Nonresident Parent Wealth Among Children Open Access

Abstract

Wealth matters for children's well-being. However, most research has considered only the wealth held in children's resident households, even though more than one quarter of U.S. children have a parent living elsewhere. We provide the first national estimates of nonresident parent wealth among children who ever lived with both parents in a shared household. We use the Panel Study of Income Dynamics (1984–2021)—whose survey design enables us to produce unbiased estimates with respect to factors that condition access to nonresident parents—to describe the characteristics of nonresident parent wealth, children's potential access to it, and its contribution to educational attainment. We find that nonresident parent wealth varies substantially and that the rank–rank correlation between nonresident and resident parent wealth is roughly .3. Nonresident parent wealth is positively predictive of more frequent parent–child contact, more consistent and higher value child support payments, and children's bachelor's degree attainment, suggesting that it is related to multiple domains of parenting and child well-being. The findings demonstrate that accounting for the spread of resources across households, but within a family system, is critical to understanding broader patterns of inequality in American family life.

Introduction

Household wealth—the combined wealth of all members living in a household—is an important stratification mechanism among children. Net of household income and parental education, household wealth matters for children's healthy development, including behavior (Berger and Houle 2016, 2019; Gibson-Davis et al. 2022; Miller et al. 2021), body mass index (Boen et al. 2021), academic performance (Gibson-Davis et al. 2022; Miller et al. 2021; Shanks 2007), and college attendance and completion (Pfeffer 2018). The pathways through which household wealth benefits children include objective and subjective financial well-being, reduced parental stress, and greater expectations of children (Gibson-Davis and Hill 2021). On a broader scale, wealth inequality among children is increasing, and households with children hold less wealth than other household types (Gibson-Davis and Percheski 2018; Pfeffer and Schoeni 2016). Understanding children's access to wealth is integral to understanding determinants of children's well-being and depicting broader patterns of inequality.

Research on wealth inequality and its consequences has focused either on (1) the wealth in children's households without considering nonresident parent households or (2) the wealth of biological parents without considering whether parents and children live together. These approaches overlook nonresident parent household wealth as a distinctive and potentially critical economic resource in a child's family system. More than one quarter of children have a parent living outside their residential household (Grall 2020) whose wealth might contribute to their development, well-being, and life chances.

We use the Panel Study of Income Dynamics (PSID) to provide the first estimates of nonresident parents’ wealth from the perspective of minor children. The PSID is the only national U.S. survey that would permit this kind of analysis. Our sample of nonresident parents includes biological parents of minor children who lived with the focal children previously but no longer do. We describe the magnitude and distribution of nonresident parent household wealth, its child and parent sociodemographic correlates, and its similarities and differences relative to resident parent household wealth. We then consider whether nonresident parent wealth is accessible to children and consequential to their early status attainment. Our study makes three contributions. First, we illustrate a more thorough picture of the economic resources potentially available to children, recognizing that families and their resources are not necessarily bound by households (Seltzer 2019). We also find moderately strong patterns of wealth concordance between resident and nonresident parents, with implications for research on wealth sorting into marriage and wealth redistribution across households following a union dissolution. Second, we demonstrate that nonresident parents’ wealth is likely accessible to children, as reflected in its positive association with the frequency of parent–child contact and child support payments. Third, we provide evidence that nonresident parent wealth is consequential for children through its positive association with the probability of children's bachelor's degree attainment, contributing to research on social stratification and the intergenerational transmission of (dis)advantage outside of two-parent, shared-household families (Tach 2015).

Background

Children experience high levels of wealth inequality (Gibson-Davis and Percheski 2018), particularly in the United States (Pfeffer and Waitkus 2021). One driver of wealth inequality among children is family structure. Households headed by married parents have substantially more wealth than single-parent households (Gibson-Davis and Percheski 2018; Percheski and Gibson-Davis 2022). In 2019, married households with children had a median net worth of $70,839, whereas unmarried households with children had a median net worth of just $975 (Gibson-Davis and Hill 2021:8). Additionally, wealth inequalities by family structure grow across the wealth distribution; inequalities are greater at the median than at the 25th percentile, and few never-married mothers’ households occupy the 75th percentile (Percheski and Gibson-Davis 2022). Wealth is also predictive of divorce; couples with low wealth holdings (especially close to 0) are more likely to divorce than couples with high wealth holdings (Killewald et al. 2023).

However, our understanding of wealth inequality among children is limited by using measures of resident householdwealth, which at least implicitly assumes that the resources important for children exist only in their household. This assumption might be incorrect, especially for the 26.5% of children who have one nonresident parent (Grall 2020). Separate strands of research on grandparent wealth and nonresident parent economic support indicate that family economic resources outside the child's household shape children's well-being and life chances. Grandparent wealth is positively associated with grades earned in high school (Hällsten and Pfeffer 2017) and wealth in adulthood (Pfeffer and Killewald 2018). Nonresident parents’ formal and informal income support are each positively associated with children's standardized test scores (Choi and Pyun 2014; Nepomnyaschy et al. 2012). These findings suggest that nonresident parent wealth, too, might influence children's well-being. An exclusive focus on household wealth likely underestimates, for a sizable minority of children, the total wealth in their family system.

