child support's role. Other studies have examined the effects of child support payments on the poverty rate of American nonresident fathers who were previously married (Bartfeld, 2000; Nichols-Casebolt, 1986), but these studies typically only consider payments to one family. The two studies focused on formerly married couples find that child support payments do increase the poverty rate among nonresident fathers (Bartfeld, 2000; Nichols-Casebolt, 1986), but the maximum increase observed across different scenarios is 2.7 additional percentage points in their poverty rate (Bartfeld, 2000). These studies also show that after support has been paid, the poverty rate among resident mothers is substantially higher than the poverty rate among nonresident fathers. Finally, a cross-national study conducted by the OECD suggests that child poverty does not change after child support payments have been considered. In fact, in all countries child poverty appears to be lower after advances on maintenance payments, child support receipt, and child support payments have been considered (OECD, 2011). The two studies of the effect of payments on nonresident parent poverty that use national data are limited: the Meyer (1998) study uses old data (from the 1980s) and Renwick and Fox (2016) provide only the simple percentage change and only use a U.S.-specific poverty measure that makes comparative analysis difficult. Moreover, the Meyer study focuses on nonresident parents as the unit of analysis, so does not provide explicit measures for children themselves. This is problematic because nonresident fathers living with one child are counted the same as nonresident fathers living with two or more children, which does not provide an accurate view of child poverty. The other studies that consider the impact of paid support on nonresident fathers and children also have limitations. First, the focus on formerly-married couples is likely causing an underestimation of the effects of child support payments on nonresident fathers' poverty because recent union instability has been mostly driven by disruption of cohabiting unions, which are generally more economically vulnerable than married couples (Smock, 2000). Moreover, these studies have been based on ex-couples, so do not consider the full extent of child support for those paying to more than one ex-partner. Second, none of these studies actually estimate the effect of child support payments by whether the child or family lives in a family receiving child support versus a family paying child support versus a family that is both paying and receiving. This distinction is important to understand the extent to which child support policy is reducing child poverty among RP families while increasing child poverty among NRP families.

4. Data, sample, methods, and measures 4.1. Data We use the Colombian Longitudinal Survey (ELCA for the Spanish acronym) and the U.S. Current Population Survey (CPS). ELCA is a household survey that began in 2010 and will follow up participants for at least 12 years. Households are interviewed every three years regarding aspects such as employment, income, property ownership, education, health, and family composition. The survey also includes information on amounts of child support paid and received. It is representative of the national socioeconomic strata one through four (3% of households have incomes too high to be included in the sampling frame for the survey). We use the 2013 wave, which is the most recent publicly available data. The CPS is the primary source of monthly labor statistics in the U.S. and has been conducted monthly since 1948. For the purpose of this study we use the Annual Social and Economic Supplement (CPS-ASEC) that is collected in March and has comprehensive information on work experience, income, non-cash benefits, and migration. The CPS-ASEC also includes information on family composition and amounts of child support paid to all ex-partners and received from all ex-partners. 4.2. Sample We limit our samples to families with own children (both biological and adoptive) under 18 years old. Our unit of analysis is the child. The ELCA has records for the families of 13,036 children, including 1781 who receive support and 419 who pay. The CPS-ASEC has 53,480, 5750, and 1046, respectively. 4.3. Measures 4.3.1. Child support received We use a continuous measure of the amount of child support received in the year prior to the survey (in 2012 for most cases in Colombia, but families interviewed between February and May of 2013 may have reported information on 2013; in 2012 for the U.S. sample). In the 2013 ELCA, child support variables are reported at the household level. In order to obtain a measure of child support received at the family level, we assign the household's amount of child support received to each family living in a household who received any child support. In the 2013 CPS-ASEC we are able to observe child support income received at the family level. In both countries, this amount includes cash child support, whether formal (with a legal order) and informal (without a legal order) child support. In Colombia, the amount also includes the value of in-kind transfers. 4.3.2. Labor income We use a continuous measure of the amount of labor income received in the child's family in the year prior to the survey (2012–2013 146

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

paid poverty gap is calculated with the post-child support paid income.

in Colombia, 2012 in the U.S.). In the U.S., we are able to observe annual labor income at the family level. In Colombia, the measure of labor income is for the household, rather than the family and it is based on a typical month, rather than annual. We multiply the monthly amount by 12 to approximate the annual amount of household's labor income; we assign this amount to each family living in the household.

4.4. Methods We first examine the extent to which families with children rely on earnings, child support, and other income sources, and whether child support payments represent a significant nondiscretionary expenditure for these families. To analyze the antipoverty effectiveness of child support, we use standard methods adapted to include expenses. Thus, we examine whether a child is poor based on his/her family income without child support paid and received. We then examine the children below the poverty threshold and calculate how many children are brought above poverty when we add in child support received. This analysis confirms prior work so is an important preliminary step. We then examine children above poverty and document how many are put into poverty based on the child support their family pays. Note that the ordering of the calculations could matter: in order to test the robustness of our findings, we estimate an additional set of analyses in which we change the order in which child support received and paid is included in this accounting exercise. Specifically, instead of starting with adding child support to the pre-child support income, we begin by subtracting child support paid from total family income and then calculate how many children are falling into poverty based on the support that is paid. For those who are below poverty based on post-child support paid income, we then add in child support received and determine how many are no longer poor. We make comparable calculations for the poverty gap.

