Predicting the Impact of Savings On Educational Outcomes

Background: The gap in educational achievement between children in poor families and their counterparts in wealthy families has sharply widened over the last few decades (Rank, 2004). An effective intervention for improving educational development and outcomes of children of the poor is asset-building that supports the poor for overcoming the barriers that keep them from building and sustaining assets (Sherraden, 1991). Savings as asset-building strategies can play an important role in supporting low-income households by providing incentives to encourage them to invest in assets for their future and for their children’s education (Grinstein-Weiss et al., 2012). However, some researchers pointed out that the effectiveness of savings on educational outcomes was unclear. There was little evidence that the poor were able to achieve postsecondary education such as enrolling in and completing college (Richards & Thyer, 2011). Savings from IDAs, as well, did not account for college completion and high degree attainment (Grinstein-Weiss et al., 2012). Thus, this study seeks to predict the impact of savings on young adults’ educational outcomes.

Methods: This study used data from the National Longitudinal Survey of Youth 1997. The 3,109 samples were derived from two Rounds of surveys in 1997 and 2007. Propensity score matching (PSM) was used to create treatment and control groups that are statistically similar at baseline. Four different types of PSM method were run to evaluate the savings’ impact as treatment. Generalized logistic regression and ordinary least square (OLS) regressions were run to predict the main factors for educational outcomes and examine the mediating effects according to Baron and Kenny’s approach, respectively.

Results: Results from Radius, Kernel, and Stratification matching indicated that the group holding savings for education is more likely to achieve a college degree by about 25%, 12%, and 8%, respectively, than the group without savings. Based on the results from generalized logistic regression, savings for education (OR=1.680, p<.001), young adults’ educational expectations (OR=1.023, p<.001), parents’ educational expectations (OR=1.025, p<.001), the mother’s age at birth (OR=1.048, p<.005), mother’s education level (OR=1.157, p<.001), and parents’ net worth (OR=1.386, p<.001) are statistically significant for college graduation. However, the significance of permanent household income disappeared as the main variables, such as savings and young adults’ and parents’ educational expectations, were added. Finally, results from OLS regressions indicated that the mediators were not statistically associated with savings.

Implications: This study found that savings for education are effective in encouraging young adults to achieve a college degree. In addition, young adults’ and parents’ educational expectations are strong in predicting young adults’ college graduation, even though they do not mediate the association between savings and educational outcomes. Interestingly, the parents’ net worth is important indicators influencing achievement of a college degree, whereas permanent household income is not associated with such achievement. Thus, these findings suggest that asset-based policies and programs encouraging the poor to hold savings and build assets may be a desirable policy strategy to help improve higher educational attainment of young adults from low-income households.