Predictors of Nonresident and Resident Parent Wealth

Prior research has documented sociodemographic and family demographic inequalities in children's household wealth, but how these associations might translate to nonresident parents’ wealth is not immediately clear. On the one hand, known correlates of wealth suggest that differences in household composition and household processes might yield different patterns of wealth accumulation in resident versus nonresident parent households. Resident households are more likely to be female-headed, and women, on average, have less wealth than men (Kapelle and Baxter 2021). Nonresident parents might also be more likely to repartner—because men are more likely to repartner than women and because nonresident parents are more likely to repartner than resident parents (Di Nallo 2019; Goldscheider and Sassler 2006)—and repartnering contributes to wealth accumulation relative to remaining unpartnered (Frech et al. 2017). The advantages of marriage for wealth accumulation suggest that parental marital status at the time of the child's birth and current union status each might shape wealth holding for (non)resident parents (Addo and Lichter 2013; Kapelle and Baxter 2021). New partner fertility might reflect higher wealth levels if couples feel financially ready to have a child (Su and Addo 2018). Conversely, it might lead to lower wealth levels, especially among parents in the bottom half of the wealth distribution, given that children are expensive (Maroto 2018). Finally, the time since a child last lived with their nonresident parent might influence wealth holding because time allows for wealth accumulation.

On the other hand, we might expect wealth holding among resident and nonresident parents to be similar. Couples often match on their parents’ wealth (Charles et al. 2013)—especially adult children of wealthy parents (Wagner et al. 2020)—suggesting wealth homophily in union formation. Additionally, among parents who were married, the relatively equal division of assets during divorce (in most U.S. states; Hersch and Shinall 2020) further increases the likelihood of wealth similarity between resident and nonresident parents after union dissolution.

Our sample of resident and nonresident parents of shared children also allows us to provide the first estimates of postseparation wealth concordance between resident and nonresident parents of shared children. These estimates reveal the stickiness of wealth between parents who split up and suggest how much we can infer about children's nonresident parent wealth from their resident household wealth information, the only information usually available in large-scale datasets. Although prior research has not examined wealth correlations between separated or divorced parents, it has established wealth correlations between other family members. For example, Charles and colleagues (2013) identified the parental wealth correlation of married spouses to be roughly .40. Pfeffer and Killewald (2018) identified the parent–child wealth correlation to be .39 and the grandparent–grandchild wealth correlation to be .23.

Children's Access to Nonresident Parent Wealth

Nonresident parents’ wealth might matter for children only to the extent that children have access to it. This access can occur partly through financial transfers from nonresident parents directly to their children living elsewhere or on their behalf. Formal child support payments from noncustodial parents are one such mechanism. Roughly one half of custodial parents have a formal childcare arrangement, and less than half of that group receives full payments consistently (Grall 2020). Nonresident parents’ wealth might enable more regular and complete payments. Additionally, because child support orders are determined by income rather than wealth (Garfinkel et al. 1994), nonresident parents might use savings or other types of liquid wealth to maintain regular payments when experiencing income shocks.

Nonresident parents might also use wealth to invest in children according to intensive parenting norms, whereby parents invest time and money in developmentally appropriate and accepted-as-best-practice parenting activities to improve their children's life chances (Hays 1996; Sayer et al. 2004; Schneider et al. 2018). They might purchase goods, services, and experiences that are productive for children's development, including private education, tutoring, and extracurricular activities (Yeung and Conley 2008), either occasionally or routinely as informal child support. Wealth, in particular, might allow parents to purchase these investments in times of economic distress (Gibson-Davis et al. 2022). Wealth is also likely a driver of class differences in parenting—rather than solely a means of parental investment—given that wealth is a symbol of class status that signals socioeconomic achievement to other families and to children's peers (Conley 1999; Gibson-Davis and Hill 2021; Killewald et al. 2017). To date, much of the work describing intensive parenting has focused on coresident parents and children, but nonresident parents might be similarly incentivized by parenting norms to use wealth to their children's advantage.

Children might also benefit indirectly from nonresident parents’ wealth if it diminishes parents’ perceived stress and economic uncertainty (Conger et al. 2010; Gibson-Davis and Hill 2021; Gibson-Davis et al. 2022). Crucial to the viability of these less tangible, nonpecuniary pathways is the expectation that wealth need not be converted to cash or directly transferred to benefit children. Indeed, the parental stress perspective is an oft-cited mechanism for how household wealth benefits children through its effect on parenting (Gibson-Davis et al. 2023), and we expect that nonresident parents’ stress and parenting behavior, too, would be influenced by their own household wealth.

Additionally, children might experience nonresident parents’ wealth as access to stability through housing apart from the child's residential home.¹ Stability and quality of the home environment are positively linked with children's healthy development (Evans 2004). A nonresident parent home might also provide educational opportunities to children. For example, children might use a nonresident parent's home address to attend a different (and potentially higher quality) school than in their resident parent's catchment area.