4.3.3. Other income The U.S. has a measure of annual family income in the year prior to the survey (2012) that includes a wide variety of sources, including public assistance, rental income, other private transfers, etc. Our measure of “other income” is continuous and is the sum of all sources of family income other than labor income or child support received – or, equivalently, the amount of family income minus labor income and child support. We follow analogous procedures in Colombia, except that the measure of other income is for the household, rather than the family and it is based on a typical month, rather than annual. We multiply the monthly amount by 12 to approximate the annual amount of household income, and subtract annual labor income and child support received to get other income; we assign this amount to each family living in the household. 4.3.4. Child support paid We use a continuous measure of the amount of child support paid by the family in the year prior to the survey (2012–2013 in Colombia, 2012 in the U.S.). Again, in Colombia we assign the household's amount of child support paid to all families living in a household who paid any child support in the year prior to the survey. In the U.S., we assign the individual's amount of child support paid to this individual's family. In both countries this amount includes formal cash support (with a legal order) and informal cash support (without a legal order); in Colombia this also includes in-kind contributions.

4.5. Expectations We expect that child support payments will be pushing some children into poverty. Based on recent patterns in union dissolution and repartnering in both countries, we expect that more children will be living with those who pay child support and thus child support will have a larger effect in pushing some into poverty than in some of the older research. Based on the prior literature's findings that mostly mothers are receivers of child support and fathers are payers, and that women generally have had lower incomes than men, in both countries we expect that child poverty rates before child support will be higher than they are after child support is considered. (In other words, child support received will pull more children out of poverty than child support paid will push them into it.) However, recent declines in the economic status of men, especially young men (Smeeding, Garfinkel, & Mincy, 2011), may lead to different conclusions. In terms of whether we expect larger effects on poverty in the U.S. or Colombia, the literature review and our data point in different directions. Two factors suggest the effect of child support on poverty in the U.S. would be greater than in Colombia. First, if union instability and re-partnering are relatively more common in the U.S. than in Colombia, then American children may be more likely than Colombian children to grow up in a family in which there is an adult paying child support. Second, the child support system in the U.S. is more routinized than in Colombia, which we expect would lead to higher child support payments in the U.S., and thus a larger effect on poverty. However, there are three offsetting factors. First, the Colombian focus on providing for needs may lead to requiring more child support for those with low incomes, whose ex-partners may also have low incomes; this would lead to larger effects of child support on poverty in Colombia. Second, the Colombian child support data include in-kind contributions, while the U.S. data do not, which could lead to a larger effect on poverty in Colombia. Third, we use a poverty threshold that is relative to others within the country; because incomes in Colombia are lower than the U.S., the Colombian poverty threshold is lower, which may make it easier to cross. Finally, an additional important factor is the cost recovery policy in the U.S.: not all child support that is paid is received by resident parents, some is retained by the government. This should

4.3.5. Pre-child support and post-child support income Pre-child support income is a continuous measure that is a simple sum of other income and labor income. Note that we use a simple accounting framework that does not consider behavioral effects (e.g., in the absence of receiving child support, other income sources may have been different). This is the standard approach in this field (Bartfeld, 2000; Cuesta & Meyer, 2014; Hakovirta, 2011; Meyer, 1998; Meyer & Hu, 1999; Nichols-Casebolt, 1986). Post-child support received income is the straightforward addition of child support received; post-child support paid income merely subtracts child support paid. These amounts are adjusted for economics of scale using the OECD-modified scale as in Hakovirta (2011) and Skinner, Bradshaw, and Davidson (2007). The OECD-modified scale assigns a value of 1 to the household head, of 0.5 to each additional adult member and of 0.3 to each child (Hagenaars, de Vos, & Zaidi, 1994). All amounts are converted to purchasing power parities (PPP) U.S. dollars (World Bank PPP conversion factor for 2012: 1174.9 Colombian pesos = 1 U.S. dollar). 4.3.6. Poverty rates For our international comparisons, we mainly use a poverty threshold based on 50% of median annual family income of all families; in a sensitivity test we also consider a poverty threshold based on 60% of median annual family income of all families. The median annual income is adjusted for economics of scale using the OECD-modified scale as in Hakovirta (2011) and Skinner et al. (2007). 4.3.7. Poverty gaps This is measured as the amount of money that it would take to bring children whose family's annual income is below the poverty line up to the poverty threshold. The pre-child support poverty gap is calculated with the pre-child support income. The post-child support poverty gap is calculated with the post-child support income. The post-child support 147

148

84 6,357

270 15,121 3,011

100.0% 91.1% 36.3%

1.7% 511

0.5% 4,239

4.7% 851

100.0% 96.5% 39.1%

2.5% 88.0% 9.5%

0.5% 6,357

1.5% 82.2% 16.4%

26 511

1,178 5,166 749

95 851

1,702 12,554 1,820

Me an Amount (in PPP)

0.4% 511

16.6% 72.8% 10.6%

0.6% 851

10.6% 78.1% 11.3%

% Total Family Income

100.0% 109

8.2% 100.0% 27.4%

100.0% 224

17.9% 99.9% 56.6%

% >0

Me an Amount (in PPP)

