Public Policy and Postsecondary Education

Chapter 1 This paper presents estimates of the impact of educational saving incentives on college saving decisions from the Michigan SEED Program, a field experiment that targeted low-income families with young children. The treatment group was offered initial deposits of as much as $1,000 in the child\u27s name into the state-sponsored 529 college savings plan when the child was 4 years old. Additional deposits by the treatment group were matched dollar for dollar. After four years, or when the children were 8 years old, 529 plan ownership was 59 percentage points higher for treatment- relative to control-group families. In addition, 22% of treatment-group families contributed their own funds, and, among these, the average family contribution was $440 during that period. For every dollar saved in a 529 plan, total college saving increased 25-45 cents, indicating 55-75% crowd-out. Chapter 2 This paper presents estimates of the impact of student eligibility for federal financial aid on total student debt. I exploit a natural experiment created by the 1992 Amendment to the Higher Education Act of 1965, which removed assets from the calculation of federal aid eligibility for families with earned income of less than $50,000. Overall, my estimates suggest that expansions in federal student aid eligibility have large, significant effects on total undergraduate debt accumulation. Each dollar of federal aid eligibility increases student debt by about 40 cents. This relationship is due to new, rather than existing, borrowers. A $1,000 increase in eligibility is predicted to increase the ratio of debt to total price of attendance by 4 percentage points. I find that the effect of federal aid eligibility on debt is significantly higher for those with more exposure to higher education. Based on my central estimates, the expansion in aid eligibility caused by the 1992 Amendment is responsible for about one-third of the growth in total student debt from 1989-90 to 1993-94. Chapter 3 The experimental and quasi-experimental literature finds few positive effects of financial education on financial behavior and wealth. I posit that the lack of measurable effects may be a result of the underlying relationship between financial literacy and asset accumulation. Theoretical literature suggests that financial education and traditional education are complements in the wealth production function, so the effect of financial education is conditional on years of schooling. To empirically test this theoretical hypothesis, I provide the first experimental evidence on the elasticity of substitution between traditional schooling and financial education in terms of wealth production in this paper. I am able to credibly identify this elasticity with data from the Learn$ave IDA program, a randomized control trial that exogenously shifted the costs of these inputs to wealth production. Using a dual translog cost function specification, I estimate an iterated 3SLS strategy to measure individual elasticities of substitution between wealth production inputs. The mean value of all of the individual estimated elasticities of substitution is greater than zero, indicating that, on average, the inputs are substitutes