Skip to main content

Advertisement

Log in

Children’s Gender and Investments from Nonresident Fathers

  • Original Paper
  • Published:
Journal of Family and Economic Issues Aims and scope Submit manuscript

Abstract

Evidence suggests that fathers have stronger ties to sons than daughters, which may result in differential investments in their children. This paper investigated whether girls’ gender restricts their access to fathers’ contributions if they do not live together. The data used were the 1994–2008 March/April Match Current Population Survey Child Support Supplements, a large, nationally representative sample which identifies child support eligible mothers of all marital statuses and collects information on nonresident fathers’ financial and social investments in their children. Results for court-mediated outcomes such as the existence and amounts of child support orders showed that courts do not allocate child support differentially by child gender. Small but suggestive effects of child gender were found on fathers’ post-dissolution investments, but these effects disadvantaged boys rather than girls.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

Similar content being viewed by others

Notes

  1. Some examples are Aizer and McLanahan (2006), Allen et al. (2011), Beller and Graham (1993), Case et al. (2003), Freeman and Waldfogel (2001), Hanson et al. (1996), Smock and Manning (1997), Sorensen and Hill (2004). “Child support enforcement environment” refers broadly to the laws in place in a given city, state, or region, the resources allocated to enforce them (Rich et al. 2007), judicial and individual attitudes toward the provision of child support, and other factors such as the availability of off-the-books employment, which impedes wage withholding as a mechanism for child support collection (Kurz 1995; Rich et al. 2007).

  2. Lundberg (2005) and Raley and Bianchi (2006) provide reviews.

  3. Married women with child-support eligible children in the CPS-CSS data are referred to as remarried mothers in this paper for parsimony, although we cannot observe whether the previous parental union was marriage, cohabitation, or not coresidential.

  4. The CSS asks questions in April of the survey year about events of the previous year. This paper consistently refers to the data by the year the survey was fielded, but it should be remembered that the responses reported refer to the previous year.

  5. Note, the mothers of these fathers’ children have been even more disadvantaged than the fathers, because they must both parent and work, and because women have been paid less than men (Cancian et al. 2011).

  6. Abundant qualitative evidence has supported this notion, e.g., Edin et al. (2009).

  7. Contact can be beneficial in a father-child relationship, but detrimental if there are negative interactions with the child or conflict between the parents (e.g., Hofferth et al. 2010).

  8. The father effects and child effects terminology originated in the psychology literature (e.g., Bell and Chapman 1986; Russell and Russell 1992). An early study of child effects in the child support literature was Aughinbaugh (2001), who found that higher scores on children’s achievement tests increased the likelihood of receiving child support and the amount received.

  9. Initially the April supplement was directed only towards mothers; fathers were added in the 1992 survey (Scoon-Rogers and Lester 1995).

  10. Because the sample in the current study includes children from mothers of all marital statuses, the results were not biased by selection into marital status by child gender.

  11. Beller and Graham (1993) presented time series for survey years 1979–1986. As would be expected, the Beller Graham numbers, although not included in the figures, conformed almost exactly to the corresponding dates in the CPS-CSS series in Figs. 1 and 2.

  12. Surprisingly little is known about the prevalence of joint physical and joint legal custody at the national level (Bartfeld 2011); the CSS is unusual in collecting nationally representative information. Earlier national figures showed that about 5% of divorced families had joint physical custody and 20% had joint legal custody (Nord and Zill 1996, Table 1). In the CSS sample in the current study the incidence of both types of custody were higher for divorced mothers, reflecting the increase over time in joint custody arrangements after divorce (Bartfeld 2011; Cancian et al. 2014).

  13. Tables 5 and 6 in the Appendix show averages for the 1994 and 2008 samples respectively. Results for the separated mothers will not be discussed because the sample size is insufficient to estimate credible time trends.

  14. Cancian et al. (2014) studied the increase in joint physical custody arrangements in Wisconsin through 2008 and concluded that it was largely driven by changes in social norms.