Here, we consider two mechanisms through which children could be exposed to nonresident parents’ wealth: frequency of parent–child contact and nonresident parents’ regular payment of formal and informal child support.

Nonresident Parent Wealth and Children's Educational Attainment

Nonresident parents’ wealth might be consequential for children only if it is independently related to children's well-being and attainment, net of resident parent wealth. Although we know little about the later-life correlates of nonresident parent wealth, the association between childhood wealth and educational outcomes is well-established. Children's residential household wealth is positively predictive of their grades (Hällsten and Pfeffer 2017), test scores (Gibson-Davis et al. 2022; Miller et al. 2021; Orr 2003; Shanks 2007), and degree attainment (Conley 2001; Pfeffer 2018). Here, we focus on whether nonresident parent wealth is related to children's educational attainment in young adulthood.

Nonresident parent wealth might influence educational attainment for at least two reasons. First and most directly, nonresident parent wealth might facilitate children's educational attainment through economic support to the child (i.e., a monetary transfer). Parental transfers for education are positively associated with college graduation (Hamilton 2013). However, transfer amounts vary depending on children's coresident family structure. Children with a nonresident parent receive less financial support for postsecondary education from resident parent(s) than children living with two biological parents (Amorim and Deming 2022; Fomby and Kravitz-Wirtz 2019; Henretta et al. 2012), and these contributions are contingent on family composition. For example, unpartnered parents transfer slightly less money to children than repartnered parents (Fomby and Kravitz-Wirtz 2019), and in repartnered families, stepchildren receive less support than biological children (Henretta et al. 2014). Transfers from nonresident parents might at least partially compensate for these shortfalls relative to children in two-parent households.

Similarly, nonresident parents’ wealth might compensate for low wealth in children's residential households to finance educational attainment. The compensation perspective posits that extended family members can provide support when the immediate family lacks resources (Jæger 2012). Children living in families with lower socioeconomic status (SES) have better educational outcomes when extended kin have relatively high SES, but extended kin's SES is unrelated to children's education when their resident family also has high SES (Anderson et al. 2018; Jæger 2012; Zang et al. 2024). Nonresident parents’ willingness to compensate and children's expectations of compensation, however, might vary by the nonresident parent–child relationship history. For example, parents transfer less money to children in young adulthood when parents separate early in the child's life course (Furstenberg et al. 1995).

Second, nonresident parent wealth might increase children's likelihood of pursuing additional education by improving their educational success during childhood or through a normative expectation that they will remain in school. Fathers’ formal child support payments are positively associated with academic success (standardized test scores, grades, teachers’ ratings, degree attainment, and years of education; Amato and Gilbreth 1999), and informal (but not formal) support is positively associated with children's standardized vocabulary test scores (Nepomnyaschy et al. 2012). Wealth might be particularly salient to informal transfers because the size of formal child support payments is determined by income rather than wealth (Garfinkel et al. 1994).

Nonresident parent wealth might also affect children's pursuit of educational attainment via parents’ expectations for children, as children will one day inherit and manage their parents’ wealth. Anticipation of future wealth transfers might shape children's educational, occupational, and financial goals and outcomes, either through parents’ or children's behaviors (Gibson-Davis and Hill 2021; Hällsten and Pfeffer 2017). Grandparent wealth is positively associated with grandchild academic performance (net of parent characteristics) in Sweden, where high-quality public schools limit wealth's purchasing power (Hällsten and Pfeffer 2017). This finding suggests that wealth outside the home shapes children's educational success beyond what it can buy, perhaps by shaping parents’ (and grandparents’) expectations and children's attitudes and behaviors. We expect that nonresident parents, too, intend to pass their wealth on to their children and thus might express expectations that shape their children's educational attainment.

Research Questions

We evaluate nonresident parent wealth as a potentially important economic resource for a large fraction of children. To our knowledge, our study is the first to directly examine nonresident parent wealth. We contend that nonresident parent wealth benefits children and that it is excluded from conventional measures of household wealth for most of the 26.5% of children who have a biological parent residing outside of their home. Given the high wealth inequality in the United States, we ask the following:

Methods

Data

We use data from the 1984–2021 PSID (2021a). The PSID is the longest-running panel study of U.S. households. It began in 1968, surveying approximately 5,000 households on a battery of socioeconomic, demographic, and household indicators, including income, employment, family and household composition, and government program participation. The PSID collected information on household wealth, assets, and debts every five years between 1984 and 1999 and biennially thereafter, providing 15 waves of data on children's resident and nonresident parents’ wealth through 2021 (PSID 2023).

