Association for Public Policy Analysis & Management

The structure of the average American family has changed over the last several decades.  In 1980, 73% of children lived with two married parents in their first marriage, compared to only 46% in 2013 (Wang & Parker, 2014). In 2014, over 24 million children, or 35% of children in the United States, were living in single-parent families. The composition of single-parent households has also changed, with a growing number of parents having never been married (Caumont, 2013). Single parent households and unmarried parents in particular are more likely to live in poverty (McLanahan, 2010). Child support is an important source of income for single parent households. Child support may be provided through the formal system but may also be provided informally by fathers, which is particularly common among lower-income families. Yet little is known regarding the trajectories of father’s provisions of child support as the time since the parents’ separation increases, particularly among never married parents.

Previous cross-sectional research has found that the number of years since divorce is associated with a decrease in formal child support payments among previously married parents (Seltzer, Schaeffer, & Charng, 1989). Nepomnyaschy and Garfinkel (2010), using longitudinal data, analyzed formal, informal, in-kind, and total child support trajectories up until the child was 5 years of age among non-marital parents.  Findings revealed that that total child support decreased for the first 15 months after parents stopped cohabiting and then began increasing after 45 months. These studies offer important preliminary information regarding the trajectories of child support. However, previous research has not yet provided a comprehensive picture of child support trajectories for both previously married and never married parents and their long-term trajectories.

This study seeks to address this gap by examining the total package of child support (formal, informal, non-cash or in-kind, and total cash support) that residential mothers receive from non-resident fathers up to when the focal child is 15 years old.  Using data from the Fragile Families and Child Wellbeing Study, we examine child support among parents who were previous married, never married and never cohabited, and never married but previously cohabited.  This study will analyze how these contributions vary as time since parents’ separation increases and how the changing levels of support interact to affect the total child support received.  Results will be estimated using locally weighted polynomial (lowess, or loess, regressions), random effects, and individual fixed effects regressions.

Preliminary lowess figures reveal a consistent, and surprising, increase in formal child support received per month through age 9 among previously married, previously cohabiting, and never cohabiting parents.  However, this trend may be driven by attrition bias. By year 9 respondents receiving lower levels of support are more likely to have left the sample among previously married and previously cohabiting parents. We will extend these preliminary analyses and use multiple imputation to test and correct for the effects of attrition bias.