% Total Family Income

1,053 109

41 6,901 271

2,001 224

221 21,982 7,554

14.6% 109

0.6% 95.7% 3.8%

6.7% 224

0.7% 73.9% 25.4%

1.2% 4,842

14.5% 63.7% 42.4%

2.1% 28,830

10.3% 91.1% 59.0%

33 4,842

496 7,691 2,590

130 28,830

609 71,327 4,757

Me an Amount

% Total Family Income

0.3% 4,842

4.6% 71.4% 24.0%

0.2% 28,830

0.8% 93.0% 6.2%

0.7% 721

100.0% 55.7% 51.5%

3.7% 2,988

100.0% 85.2% 59.5%

16 721

3,419 4,810 2,928

290 2,988

5,891 39,150 4,489

Me an Amount

% Total Family Income

0.1% 721

30.6% 43.1% 26.3%

0.6% 2,988

11.9% 79.1% 9.0%

100.0% 57

8.8% 71.9% 47.0%

100.0% 587

18.5% 96.4% 61.8%

2,877 57

95 9,536 2,833

6,307 587

998 71,634 5,013

Me an Amount (in PPP)

23.1% 57

0.8% 76.5% 22.8%

8.1% 587

1.3% 92.3% 6.5%

% Total Family Income

Source: Authors' calculations based on Encuesta Longitudinal Colombiana de la Universidad de los Andes – ELCA public data, and U.S. Current Population Survey, Annual Social and Economic Supplement.

Low income families with children Income source s Child support 14.9% 175 Labor income 95.1% 6,121 Other income 39.1% 662 Nondiscretionary expenses Child support 3.1% 32 Unweighted sample 4,239 4,239 (families)

All families with children Income source s Child support 15.8% Labor income 97.1% Other income 43.4% Nondiscretionary expenses Child support 4.2% Unweighted sample 6,357 (families)

% >0

% >0

% Total Family Income

% >0

Me an Amount (in PPP)

% >0

% >0

Payers

Recipients

All

Payers

All

Recipients

United State s

Colombia

Table 1 Child support and other income sources. Colombia and the United States. 2013. Weighted.

L. Cuesta, D.R. Meyer

Children and Youth Services Review 93 (2018) 143–153

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

while only 2.1% of families with children are paying child support. Among those receiving child support, this transfer represents 11.9% of the total family income. On average, the annual amount of support received for those who receive is substantial, $5891. Of the American families who receive child support, a small percentage are in families who are also paying child support (3.7%). Because so few pay, the annual amount they pay is small ($290) when averaged over all those who received; but if we calculate the average among recipients who are also payers, it is quite large, over $7000. In addition, about one in every five American families that pay child support are also receiving support from an absent parent (18.5%). The annual amount of support that payers transfer to other families is, on average, $6307, which is equivalent to about 8% of the annual family income. Like in Colombia, the large difference in the annual earnings of families receiving child support ($39,150) and the annual earnings of families paying child support ($71,643) suggest that children living in recipient families are significantly more disadvantaged than children living in payer families. The analyses for low-income families in the U.S. shows that the amount of child support received by recipients ($3419) is approximately equivalent to four-fifths of the annual family earnings ($4810). A significant difference observed between low-income families in Colombia and the U.S. is that the vast majority of Colombian child support recipients live in families with labor income (91.1%) while only 55.7% of American recipients live in families with any earnings. Our results also show that in the U.S., low-income payers are relatively more advantaged than low-income recipients. The average annual earnings of families paying support ($9536) was about twice the annual earnings of those receiving support ($4810). However, the percentage of earnings that goes to child support payments among low-income payers is almost one fourth (23.1%) of the annual family earnings in payer families. Overall, our results show that a higher proportion of families with children in Colombia are receiving support than in the U.S., and a higher proportion are paying support. However, amounts of child support in the U.S. are substantially higher; perhaps not surprising given how much higher earnings are. The results also show that in both countries child support receipts and payments are substantial when they exist. In both countries, recipients have lower earnings and income than payers. Finally, these results demonstrate that some children are in families that are both receiving and paying support, suggesting that prior analyses that implicitly assume that families are only receivers or only payers are incomplete. Table 2 shows the mechanical effect of child support on child poverty rates when the poverty threshold is set at 50% of the annual family median income. Pre-child support received and paid poverty rates among children in Colombia are 24.7%, higher for those in families receiving child support (33.3%) and lower for those in families paying support (13.5%). Our analyses for Colombia indicate that child support receipt has a modest effect on child poverty among all children (5.6%). Not surprisingly, the largest effect is observed among children

