Marital Status and the Transition Into Low-Income Homeownership: Evidence From a Propensity Score Analysis

Purpose: Homeownership has been one of the most widely used approaches to increase economic and social stability for the past twenty years. Recently, another approach to reduce poverty has emerged with an emphasis on promoting marriage and strengthening relationships. The purpose of this study is to explore the association between marital status and homeownership among a group of low- and moderate-income (LMI) renters. This study contributes to social work policy by providing a critical link between two historically separate poverty reduction strategies. Moreover, it contributes to social work research as an example of the application of a highly rigorous analytical technique.

Method: This study utilizes five waves of data (2004-2008) from the Community Advantage Panel Study (CAP). CAP is a secondary-mortgage pilot program for low- and moderate-income (LMI) households. Since this study examines the transition to homeownership, the sample consists of low-income married and unmarried renters. One of the primary criticisms of marriage research is the role of selection bias into marriage. In order to address this limitation, and to address the problem that the married group differed from the unmarried group in this sample, we employed propensity score matching (PSM) to control for selection bias into marriage. Survival analysis was then used to estimate the effects of marital status on the transition into homeownership (N = 923). Multiple imputation was used to reduce bias because of missing data. The data are collected yearly and display interval censoring. Thus, a discreet-time proportional hazard model, controlling for static and time-varying covariates, estimates the effect of marital status on the “risk” of homeownership over time.

Results: Results indicate married couples have higher odds of buying a home, and do so at faster rates, than their unmarried counterparts. These findings were robust to the control of selection bias between the married and unmarried groups using propensity score matching. Specifically, using matched samples created by PSM, married participants purchased houses at a rate that was 158%-171% higher than their unmarried counterparts (p < .001). Other significant predictors of timing of house purchase included participant's age at baseline (p < .05), educational attainment, (p < .05), and income (p < .001). In general, participants who were younger, had attained higher levels of education, and had higher income than their study counterparts were also more likely to purchase houses at a faster time-to-event rate.

Implications: This study provides evidence of the effects of marital status on the homeownership trajectory of LMI households. These findings underscore the barriers that unmarried disadvantaged couples face when attempting to accumulate assets. Given the scarcity of research on the impact of marital status on homeownership and that homeownership is an important proxy of economic stability this study offers an important contribution to the social work field. This study also presents one of the first attempts to address self-selection into marriage using propensity score analysis. This innovative and rigorous method allows us to draw a strong association between marital status and the transition to homeownership.