Association for Public Policy Analysis & Management

Financial circumstances and living arrangements are closely linked, and there is a long debate about the role of government assistance in either facilitating or discouraging certain living arrangements. Research on the effects of the EITC on coresidence is just beginning. The limited set of studies that has examined how the EITC influences living arrangements has focused primarily on the likelihood of cohabitation versus marriage (Dickert-Conlin, Houser, & Li 2002; Ellwood 2000; Michelmore 2014), despite evidence that cohabitation increasingly serves as an alternative to dating, rather than a marriage-like relationship, for many couples (Guzzo 2014; Heuveline & Timberlake 2004; Vespa 2014). Further, cohabitation is just one of the coresidential economic coping strategies available to families; coresidence through doubling up is another option, which can be just as economically beneficial (Brown & Lichter 2004). Improved finances via EITC receipt could allow mothers to avoid or exit coresidential relationships; alternatively, decreased financial stress could ease the strain on relationships and encourage dating partners to enter cohabitation. 

The present study investigates whether and how EITC receipt is associated with changes in unmarried mothers’ living arrangements. We examine whether EITC receipt coincides with a change in the fraction of unmarried mothers coresiding with another adult. If there are declines in coresidence around tax time, we then ask whether this is driven primarily by higher rates of exit from or lower rates of entry into coresidential relationships. 

We use the 1996-2004 panels of the Survey of Income and Program Participation (SIPP), which include household rosters that measure coresidence and cohabitation on a monthly basis as well as the information necessary to estimate EITC receipt. To identify the effect of the EITC on coresidence, we compare monthly changes in living arrangements for our “treatment” group—unmarried mothers who received the EITC—to changes in the living arrangements for our “control” group—single women without children (who would have qualified for the EITC if they had children). We estimate a difference-in-difference model that compares differences in living arrangements between the two groups during tax time (February-April) to differences during May-January. Although the fraction of women coresiding differs initially between the “treatment” and “control” groups, the model assumes that any additional differences observed during tax time are due to EITC receipt. We employ a variety of robustness checks, alternative specifications, and placebo tests to verify this assumption. Next, we use decomposition methods to assess whether the differences in coresidence at tax time are driven by differential entries or exits from coresidence, and whether the difference is driven primarily by coresidence with a romantic partner, relatives, or nonrelatives. 

This research will contribute to the fields of policy and family studies, providing a more complete picture of the impacts of EITC policy on coresidential behavior and of how romantic and coresidential decisions can be shaped by the larger policy environment.