Crucial to our purposes, the PSID features a genealogical design. When households are newly recruited to participate in the study, all coresident family members are tagged as followable sample members, meaning that they remain eligible to participate in subsequent interview waves regardless of where they live. Children who are later born to or adopted by a followable PSID sample member and who have lived with that parent at any interview inherit this followable status, the PSID gene. Gened parents and their children who cease living together remain eligible to participate in PSID in their separate households. We hereafter use the term followable nonresident parents to describe the parents in our sample.² Because followable nonresident parent sample status is determined before the birth of their child, continued study contact with them is unrelated to their relationship with the child's resident parent. This sample of followable nonresident parents is free from the sources of bias present in child-centered studies that have sought to contact nonresident parents through a child's coresident parent, who may lack or decline to provide accurate contact information (Garfinkel et al. 1998). Therefore, the PSID provides a novel opportunity to examine the economic context of children's nonresident parent households.³

We use data from two supplements to the biennial PSID interview to estimate how nonresident parent wealth flows to children and is associated with children's educational attainment. First, the PSID's Child Development Supplement (CDS) includes children born since 1985. The original CDS includes up to two gened children per family aged 0–12 in 1997 who were followed at five-year intervals over up to three waves (1997, 2002, and 2007; N = 3,563). The ongoing CDS includes all gened children aged 17 or younger in 2014 (N = 4,333) or 2019 (N = 4,629), with children added as they were born into or entered PSID families between CDS waves. The CDS contains information on children's well-being, family and peer relationships, time use, and other information pertinent to children's development. We use the CDS to estimate how wealth shapes nonresident parents’ contact with children (PSID 2022).

Second, we use the PSID's Transition into Adulthood Supplement (TAS) to identify the association between nonresident parent wealth and children's educational attainment. The biennial TAS began in 2005 and was designed to collect information on school, work, and family formation during early adulthood, at ages 18–28 (PSID 2021b). We use information on postsecondary enrollment and completion from the TAS and completed years of education from the main PSID questionnaire to determine whether youth (1) have ever been enrolled in college and (2) earned a bachelor's degree.

Samples

We use four analytic samples. The first is children who have one resident parent and one followable nonresident parent with reported wealth available from both households. We begin with the full sample of gened children present in at least one wave where wealth was reported (n = 22,683). We limit the sample to children ever observed with one resident and one nonresident parent (n = 10,611). Table A1 (all tables designated with an “A” appear in the online appendix) describes resident household and nonresident parent wealth distributions for PSID children with two nonresident parents. We further limit the sample to children whose nonresident parent is followable (n = 4,112). Finally, we limit the sample to children for whom followable nonresident parent household wealth was observed at least once, producing a sample of 2,342 children (and 5,960 child-waves). The analytic sample includes 5,536 child-waves after listwise deletion in regression models. Estimates from the quantile treatment effect models (Table 3) are robust to multiple imputation as an alternative approach to handling item nonresponse on covariates (online appendix, section B). We use this sample to answer Research Questions 1a, 1b, and 2.

The second analytic sample is a subset of the first. It includes eligible children who ever participated in the CDS and is restricted to those who have information on contact with the nonresident parent in the past year (N = 797, including 988 child-waves with no missing data on covariates).⁴ We use this sample to answer Research Question 3.

The third and fourth analytic samples are also subsets of the first. They include eligible children who reached age 19 for college attendance and age 24 for bachelor's degree attainment by 2021. We limit the sample to youth whose college enrollment status could be determined from PSID and TAS and who had nonmissing values for all control variables (N = 1,365 for college attendance and N = 1,022 for bachelor's degree attainment). We use these samples to answer Research Question 4.

Because PSID followable status is random with respect to household wealth and its determinants, children whose nonresident parent is followable should not have substantively different wealth levels than children whose nonresident parent is not followable. Indeed, we find that household wealth for children whose nonresident parent is not followable is similar to that of children in our analytic sample (Table A2). However, we might overestimate nonresident parent wealth in two ways. First, children must have been observed living with a parent who eventually became nonresident during at least one survey wave, which characterizes the experience of 30% of children born to married parents (Brown et al. 2016) and 71% of children born to cohabiting parents (Bzostek and Berger 2017). Formerly coresident parents might positively select on characteristics related to wealth relative to parents who never lived together. Second, estimates of nonresident parent wealth might be biased by survey (non)response. Table A3 compares child household and nonresident parent socioeconomic characteristics before and after separation by followable nonresident parents’ survey participation status.

Table 1 shows weighted descriptive statistics for person-waves when the nonresident parent's wealth is reported, stratified by child-level characteristics (panel A), nonresident parent household characteristics (panel B), and resident parent household characteristics (panel C). On average, children lived with their nonresident parent for 5.31 years. Children's nonresident parents are most often fathers (82%). On average, resident parents have more children in their family unit (2.27) than nonresident parents (0.51). Nonresident and resident parents are roughly equally likely to be remarried, but nonresident parents are somewhat more likely to be cohabiting than resident parents (16% vs. 10%). Fourteen percent of nonresident and 15% of resident parent households include a child who is a half-sibling to the focal child (i.e., the parent had a birth with a new partner). Family income is approximately $20,000 higher in nonresident parents’ households than in resident parents’ households despite their similar educational attainment levels and union status distributions.