  15. Chen (2015) studied the increasing incidence of joint legal custody for nonmarital children in Wisconsin in the period 1988–2009. She concluded that increasing preference for this custody arrangement, encouraged by a policy change that made it presumptive, brought about this result, rather than a change in the demographic composition of never-married parents.

  16. The remaining significant differences were an increase in the average number of children for separated women, a decrease in the presence of never-married women (and the all-women sample) in central-city MSAs, and an increase in the household size of divorced women (and the all-women sample). These characteristics were controlled for in the multivariate analysis.

  17. The variables of interest \( \varvec{x} \) reported in the regressions in the current study were all binary, so the AME for each regressor \( x_{i} \) was calculated as follows: for the first observation in the data, \( x_{i} \) is set to one, all other regressors are at the observed values for this observation, and \( P(y = 1|x_{i} = 1) \) was calculated. Then \( x_{i} \) was set to zero, all other regressors remained at the observed values for this observation, and \( P(y = 1|x_{i} = 0) \) was calculated. The difference \( P\left( {y = 1 |x_{i} = 1} \right) - P\left( {y = 1 |x_{i} = 0} \right) \) was the marginal effect for the first observation. This calculation was performed for each observation, and the resulting n marginal effects were averaged to obtain the AME. If \( x_{i} = 1 \) represented female and \( x_{i} = 0 \), male, the AME gave the estimated difference in \( P\left( {y = 1} \right) \) between women and men, controlling for all other variables. In this paper the average marginal effects are referred to as simply the marginal effects. There are a number of ways to evaluate marginal effects for nonlinear binary response models, for which the following terminology has evolved: calculating average marginal effects at observed values of the regressors (AME), marginal effects at the means of the regressors (MEM), or marginal effects at representative values of the regressors (ME). Noted advantages of AMEs were that they use all of the data and the use of observed values of the regressors yields more “realistic” estimates (Bartus 2005; Cameron and Trivedi 2010, p. 343; Williams 2012).

  18. These and subsequent statements of statistical differences between coefficients for the remarried, separated, and never-married mothers groups were based on unreported computations, and significant at the 5% or 1% levels.

  19. Thanks to the referee who suggested investigating father contact for the older children.

  20. Mammen (2008) also found that never-married mothers with at least one son received smaller amounts of child support than did never-married mothers of all daughters.

  21. An arguably related finding was that being born outside of marriage significantly increased sons’ but not daughters’ externalizing problems.

References

Download references

Acknowledgement

I wish to thank Elsie Pan, Doug Miller, and two referees for their valuable comments; and Carly Dennis and Emma Joslyn for excellent research assistance.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Kristin Mammen.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendices

Appendix 1

See Tables 5 and 6.

Table 5 Sample averages for 1994 sample
Table 6 Sample averages for 2008 sample

Appendix 2: Data Appendix

The CSS data had a number of strengths.

  1. 1.

    The CSS contained measures of outcomes of the legal system such as the existence and amounts of child support orders that mothers have been awarded by the court, and whether the father had joint physical custody or joint legal custody of the child(ren), for women of all marital statuses.

  2. 2.

    The CSS identified children living with a mother and a stepfather who were child-support eligible - that is, which children in married-parent homes were born in previous unions. Married women were sometimes omitted from studies of post-dissolution outcomes when this information was not available. For parsimony these women are referred to as remarried mothers in this paper, although the previous parental union could have been marriage, cohabitation, or not coresidential. This was an important group to study because in well-being measures, children living with stepfathers do no better than children of single mothers (e.g., McLanahan and Sandefur 1994, p. 90). Evidence has also suggested that remarriage reduced amounts of child support from nonresident fathers (Hill 1992).

  3. 3.

    It is likely that child support amounts were measured more accurately in the CSS than in some other surveys, because the CSS questions were asked directly of all household members 15 and over who had children living with them (CPS 2010). In contrast, in the March CPS, for example, one household member was asked about all other members (Campanelli et al. 2005, p. 425).

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Mammen, K. Children’s Gender and Investments from Nonresident Fathers. J Fam Econ Iss 41, 332–349 (2020). https://doi.org/10.1007/s10834-019-09654-y

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10834-019-09654-y

Keywords

Navigation