make the effect of child support receipts on poverty in the U.S. lower than in Colombia, though it will not differentially affect the relationship between child support paid and poverty. 5. Results Table 1 shows the frequency and mean of annual amounts of child support, earnings, and other family income, and the share of child support, earnings, and other family income in the total family income of families with children. We present results for all these families, and then separately for families receiving child support, and for families paying child support. The first panel shows findings for all families with children and the second shows results for low-income families with children, using each country's official definition of low-income based on post-child support received pre-child support paid income. These results show that a little less than one in every six Colombian families with children were receiving child support (15.8%) while only 4.2% families with children were paying child support. (Some households paying support do not have resident children, so are not part of our focus.) The share of child support in the total income of all families with children is small (1.5%). However, among those who receive this transfer, the amount received is about $1700, which represents 10.6% of their total family income. A small percentage of those receiving child support are also paying support to another household (4.7%); while this amount is small when averaged over all those who receive child support, it is not trivial when it is present, averaging over $2000. In addition, one in every six payer families are also receiving child support from an absent parent (17.9%). The annual amount of support that they receive is only $221 averaged over all those in paying families, but averages over $1200 when it is present. Child support payments take an important portion of the NRP family earnings. On average, the annual amount of support paid by families paying support is $2001. These payments also represent 6.7% of the total family income of payer families. Descriptive analyses show that, on average, families receiving child support are more disadvantaged than families paying child support. For example, the annual earnings of the families paying child support ($21,982) are almost twice the annual earnings of families receiving child support ($12,554). The second panel shows that 14.9% of low-income families with children are receiving support, about the same percentage as among all children (15.8%). Some of the low-income recipients of child support also pay child support (1.7%). Child support represents a larger percentage of total family income among low-income recipients (16.6%) than among all recipients (10.6%). These results also show that lowincome payers transfer a larger proportion of their income (14.6%) to other families than all payers do (6.7%). The average of annual paid support among low-income payers is $1053. The results for the U.S. show that approximately one in every ten American families with children receive child support income (10.3%)

Table 2 Mechanical effect of child support on child poverty rates, using poverty threshold set at 50% of median income. Colombia and the United States. Weighted. Colombia

Poverty rate pre-child support receipt and paid Poverty rate post-child support receipt Poverty rate post-child support receipt and paid Children brought out of poverty after child support receipt (pp) % of children brought out of poverty after child support receipt Children falling into poverty after child support paid (pp) % of children falling into poverty after child support paid Unweighted sample (children)

United States

All

Recipients

Payers

All

Recipients

Payers

24.7% 23.4% 23.7% 0.01 5.6% 0.00 0.4% 13,036

33.3% 25.4% 25.6% 0.08 23.7% 0.00 0.3% 1781

13.5% 13.0% 20.2% 0.01 3.8% 0.07 8.4% 419

32.1% 31.3% 31.4% 0.01 2.5% 0.00 0.1% 53,480

50.3% 42.8% 42.9% 0.08 14.9% 0.00 0.2% 5750

22.3% 22.1% 26.8% 0.00 0.8% 0.05 6.0% 1046

Source: Authors' calculations based on Encuesta Longitudinal Colombiana de la Universidad de los Andes – ELCA public data, and U.S. Current Population Survey, Annual Social and Economic Supplement. Note: pp.: percentage points. 149

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

recipients also paying child support, the effect on the poverty gap remains very high (29%) even after paid support is considered. However, our analyses also show that child support widens the child poverty gap among poor families paying child support. Child support payments increase the child poverty gap among payer families by 86%. In the U.S., the impact of child support income on recipient families (22%) and payer families (35%) is lower than it is in Colombia (29% and 86%, respectively). Some comparative studies use a poverty threshold of 60% of median income (Förster, 2005; Hakovirta, 2011). Our key results are presented in Table 4, with selected results from Tables 2 and 3 (which use a poverty threshold of 50% of median income) and the corresponding results if 60% of median income is the threshold. Of course with a higher threshold, more children are counted as poor, with generally larger changes in levels in Colombia than in the U.S. In both countries, the percentage of children brought out of poverty by child support is dampened when the higher threshold is used. Similarly, poverty gaps are higher with the higher threshold, but the poverty-reducing effects of child support are dampened somewhat. Overall, though, our main conclusions are robust to the alternative poverty line: child support recipients have higher poverty rates than child support payers; the antipoverty effect of child support receipt is larger than the povertyinducing effect of child support payments, and these findings hold in both countries.

receiving this transfer. Among pre-child support poor children who receive child support, 23.7% are brought out of poverty by child support alone. Given the low percentage of families that are paying support, the percentage of children falling into poverty after child support paid is very small (0.4%). Our results also show that child support payments have a negative impact on child poverty among payer families. Approximately 8% of Colombian children in payer families are falling into poverty after child support is transferred to other families. While the child poverty rate among payers increases (from 13.5% to 20.2%), the rate observed among recipients remains higher even after child support is considered (25.6%). In the U.S., child poverty rates before child support is paid or received are 32.1%, surprisingly higher than Colombia, but consistent with the high poverty rate among children in the U.S. compared to other countries when the poverty threshold is based on median income (OECD, 2017). Similar to Colombia, the pre-child support rate is higher for recipients (50.3%) and lower for payers (22.3%). Findings for the U.S. are similar to those observed for Colombia although the size of the effects is generally smaller in the U.S. than in Colombia. Child support reduces income poverty among American children receiving child support (almost 15% of children in this group are brought out of poverty by child support alone) but 6% of children living in payer families are falling into poverty after child support is transferred to other families. Like in Colombia, poverty rates of recipients in the U.S. (42.9%) remain higher than poverty rates of payers (26.8%), even after child support income is received. However, the difference in child poverty rates between recipients and payers are significantly larger in the U.S. (42.9% vs. 26.8%) than in Colombia (25.6% vs. 20.2%). Our findings are robust to a sensitivity test in which we switch the order in which child support received and paid are added and subtracted. For example, in our base results (in which child support received is considered before child support paid, and which allow us to compare our results to the prior literature), 5.6% of poor children in Colombia and 2.5% of poor children in the U.S. are brought out of poverty by child support received; switching the order leads to comparable numbers of 5.5% and 2.5%. Similarly, in our base results, 0.4% of non-poor children in Colombia and 0.1% of non-poor children in the U.S. become poor when payments are considered; switching the order leads to comparable numbers of 0.4% and 0.2% (Results available from authors upon request). Table 3 shows the effects of child support on the child poverty gap when the poverty threshold is set at 50% of the annual family median income. Our results indicate that child support is helping some poor families by closing the gap between their income and the poverty threshold. Looking at all children, child support receipt lowers the poverty gap by 6% in Colombia, and in the U.S. by 3%. Accounting for payments does increase the gap slightly, but the decline from receipts outweighs the increase from payments. The largest impact is observed among recipients in both countries. In Colombia, child support income reduces the child poverty gap by 30%. Given the low percentage of