Measures

Nonresident and Resident Parent Household Wealth

We use the imputed value of household wealth created by PSID for public use.⁵ It is the combined wealth of all persons living in the household at the time of the survey and is the summed value of eight housing and asset components: equity in business; transaction and savings accounts; value of debt, except mortgage and vehicle debt; equity in real estate; equity in stock; equity in vehicles; other assets; and equity in individual retirement accounts. Depending on the research question, we index household wealth to either the nonresident or the resident parent's household. We transform all wealth values to 2021 U.S. dollars using the Consumer Price Index.

Homeownership is represented by an indicator of whether anyone living in the (non)resident parent's home is the full or joint owner of the property (vs. renting or neither renting nor owning the property).

Nonresident and Resident Parent Sociodemographic Characteristics

We index parent-level sociodemographic variables to the resident or nonresident parent's household. Indexed variables include whether the parent in the household is the father (vs. mother), the number of children in the family unit, the parent's current union status (not in union, cohabiting, or married), whether any younger half-siblings are in the household (signaling fertility after union dissolution), parental education (in years), whether the household reference person is employed, and household income adjusted for family size in thousands of 2021 dollars. Information about the focal child includes age, race and ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and other), the mother's marital status at the child's birth, the number of years the child lived with the nonresident parent, and whether the child lives in a doubled-up household.

Nonresident Parent Contact

Nonresident parent contact is reported by the focal child's primary caregiver in the CDS for children with a living nonresident mother or father. Response options range from not at all to several times per week. The two modal categories were one to three times per month (19.74%) and several times per year (21.26%). We categorized one to three times per month or more as “higher frequency” and anything less than one to three times per month as “lower frequency.”⁶ Slightly more than half the sample (55.77%) was categorized as higher frequency.

Child Support Payment and Receipt

The main PSID interview gathers information on the household reference person's and spouse's/partner's child support receipt and payment⁷ in each month of the prior calendar year. We collapse the total number of months into a two-category variable (1 = received all months, 0 = received some months or none). We also sum reported values across all months to estimate the annual child support received by a child's coresident biological parent and paid by their nonresident parent (transformed to 2021 dollars). Because we do not directly observe transfers between households, we cannot determine whether the nonresident parent's child support was paid to the child's resident parent or whether the amount the resident parent received was paid by the child's nonresident parent.

Children's Educational Attainment

The children's educational attainment measure is constructed from 11 items in the TAS to ascertain the respondent's highest level of education, including categories for currently enrolled in college and completed some college. When this information is unavailable, we use the number of years of completed education from the main PSID interview, coding those with 13 or more years of education as college attendees and those with 16 or more years of education as college graduates. We use the highest level of education observed at ages 19–28 for college attendance and 24–28 for college graduation and control for the age at which each is observed.

Analytic Approach

Our analysis proceeds in five stages. We weight descriptive statistics and analyses by the child's individual longitudinal weight and cluster standard errors on the PSID family of origin. First, we use box and whisker plots to describe children's resident and nonresident parent household wealth distributions.

Second, we describe the wealth concordance between children's resident and nonresident parents using rank–rank regressions. For resident and nonresident parents separately, we rank household wealth from lowest to highest value, divide the rank position by the total number of children (N = 2,342), and multiply by 100 to obtain percentile rank. We calculate percentile ranks for children's first observation after parental separation and last observation in childhood. Following Pfeffer and Killewald (2018), we regress nonresident parent wealth percentile rank on resident parent wealth percentile rank, controlling for resident and nonresident parents’ age and age squared.

Third, we use quantile treatment effects (QTEs) to estimate how sociodemographic characteristics are associated with household wealth holding at the 25th, 50th, and 75th percentiles, separately for children's resident and nonresident parents (Borgen et al. 2022; Rios-Avila and Maroto 2022). In these models, coefficients associated with categorical variables represent group differences in wealth at each quantile.⁸ Coefficients associated with continuous variables represent wealth differences at a given quantile in response to a one-unit change in the value of the independent variable. We estimate QTEs using residualized quantile regression, a two-step approach that regresses the independent variable of interest (what QTE models call the treatment variable) on control variables in the first step (ordinary least-squares [OLS] regression) and then uses the residualized outcome as the independent variable in the second step (conditional quantile regression) (Borgen et al. 2021:3). QTEs are advantageous for two reasons. First, given that they do not depend on the conditional mean to estimate coefficients, extreme outliers do not disproportionately influence the results, which is important because wealth is highly positively skewed. Second, QTEs allow us to identify variation in the correlates of (non)resident parent wealth across the wealth distribution.

We rotate through the sociodemographic and family demographic variables listed in Table 1, assigning each as the independent variable of interest and estimating its association with (non)resident parent household wealth at each quantile. We separate these variables into two categories: (1) those that are time-invariant or observed before separation (child's race and ethnicity, parent's gender, mother's marital status at child's birth, years lived with the nonresident parent, parent's education [indexed to time of separation]) and (2) those observed after separation (child's age, number of children in the family unit, parent's current union status, younger half-sibling in the household, household reference person's employment, total family income, whether the focal child and nonresident parent live in the same state, and whether the focal child lives in a doubled-up household). We first focus on preseparation variables, including only survey year as a control to account for year fixed effects. We then turn to postseparation variables, controlling for the survey year and all preseparation variables. Because no prior research has explicitly examined nonresident parent wealth, our goal is descriptive, and we discuss the results as such. Yet, we can rule out the potential influence of some confounders by adjusting for preseparation variables in our postseparation model.

