Methods: Using the data from the SEED for Oklahoma Kids (SEED OK), a CDA experiment in the United States, the study examines the separate impacts of parents’ saving behavior and asset holding on early childhood outcomes. 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. Control participants are allowed to open their own college savings accounts as well, but SEED OK did not provide information or incentives. Parents’ saving behavior is measured by whether parents have opened a college savings account for their children, and asset holding is indicated by treatment status because all treatment children received a $1,000 seed deposit. Outcomes include parental educational expectations, parenting practices, parental mental health, and children’s social-emotional development measured in the follow-up survey collected in 2011. We use both structural equations modeling and principal stratification analysis to test the role of parents’ saving behavior as a causal mediator between the CDA treatment and outcomes.
Results: The CDA in SEED OK has a positive impact on parents’ saving behavior, encouraging parents to open college savings accounts (16.8% vs 1%). In terms of asset-holding effects, treatment parents are more likely to maintain high educational expectations for their children, have more positive parenting practices, and report less maternal depression; children in the treatment group also display a higher level of social-emotional development. Both structural equations modeling and principal stratification analysis do not support the hypothesis that opening a college savings account mediates asset-holding effects on outcomes except in the case of educational expectations.
Implications: The study shows that CDAs have important impacts on parents and children. These impacts in early childhood are mainly generated through holding program-provided assets for children’s long-term development not through parents’ saving behavior. It also suggests that CDAs could have greater impacts for disadvantaged children whose parents may not have sufficient resources to save. Thus, a key feature of CDAs is the inclusion of all children, through automatic account opening and automatic deposits.