6. Conclusions Our findings should be interpreted in light of the following limitations. First, we do not account for the effects of child support on RP and NRP behavior that may change family income. For instance, in the absence of child support the RP may find another partner to compensate for the lack of NRP contributions. This situation would make our pre-child support income measure an inaccurate proxy for what family income would be in the absence of child support. While some research has examined the potential behavioral effects of child support received (Huang and Han (2012) and child support paid (Garfinkel, McLanahan, Meyer, & Seltzer, 1998b), accounting for them is beyond the scope of this article. Another limitation of our study is that we are using a household-reported measure of child support in Colombia instead of a family-level measure; this is a potential issue because 23% of the families in our Colombian sample are living in multiple-family households. If members of households do not pool their resources, then we are attributing child support receipt to some children who are not benefiting from it and counting it as an expense for some children whose financial situation may not be affected by the payment. If this is the case, in addition to overestimating the number of children receiving or paying child support, we are underestimating the amount of support received and paid per child. In a robustness test, we found that our conclusions were not sensitive to whether children in extended-families were included or excluded. Third, there are some differences between the two data sources that make comparisons less accurate: the

Table 3 Mechanical effect of child support on child poverty gaps, using poverty threshold set at 50% of median income. Colombia and the United States. Weighted. Colombia (in PPP)

United States

(Millions of dollars)

All

Recipients

Payers

All

Recipients

Payers

Poverty gap pre-child support receipt and paid Poverty gap post-child support receipt Mechanical effect of child support received on poverty gap Poverty gap post child support receipt and paid Mechanical effect of child support received and paid on poverty gap Unweighted sample (children)

2519 2357 −6% 2400 −5% 13,036

540 378 −30% 383 −29% 1781

47 46 −3% 88 86% 419

163,000 158,000 −3% 158,000 −3% 53,480

24,500 19,100 −22% 19,200 −22% 5750

1790 1740 −3% 2420 35% 1046

Source: Authors' calculations based on Encuesta Longitudinal Colombiana de la Universidad de los Andes – ELCA public data, and U.S. Current Population Survey, Annual Social and Economic Supplement. 150

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

Table 4 Mechanical effect of child support on child poverty and child poverty gaps: Sensitivity test to alternate poverty threshold. Colombia and the United States. Weighted. Colombia (in PPP)

Poverty Rates: Base (Threshold at 50% of Median) Poverty rate pre-child support receipt and paid Poverty rate post-child support receipt and paid % of children brought out of poverty after child support receipt % of children falling into poverty after child support paid Poverty Rates: Sensitivity Test (Threshold at 60% of Median) Poverty rate pre-child support receipt and paid Poverty rate post-child support receipt and paid % of children brought out of poverty after child support receipt % of children falling into poverty after child support paid Poverty Gaps: Base (Threshold at 50% of Median) Poverty gap pre-child support receipt and paid Mechanical effect of child support received on poverty gap Mechanical effect of child support received and paid on poverty gap Poverty Gaps: Sensitivity Test (Threshold at 60% of Median) Poverty gap pre-child support receipt and paid Mechanical effect of child support received on poverty gap Mechanical effect of child support received and paid on poverty gap Unweighted sample (children)

United States

All

Recipients

Payers

All

Recipients

Payers

24.7% 23.7% 5.6% 0.4%

33.3% 25.6% 23.7% 0.3%

13.5% 20.2% 3.8% 8.4%

32.1% 31.4% 2.5% 0.1%

50.3% 42.9% 14.9% 0.2%

22.3% 26.8% 0.8% 6.0%

34.6% 32.8% 5.2% 0.1%

41.2% 31.2% 24.8% 0.3%

22.0% 22.7% 2.0% 1.5%

38.0% 37.5% 1.8% 0.2%

56.6% 50.4% 11.2% 0.2%

26.6% 33.4% 0.3% 9.4%

2519 −6% −5%

540 −30% −29%

47 −3% 86%

163,000 −3% −3%

24,500 −22% −22%

1790 −3% 35%

4409 −5% −3% 13,036

868 −24% −23% 1781

106 −2% 52% 419

235,000 −3% −3% 53,480

34,900 −20% −19% 5750

2850 −2% 33% 1046

Source: Authors' calculations based on Encuesta Longitudinal Colombiana de la Universidad de los Andes – ELCA public data, and U.S. Current Population Survey, Annual Social and Economic Supplement.