Fourth, we estimate the association of nonresident parent wealth with parent–child contact and child support receipt/payment. Wealth is measured in quartiles. We use logistic regression to predict the nonresident parent's frequency of contact with the child (higher vs. lower) and the resident parent's past-year child support receipt (all months vs. some or none). We use OLS regression to predict the log-transformed amount of child support paid and received (conditional on paying/receiving any child support).

Fifth, we use logistic regression to estimate the association of nonresident parent wealth with children's college attendance and bachelor's degree attainment in young adulthood. We average nonresident parent wealth across observed waves and create quartiles from this average. For this analysis and analyses predicting contact and support, the first model presents unadjusted estimates. The second model controls for the nonresident parent's and child's sociodemographic characteristics that might be confounders, as well as for the resident parent's household wealth. In analyses predicting education, we average or create childhood-long indicators of sociodemographic characteristics (e.g., child's age averaged across observations; whether the nonresident parent ever married a new partner).⁹

Results

Distributions of Nonresident and Resident Parent Wealth

Figure 1 presents box plots illustrating the distribution of nonresident (panel a) and resident (panel b) parent household wealth in thousands of 2021 dollars by current marital status. For nonresident parents, wealth holding is $8 at the 25th percentile, $16,420 at the median, and $94,272 at the 75th percentile. These figures largely match the wealth distribution among resident parents, who have slightly more wealth at each point in the distribution. Nonresident and resident parent wealth is similarly stratified by current marital status, with married parents having more wealth across the distribution than nonmarried parents. Married nonresident parents hold more household wealth at the 75th percentile than resident parents ($212,585 vs. $184,404), whereas the opposite is true for nonmarried parents ($58,478 vs. $82,962).

Wealth Concordance Between Nonresident and Resident Parents

The distributions of household wealth for resident and nonresident parents are similar, but how does this translate to wealth concordance between resident and nonresident parents within families? Table 2 shows results from rank–rank regressions. The correlation between nonresident and resident parent wealth is .30 in the first observation after parental separation and .28 in the last observation in childhood. To contextualize this finding, we note that Pfeffer and Killewald (2018) found rank slopes of .39 for parent–child wealth and 0.23 for grandparent–grandchild wealth using the same regression approach.

Correlates of Nonresident and Resident Parent Wealth

Table 3 shows results from QTE regression models describing differences in (non)resident parent household wealth on preseparation characteristics at the 25th, 50th, and 75th percentiles, adjusting for the survey year. Nonresident and resident parent households of Black and Hispanic children have less wealth than those of White children, and these disparities grow across the wealth distribution. Mixed and other race children fall in the middle: their (non)resident parent household wealth is greater than Black children's (pairwise contrasts not shown) but less than White children's at the 50th and 75th percentiles. Nonresident parent household wealth is $8,675 greater at the median (p < .01) and $40,287 greater at the 75th percentile (p < .01) when the nonresident parent is the focal child's father (vs. the mother). Resident and nonresident parents of children born to never-married or previously married mothers have less wealth than resident and nonresident parents of children born to married mothers, and these disparities increase across the distribution. Resident and nonresident parents who lived with their children longer before separation hold more wealth, perhaps because of the benefits of cohabitation/marriage for wealth accumulation or selection into longer relationships on higher wealth. Parent education is also positively associated with wealth holding, and these disparities are larger at higher levels of the wealth distribution.

In Table 4, we turn to postseparation family and sociodemographic variables, adjusting for all time-invariant and preseparation variables as well as survey year. In nonresident parent households, a one-year increase in child age is associated with $154 more wealth at the 25th percentile, $1,855 more wealth at the 50th percentile, and $8,013 more wealth at the 75th percentile. An additional child in the family unit is negatively related to wealth among resident parents but positively related among nonresident parents. Yet, having a younger half-sibling in the household is positively associated with wealth across the distribution for both nonresident and resident parents. One explanation is that other children in a resident parent household might be the focal child's biological siblings who are present regardless of whether the resident parent has repartnered. By contrast, children in a nonresident parent's household might be half- or stepsiblings added to the focal child's family as a result of that parent's new union.¹⁰ Indeed, (re)married parents have much higher wealth throughout the distribution relative to parents not in a union, likely reflecting their positive selection into marriage as well as ongoing wealth accumulation in marriage. Resident and nonresident parents have significantly higher wealth in the middle of the distribution when they are cohabiting (relative to not being in a union). Householders’ employment is also positively associated with wealth, except for resident parent households at the 75th percentile. Nonresident parents who reside in the same state as their child have greater wealth at the 50th and 75th percentiles than nonresident parents who live elsewhere. Living in a doubled-up household is associated with significantly higher wealth for children's resident parent households, with increasing benefits at higher points in the distribution, but is not associated with nonresident parent wealth.

