The Florida Study Access Grant (FSAG) provided $1,300 of financial aid support for students below a given low of household income who enrolled in two-year or four-year public universities in Florida. Castleman and Long (2016) examine the FSAG using a regression discontinuity design that examines students just above and just below the Expected Family Contribution (EFC) cutoff that determines eligibility for grant. They calculate the impact of the FSAG on initial college enrollment, total credits received, and bachelor’s degree completion rates. Hendren and Sprung-Keyser (2020) use these estimates to calculate the grant’s MVPF. To do so, they use the estimates on college outcomes to project the impact of the policy on lifetime earnings and tax revenue. For this, they utilize estimates from Zimmerman (2014) on the impact attendance of college on earnings and assume that the returns to college are constant in percentage terms over the lifecycle.
MVPF = 7.4
Initial program costs are determined by cost of the grant scaled by the number of years of grant receipt. Using the estimates from Castleman and Long (2016), Hendren and Sprung-Keyser (2020) calculate the total grant costs to be $891. Next, Hendren and Sprung-Keyser (2020) account for the additional costs due to increased educational attainment by those eligible for the FSAG. Hendren and Sprung-Keyser (2020) use estimates from Castleman and Long (2016) on the FSAG’s impact on additional years of educational attainment. They calculate government costs per full time enrollee based on data from the Delta Cost Project. This increases costs by $1607. Finally, Hendren and Sprung-Keyser (2020) account for the changes in taxes paid and transfers received based on the earnings gains calculated in the willingness to pay section. That results in a fiscal externality of $1609 and a total net cost of $889.
In their primary estimates Hendren and Sprung-Keyser (2020) calculate willingness to pay for the FSAG using estimates from Castleman and Long (2016) on the total number of additional credits received as a result of the grant. From there, Hendren and Sprung-Keyser (2020) translate those changes in credits in additional years of schooling and then use estimates from Zimmerman (2014) to estimate the impact on earnings. In particular, Hendren and Sprung-Keyser (2020) use the results from Zimmerman to estimate a decline in earnings in years 1-7 after enrollment and then an increase in earnings over the rest of the lifecycle. Zimmerman (2014) observes earnings gains in years 8-14 and Hendren and Sprung-Keyser (2020) project those gains over the rest of the lifecycle. Hendren and Sprung-Keyser (2020)then calculate the fiscal externality at 20% over the bulk of the lifecycle. This figure comes from their calculation of the effective marginal tax rates based on estimates from the Congressional Budget Office. This results in a post-tax and post-transfer increase in earnings of $6,073. Hendren and Sprung-Keyser (2020) also subtract out individual contributions to schooling costs, as approximated by the Expected Family Contribution at the regression discontinuity cutoff. After scaling by years of enrollment, that reduces willingness to pay by $289. The projected earnings gains provide the willingness to pay for all individuals induced to change their behavior and receive more education as a result of the grants. Hendren and Sprung-Keyser (2020) calculate total willingness to pay by summing those earnings gains with the simple value of the transfer for the fraction of individuals not induced to change their behavior. Hendren and Sprung-Keyser (2020) obtain a willingness to pay of $6,592.
Combining these estimates, Hendren and Sprung-Keyser (2020) obtain an MVPF of 7.42. To obtain the confidence intervals, they bootstrap the enrollment, earning and tax revenue outcomes. They assume (conservatively) these estimates are perfectly correlated across bootstrap iterations. This leads to a confidence interval of [1.09,\infty].
Hendren and Sprung-Keyser (2020) also consider alternate approaches to calculating the earnings effect and determining willingness to pay. In the case of the earnings effect, they create a specification where their earnings projections are not based on credits completed, but rather initial enrollment and bachelor’s degrees received. Hendren and Sprung-Keyser (2020) assume that these represent distinct groups of recipients and translate each outcome into an increase in years of schooling. In that case they assume that initial enrollment or bachelors degree receipt both result in two additional years of schooling. Hendren and Sprung-Keyser (2020) then apply the same earnings projection method from Zimmerman (2014) using these schooling increases. The resulting MVPF for this alternate specification is 7.67 with a 95% confidence interval of [0.98,\infty]. Hendren and Sprung-Keyser (2020) also consider a specification where the FSAG is valued at the cost of the transfer rather than based on the change in long-term earnings. This conservative method for calculating willingness to pay ignores any effect from increases in earnings. Instead, it applies the envelope theorem to those induced to get more schooling and assumes they are indifferent to the expenditure. All those who do not change their behavior as a result of the scholarship value it as a dollar-for-dollar transfer. The resulting MVPF for this alternate specification is 0.91 with a 95% confidence interval of [0.32,\infty].
Castleman, Benjamin L. and Bridget Terry Long (2016). “Looking beyond Enrollment: The Causal Effect of Need Based Grants on College Access, Persistence, and Graduation.” Journal of Labor Economics, 34(4), 1023-1073. DOI: https://doi.org/10.1086/686643
Hendren, Nathaniel and Ben Sprung-Keyser (2020). “A Unified Welfare Analysis of Government Policies.” The Quarterly Journal of Economics, 135(3): 1209–1318. DOI: https://doi.org/10.1093/qje/qjaa006
Zimmerman, Seth D. (2014). “The Returns to College Admission for Academically Marginal Students.” Journal of Labor Economics, 32(4), 711-754. DOI: https://doi.org/10.1086/676661