child support recipients (based on pre-child support income) receives more support than those in other quintiles. In the U.S. we see the opposite pattern. Second, the relatively higher antipoverty effectiveness of child support in Colombia may be partly explained by the fact that the Colombian social welfare system is not founded on a cost-recovery scheme. As a result, low-income families eligible for public assistance and child support can receive both benefits in full, increasing the antipoverty effectiveness of child support in Colombia. In contrast, in the U.S. a portion or all child support paid on behalf of children receiving public assistance is used to offset taxpayer expenditures rather than to decrease poverty (Skinner et al., 2017).2 A third possibility is that because median incomes are lower in Colombia, the poverty threshold is lower; this means that a lower amount of child support might be effective at bringing a family above the poverty line. Indeed, sensitivity analyses using a poverty threshold of $5.50/person/day (as recommended by the World Bank for upper middle-income economies), shows some support for this explanation: effects in the U.S. are somewhat higher than in Colombia when an absolute poverty threshold is used. Our analyses on the effects of child support income confirm and expand upon prior work. Child support reduces child poverty, especially among children living in recipient families. Our findings suggest that the antipoverty effectiveness of child support among RP families has increased in both countries since the most recent analyses. Hakovirta (2011) had estimated that child support reduced child poverty in the U.S. by 6% in 2004 while Cuesta and Meyer (2014) found that child support reduced child poverty in Colombia by 12.4% in 2008. In this study we find that 11% to 15% American children and 24% to 25% of Colombian children in RP families were lifted out of poverty by child support alone in 2013. Increases in the effect of child support over time may be an artifact of different data sources or slightly different methods of analysis; data from the U.S. does not show large improvements in most child support outcomes over time (Grall, 2018). Another

Colombian data on labor income and other income cover only a typical month; the U.S. measure is annual. The Colombian data on child support include in-kind contributions; the U.S. measures do not. This does seem to be a serious limitation in that in-kind transfers are infrequent in Colombia.1 Finally, the Colombian data do not include the 3% of the population with the highest incomes; the U.S. data are nationally representative. Despite these limitations, this is one of the first studies to present information on the relationship between child support and child poverty that considers both those receiving and those paying support. Moreover, it is the first study that we know that looks at child support's effects among children living in NRP families in a developing country, and the first to compare these findings with a developed country. We expected to find that child support was pushing some children into poverty in both countries. We find this, but the effects are very small when considering all children, primarily because so few children are in families paying support. Even among those who do pay, only 8% of non-poor paying families in Colombia fall into poverty as a result of their payments; the comparable figure for the U.S. is 6%. Child support payments are increasing the poverty gap among those already poor and paying, by 86% in Colombia and 35% in the U.S. We expected that child support would pull more children out of poverty than it pushed children into poverty. This was confirmed: considering both child support paid and received, the relative poverty rates decreased by 5.6% in Colombia and 2.5% in the U.S. Moreover, the poverty gap declined in both countries after child support was considered. Based on the prior literature and policy context, we did not have strong expectations for which country would have larger effects, as several factors point in each direction. We generally find larger effects in Colombia, even though their child support system is more voluntary and reactive and incomes are lower. Why might this be? One factor may be that the Colombian system sets child support orders more based on what children need, so orders for low-income children may be higher. Some data support this: in Colombia, the lowest-income quintile of

2 This is unlikely to be related to the direct pass-through and disregard of child support in TANF, because so few families receive TANF (7% in these data). A larger issue may be the indirect cost recovery, in which child support counts as income in other means-tested programs so that recipients receive less in these other benefits. Skinner and colleagues (2017) provide a fuller examination of this issue.

1 The Colombian data provides information on the type of support received, but not the amount of each. More than three-quarter of those receiving child support receive only cash, 17% receive both cash and in-kind, and 7% receive only in-kind.

151

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

potential explanation is that in the U.S. as the public safety net has shrunk, private transfers like child support are becoming more important in lifting families out of poverty. Our findings have some implications for policy in these and other countries. First, as public programs that provide economic support to economically vulnerable families with children become more difficult to sustain (either politically or economically), it is likely that countries will look to other sources of income, including nonresident parents. These findings suggest that child support should be considered a vital potential income source: we find that it has important effects on poverty among children in families that receive it, and that while it does increase poverty among some children in families that are paying support, these numbers are substantially smaller. If policy emphasizes child support as a means of decreasing poverty among children, these results show that some care will be needed in determining policies that best ensure that children in new families are not impoverished by payments made to original families. The appropriate amount of support expected when there are new children involved is a thorny policy issue that countries and states approach differently (see Meyer et al., 2011; Meyer & Skinner, 2016). More research on policy approaches and their effects would be useful. Despite the potential importance of child support, and in the face of the difficulties in constructing policy for complicated family relationships, we want to highlight that in neither of these countries is child support on its own the main antipoverty tool. In both counties, public programs and earnings from the labor market are both more important in the number of families that receive them and in the level of income they provide. Moreover, as we show here, child support has a negative side that is substantially less relevant for other sources of income: child support payments are associated with some increase in poverty among families that pay it. Government transfers do not have this direct effect on decreasing resources among the other parent, since taxes are spread over a larger base. Thus, those concerned about child poverty will need to have a multi-faceted strategy for improving the economic well-being of children who live with only one parent and other economically vulnerable children.