Nonresident Parent Wealth and Contact With and Support of Nonresident Children

Figure 2 shows the predicted probability of frequent parent–child contact at each quartile of nonresident parent wealth before and after adjusting for sociodemographic characteristics and resident parent wealth. Nonresident parents in the highest wealth quartile are more likely to have greater contact with children than nonresident parents in all other quartiles, although these differences are not statistically significant at conventional levels in the adjusted model.

Figure 3 shows unadjusted and adjusted values for the predicted probability of regular child support receipt (panel a) and conditional (logged) amount of child support received (panel b) or paid (panel c) in the past year. We describe the adjusted estimates. Children's residential households are more likely to receive child support in all months when nonresident parents occupy the top two quartiles relative to the bottom quartile. Conditional on any receipt, children's residential households receive significantly larger amounts when nonresident parents occupy any of the top three quartiles relative to the bottom quartile. Additionally, nonresident parents in the highest wealth quartile pay more in child support than those in the three lower quartiles. In sum, wealthier nonresident parents have more frequent contact with their nonresident children, pay child support more often, and pay more child support conditional on paying any. Results from the full regression models summarized in Figures 2 and 3 appear in Tables A5 and A6.

Nonresident Parent Wealth and Children's Educational Attainment

Figure 4 shows the unadjusted and adjusted probability of children's college attendance (panel a) and bachelor's degree attainment (panel b) by nonresident parent wealth quartile. Estimates from the full regression models and analogous models estimating the log odds of high school graduation appear in Table A7. The unadjusted estimates show that children whose nonresident parents occupy the top two wealth quartiles are significantly more likely to attend college than those in the lowest quartile. After we control for nonresident parent characteristics and resident parent wealth, a positive association remains (highest vs. lowest quartile odds ratio = 1.27) but is reduced in magnitude and no longer statistically significant.

Panel b of Figure 4 shows the predicted probability of bachelor's degree attainment. In the unadjusted model, children whose nonresident parents occupied the top three quartiles during childhood are significantly more likely than those from the lowest quartile to attain a college degree. In the adjusted model, we still observe that children whose nonresident parents occupied the highest quartile in childhood have a significantly higher probability of earning a bachelor's degree (.31) compared with those from the bottom quartile (.21; p < .05).

We tested whether results for contact, support, and education varied by nonresident parent family structure (i.e., union status and the focal child having a younger half-sibling by the nonresident parent). We report and discuss these results in Tables A8 and A9.

Discussion and Conclusion

Wealth inequality has been rising and is particularly high among households with children (Gibson-Davis and Percheski 2018; Pfeffer and Waitkus 2021). Furthermore, wealth is associated with multiple indicators of children's well-being (e.g., Boen et al. 2021; Shanks 2007). Yet, this area of research has focused almost exclusively on children's resident household wealth, primarily because of data limitations. We responded to recent calls to describe dynamics in families that do not reside in a shared household (e.g., Seltzer 2019)—a vital task, given that 26.5% of children live apart from one parent (Grall 2020)—by providing the first estimates of nonresident parent wealth for minor children. We estimated nonresident parents’ wealth, how it compares with children's residential households, children's access to nonresident parents’ wealth, and whether and how it matters for one measure of life chances: educational attainment. We highlight four contributions.

First, we found sizable variation in nonresident parent wealth. At the 25th percentile, nonresident parents have just $8 in net wealth, and many are in net debt. Nonresident parents have $16,420 in wealth at the median and $94,272 at the 75th percentile. We observed a strong pattern of wealth concordance between resident and nonresident parent households across the wealth distribution. We also found that the correlation between nonresident and resident parent wealth position, net of age, is roughly .30. Wealth concordance between resident and nonresident parents complements previous work showing relatively similar characteristics—in terms of family complexity—of children whose parents live apart (Cancian et al. 2011). That work proposed assortative mating as one explanation, which we echo for our case of household wealth. Indeed, research has found that in the United States, couples sort on parental wealth (Charles et al. 2013). Moreover, in most U.S. states, married parents who divorce split assets equally (Hersch and Shinall 2020), likely further contributing to wealth similarity between households. Future research should examine how families with children accumulate and divide specific assets and debts as they form and dissolve unions, paying close attention to how child residence shapes which assets (non)resident parents retain.

Second, nonresident and resident parent wealth are moderately positively correlated but are not equal. Results from our QTE models revealed some sources of heterogeneity in wealth holding between separated parents of a shared child. For example, children doubling up was strongly positively associated with resident parent wealth (which is sensible, given that the child lives with another adult in addition to their resident parent) but not at all associated with nonresident parent wealth. This finding suggests that resident parents may double up independent of nonresident parent's wealth. Yet, many of the other variables we examined had associations that were similar in magnitude in resident and nonresident parent households. We conclude that differences in household wealth between nonresident and resident parents are likely driven by differences in household composition and parents’ characteristics rather than in the effect of such factors on wealth accumulation.

