Methods: The study uses the data from SEED for Oklahoma Kids (SEED OK), a CDA experiment in the United States with probability sampling from a full state population. SEED OK included 2,704 primary caregivers of children born in Oklahoma during 2007: 1,358 were randomly assigned to the treatment group, and 1,346 to the control group. For treatment children, the experiment automatically opened a CDA for postsecondary education with a $1,000 seed deposit, and parents were encouraged to open their own college savings accounts to save for children. Evaluated outcomes are parental educational expectations (1-5), positive parenting practices (0-150), negative parenting practices (0-75), maternal depression (0-12), and children’s social-emotional development (0-120) measured in the follow-up survey collected in 2011. We limit our sample to those participating in the Head Start or TANF programs (N = 426), and use structural equations modeling to examine the impacts of CDAs on these outcomes controlling for multiple demographic and socioeconomic characteristics.
Results: The CDA in SEED OK has positive impacts on most of the outcome variables. Treatment parents report more positive parenting practices (b=6.33, p<.10), fewer negative parenting practices (b=-.87, p<.05), and less maternal depression (b=-.44, p<.05); children in the treatment group also display a lower level of social-emotional problems (b=-3.16, p<.10). Treatment parents are more likely to maintain high educational expectations for their children (b=.11) but the difference is not statistically significant (p=.11). Overall, families in the treatment and control groups have similar baseline demographic and socioeconomic characteristics. The structural equation models fit the data well.
Conclusions and implications: The study shows that CDAs have positive non-financial impacts on parents and children of Head Start and TANF families, beyond the financial impacts reported in the literature. Results are consistent with previous findings that CDAs improve parents’ and children's attitudes, and behaviors, particularly for those with disadvantaged backgrounds. Thus, CDAs, if designed to include all children, may be an effective policy for vulnerable populations. These results inform the potential change of an exclusive and regressive policy (current 529 plans) into an inclusive and progressive policy (universal CDAs). As demonstrated in this statewide study, and because the 529 platform already exists, the policy shift to universal CDAs can be efficient and doable.