4,8,12). DANE (2016). Pobreza monetaria y multidimensional en Colombia 2015. (Boletín Técnico, Bogotá, March 2nd 2016). Retrieved from Departamento Nacional de Estadistica—DANE website http://dane.gov.co/files/investigaciones/condiciones_ vida/pobreza/bol_pobreza_16.pdf. Devaus, D., Gray, M., Qu, L., & Stanton, D. (2015). The economic consequences of divorce in six OECD countries (research report no. 31). Melbourne: Australian Institute of Family Studies. Esteve, A., Lesthaeghe, R., & López-Gay, A. (2012). The Latin American cohabitation boom, 1970–2007. Population and Development Review, 38, 55–81. Fiszbein, A., & Schady, N. (2009). Conditional cash transfers. A World Bank Policy Research Report. Retrieved from the World Bank website http://siteresources. worldbank.org/INTCCT/Resources/5757608-1234228266004/PRR-CCT_web_ noembargo.pdf. Förster, M. (2005). The European social space revisited: Comparing poverty in the enlarged European Union. Journal of Comparative Policy Analysis, 7(1), 29–48. Garfinkel, I., McLanahan, S. S., & Hanson, T. L. (1998a). A patchwork portrait of nonresident fathers. In I. Garfinkel, S. S. McLanahan, D. R. Meyer, & J. A. Seltzer (Eds.). Fathers under fire: The revolution in child support enforcement (pp. 31–60). New York: Russell Sage. Garfinkel, I., McLanahan, S. S., Meyer, D. R., & Seltzer, J. A. (1998b). Fathers under Fire: The Revolution in Child Support Enforcement. New York: Russell Sage31–60. Grall, T. (2018). Custodial mothers and fathers and their child support: 2015. Current Populations Reports P60–262. Retrieved from the U.S. Census Bureau website https:// www.census.gov/library/publications/2018/demo/p60-262.html. Hagenaars, A., de Vos, K., & Zaidi, M. A. (1994). Poverty statistics in the late 1980s: Research based on micro-data. Luxembourg: Office for Official Publications of the European Communities. Hakovirta, M. (2011). Child maintenance and child poverty: A comparative analysis. Journal of Poverty and Social Justice, 19(3), 249–262. Huang, C., & Han, K. (2012). Child support enforcement in the United States: Has policy made a difference? Children and Youth Services Review, 34, 622–627. Jacobsen, L., Mather, M., & Dupuis, G. (2012). Household change in the United States. Population Bulletin, (1), 67. Retrieved from the Population Reference Bureau website http://www.prb.org/pdf12/us-household-change-2012.pdf. Kroeger, R., & Smock, P. (2014). Cohabitation: Recent research and implications. In J. Scott, J. Treas, & M. Richards (Eds.). The Wiley-Blackwell Companion to the Sociology of Families (pp. 217–235). Wiley-Blackwell. Manning, W. D., & Smock, P. J. (2000). “Swapping” families: Serial parenting and economic support for children. Journal of Marriage and the Family, 62(1), 111–122. Martin, J. A., Hamilton, B. E., Osterman, M. J. K., Driscoll, A. K., & Mathews, M. S. (2017). Births: Final data for 2015. National Vital Statistics Reports. vol. 66. Retrieved from the Centers for Disease Control and Prevention website https://www.cdc.gov/nchs/data/ nvsr/nvsr66/nvsr66_01.pdf. Meyer, D. R. (1998). The effect of child support on the economic status of nonresident fathers. In I. Garfinkel, S. S. McLanahan, D. R. Meyer, & J. A. Seltzer (Eds.). Fathers under fire: The revolution in child support enforcement (pp. 67–93). New York: Russell Sage. Meyer, D. R. (2012). Child maintenance policies in the United States. European Journal of Social Security, 14(4), 252–266. Meyer, D. R., & Hu, M. (1999). A note on the antipoverty effectiveness of child support among mother-only families. Journal of Human Resources, 34(1), 225–234. Meyer, D. R., & Skinner, C. (2016). Privileging biological or residential relationships in separated families: Child maintenance policies approaches in 12 countries. Families, Relationships, and Societies, 5(1), 79–95. Meyer, D. R., Skinner, C., & Davidson, J. (2011). Complex families and equality in child support obligations: A comparative policy analysis. Children and Youth Services Review, 33(10), 1804–1812. Minow, M. (1998). How should we think about child support obligations? In I. Garfinkel, S. S. McLanahan, D. R. Meyer, & J. A. Seltzer (Eds.). Fathers under fire: The revolution in child support enforcement (pp. 302–330). New York: Russell Sage. National Conference of State Legislatures (2017). Child support pass-through and disregard policies for public assistance recipients. http://www.ncsl.org/research/ human-services/state-policy-pass-through-disregard-child-support.aspx#50-State %20Chart. Nichols-Casebolt, A. (1986). The economic impact of child support reform on the poverty status of custodial and noncustodial families. Journal of Marriage and Family, 48(4), 875–880. OECD (2010). Gender Brief. Retrieved from the OECD website http://www.oecd.org/els/ family/44720649.pdf. OECD (2011). Doing better for families. Retrieved from the OECD website https://doi. org/10.1787/9789264098732-en. OECD. (2017). Poverty rate. doi: https://doi.org/10.1787/0fe1315d-en (Accessed on 02 August 2017). Pirog, M., & Ziol-Guest, K. (2006). Child support enforcement: Programs and policies, impacts and questions. Journal of Policy Analysis and Management, 25(4), 943–990. Profamilia (2016). National Demographic Survey 2015. Retrieved from Profamilia websitehttp://profamilia.org.co/docs/Libro%20RESUMEN%20EJECUTIVO.pdf. Renwick, T., & Fox, L. (2016). The supplemental poverty measure. Current Population Reports P60, No 258. Suitland, MD: U. S. Census Bureau. Semega, J. L., Fontenot, K. R., & Kollar, M. A. (2017). Income and poverty in the United States: 2016 (current population reports). Retrieved from the U.S. Census Bureau website https://www.census.gov/content/dam/Census/library/publications/2017/ demo/P60-259.pdf. Skinner, C., & Davidson, J. (2009). Recent trends in child maintenance schemes in 14 countries. International Journal of Law, Policy, and the Family, 23(1), 25–52.