Moving past differences between nonresident and resident parents, many sociodemographic characteristics (e.g., child's age, parent's gender, child's race and ethnicity, parental union status, and parental education and employment) are associated with nonresident parent wealth, in line with prior research on the correlates of household wealth. Most of these disparities grew at higher levels of the wealth distribution. Describing the wealth of nonresident parents builds on past research on the relationship between household structure and wealth (Maroto and Aylsworth 2017; Ozawa and Lee 2006; Percheski and Gibson-Davis 2022; Yamokoski and Keister 2006), highlighting inequality in an understudied economic resource in children's family systems.

Third, nonresident parents’ wealth is positively related to two dimensions of parental investments that are important for children with a nonresident parent: nonresident parents: contact and child support payments. Wealth might directly facilitate contact and support by freeing up parents’ time and resources, nonresident parents with higher wealth might be more likely to make frequent contact and complete payments (i.e., selection), or both processes could operate. Our findings suggest that children have some access to their nonresident parent's wealth, and access likely increases with wealth. Prior research has documented that household wealth shapes parenting—for example, by increasing parents’ capacity to instill cultural capital in children (Orr 2003). Additionally, nonresident parents’ child support arrears are associated with children's internalizing and externalizing behaviors, demonstrating the impact of nonresident parents’ (missing) economic support on child well-being (Nepomnyaschy et al. 2021). Nonresident parent wealth, too, might be implicated in shaping children's well-being (via on-time and complete child support payments). More broadly, nonresident parent wealth might signal class status and reduce parental stress (Conley 1999; Gibson-Davis and Hill 2021), each of which might benefit nonresident children. In terms of other nonpecuniary benefits of wealth, 40% of nonresident parents in our sample were homeowners. Given the well-established association between the home environment and child well-being (Evans 2004), nonresident parent homeownership might provide advantages to children.

Fourth, nonresident parent wealth is positively predictive of college completion net of other nonresident parent characteristics and resident parent wealth. Our results for college attendance were more tenuous; after controlling for resident parent wealth, we found that nonresident parent wealth is positively but nonsignificantly associated with college attendance. This result corresponds to a recent finding by Zang and colleagues (2024), who documented a positive but nonsignificant association between grandparent wealth and college attendance. Additional analyses indicate that nonresident parent wealth is not associated with high school graduation (Table A7). This pattern suggests the purchasing power of nonresident parent wealth, which is particularly important for educational attainment when education is expensive.

Our analyses have limitations. Estimates of nonresident parent wealth are likely positively biased because nonresident parents who are survey nonrespondents are less advantaged than survey respondents. Our sample also does not contain children whose parents are nonresident because they are in institutions (e.g., prison, the military), although we do capture these parents if they reentered the PSID. This exclusion might further inflate our estimates because incarcerated individuals have much lower wealth than those never incarcerated (Sykes and Maroto 2016). Comparisons in Table A2 indicate that nonresident parents who never responded were similar—or, in some cases, more advantaged—across sociodemographic characteristics to occasional responders, suggesting that we captured some of the wealth distribution of nonresident parents who are nonrespondents.

Further, we could not discern the amount of time children lived with nonresident parents. Custody arrangements vary, and they might do so with wealth. Children might benefit more from nonresident parents’ wealth, regardless of wealth level, when they spend more time with nonresident parents. Future data collection efforts should document how much time children spend in other households and with nonresident parents.

Finally, all our analyses estimated associations, not causal relationships. This consideration is especially important for our analyses predicting contact, support, and educational attainment, given that other unobserved characteristics might bias the observed associations between wealth and these outcomes.

Despite these limitations, we offer nonresident parents’ wealth as an understudied but important economic resource, given that more than a quarter of children have a parent from whom they live apart (Grall 2020). We answer calls for research to better understand social mobility processes in nontraditional families (e.g., Tach 2015). Our findings might also lend insights into heterogeneity in health and well-being among children with a nonresident parent: we demonstrated variation in economic resources external to the child's household. Because family relationships and processes are not bound to a single household (Seltzer 2019), depicting the spread of resources across households, but within a family system, is critical to understanding broader patterns of inequality in American family life.

Acknowledgments

A previous version of this paper was presented at the 2022 annual meeting of the Population Association of America in Atlanta, GA. The authors thank Fabian Pfeffer for comments on a previous draft of the manuscript; Davis Daumler for fruitful discussions about the analyses; and Laura DeMarco, Jane Furey, and Alexandra Killewald for feedback on the presentation of results. Any errors are those of the authors. The authors acknowledge partial support from the Eunice Kennedy Shriver National Institute of Child Health and Human Development under awards R01HD088506 (PI: P. Fomby) and P2CHD041028 (University of Michigan Population Studies Center, PI: S. Burgard). The collection of data used in this study was partly supported by the National Institutes of Health under grants R01 HD069609 and R01 AG040213 and the National Science Foundation under awards SES 1157698 and 1623684.

Notes

References