Research That Matters (January 17 - 20, 2008)

Purpose: For low-income families, homeownership is not only an important strategy to move out of poverty but also an opportunity for long-term socioeconomic well-being. Individual Development Account (IDA) programs facilitate savings toward home purchase through matched savings, financial education, and case management. This study tests the effect of IDA programs on homeownership rates among low-income participants.

Method: Using longitudinal randomized experiment data from the American Dream Demonstration, we examine experiences of 1,103 participants assigned to the IDA treatment group (n = 537) or a control group (n = 566). The affect of IDA participation on homeownership rates among low-income participants is measured after 18 months (Wave 2) and 48 months (Wave 3) in the program. Nine dependent variables measure homeownership and home-search activities. Independent variables include treatment condition, demographic, household, and economic covariates. Univariate analyses and bivariate analyses are used sequentially to identify sample imbalance at baseline, and to determine whether homeownership and home-search activities differ by treatment conditions. Further, multivariate analyses (logistic regression, ordinary least square regression) estimate the impact of IDA treatment on homeownership levels and home-search activities. Last, a three-step mediation analysis (Krull & MacKinnon, 1999) is conducted to test the hypothesis that the effect of IDA treatment on homeownership at Wave 3 would be mediated through home-search activities at Wave 2.

Results: Binary logistic regression reveals participation in IDA treatment did not significantly increase homeownership at Wave 2; however, at Wave 3, the treatment group has 1.53 times greater odds of being homeowners than the control group (OR = 1.53, 1.02-2.28). Further, chi-square statistics show the treatment and control groups differ on one home search activity (clear-up old debts) at Wave 2, and on three home search activities (look at houses for sale, attend open houses, and clear-up old debts) at Wave 3. Logistic regression analyses of home search activities show similar results. OLS regression on the index of home search activities shows that the treatment groups engaged in significantly more home search activities at wave 3 than the control group (â = .40, p = .03).

Mediation analysis indicates that the effect of the treatment on homeownership was mediated through one of the home-search activities (i.e. clear-up old debts) at wave 2 (â= .60, p = .01). The result was also supported by the fact that the chi-square difference between reduced model and full model was significant at .05 level. Results indicate the effect of IDA treatment on homeownership at Wave 3 was mediated by clearing-up debt at Wave 2.

Implications: This study found that IDAs are effective in helping low-income families save for a home, which represent an important strategy for social work practice. Furthermore, this study indicates that the road to homeownership takes time. Recommendations to better design and promote programs to help low-income families save for a home are provided. Further, the importance of using longitudinal random designs in social work interventions is discussed.