Declarations of interest None. References Andersson, G., Thomson, E., & Duntava, A. (2017). Life-table representations of family dynamics in the 21st century. Demographic Research, 37(35), 1081–1230. Barrientos, J., Ramírez, S., & Tabares, E. (2015). On the poverty dynamics in Colombia, 2002-2012. Revista de Economía del Caribe, (15), 11–46. Bartfeld, J. (2000). Child support and the post-divorce economic well-being of mothers, fathers, and children. Demography, 37(2), 203–213. Bianchi, S. M., Subaiya, L., & Kahn, J. R. (1999). The gender gap in the economic wellbeing of nonresident fathers and custodial mothers. Demography, 36(2), 195–203. Buehler, C., & Gerard, J. M. (1995). Divorce law in the United States: A focus on child custody. Family Relations, 44, 439–458. Cancian, M., Meyer, D. R., & Cook, S. (2011a). The evolution of family complexity from the perspective of nonmarital children. Demography, 48, 957–982. Cancian, M., Meyer, D. R., & Han, E. (2011b). Child support: Responsible fatherhood and the quid pro quo. The Annals of the American Academy of Political and Social Science, 635, 140–162. Carlson, M. J., & Furstenberg, F. F. (2006). The prevalence and correlates of multipartnered fertility among urban U.S. parents. Journal of Marriage and Family, 68, 718–732. Child Trends (2015). Mapping family change and child well-being outcomes. Retrieved from Child Trends website https://childtrends-ciw49tixgw5lbab.stackpathdns.com/ wp-content/uploads/2015/09/2015-39WorldFamilyMap2015.pdf. Cuesta, L., & Meyer, D. R. (2012). Child support receipt: Does context matter? A comparative analysis of Colombia and the United States. Children and Youth Services Review, 34, 1876–1883. Cuesta, L., & Meyer, D. R. (2014). The role of child support in the economic wellbeing of custodial-mother families in less developed countries: The case of Colombia. International Journal of Law, Policy, and the Family, 28, 60–76. Cuesta, L., Rios-Salas, V., & Meyer, D. R. (2017). The impact of family change on income poverty in Columbia and Peru. Journal of Comparative Family Studies, 48(1), 67–96 (3-

152

Children and Youth Services Review 93 (2018) 143–153

L. Cuesta, D.R. Meyer

families, poverty, and policy. Annals of Political and Social Science, 635(1), 6–21. Smock, P. (2000). Cohabitation in the United States: An appraisal of research themes, findings, and implications. Annual Review of Sociology, 26, 1–20. Sorensen, E. (2016). The child support program is a good investment. Washington, DC: Office of Child Support Enforcement. Thomson, E., Lappegard, T., Carlson, M., Evans, A., & Gray, E. (2014). Childbearing across partnerships in Australia, the United States, Norway, and Sweeden. Demography, 51, 485–508. UNDP (2016). Regional human development report for Latin America and the Caribbean. Multidimensional progress: well-being beyond income. Retrieved from United Nations development program—UNDP website http://www.latinamerica.undp.org/content/ rblac/en/home/library/human_development/informe-regional-sobre-desarrollohumano-para-america-latina-y-e/.

Skinner, C., & Meyer, D. R. (2006). After all the policy reforms, is child support actually helping low-income mothers? Benefits. the Journal of Poverty and Social Justice, 14(3), 209–222. Skinner, C., Bradshaw, J., & Davidson, J. (2007). Child support policy: An international perspective (DWP Research Report No 405). Retrieved from the University of York website https://www.york.ac.uk/inst/spru/pubs/pdf/rrep405.pdf. Skinner, C., Cook, K., & Sinclair, S. (2017a). The potential of child support to reduce lone mother poverty: Comparing population survey data in Australia and the UK. Journal of Poverty and Social Justice, 25(1), 79–94. Skinner, C., Meyer, D. R., Cook, K., & Fletcher, M. (2017b). Child maintenance and social security interactions: The poverty reduction effects in model lone parent families across four countries. Journal of Social Policy, 46(3), 495–516. Smeeding, T. M., Garfinkel, I., & Mincy, R. B. (2011). Young disadvantaged men: Fathers,

153