The John and Abigail Adams Scholarship Program, administered by the Commonwealth of Massachusetts, is a merit aid program that provides tuition wavers to students, regardless of financial need. The scholarship is provided to all students who meet minimum test score criteria in their sophomore year of high school and who subsequently attend an in-state public college or university. Cohodes and Goodman (2014) analyze the program using a regression discontinuity design. They exploit the scholarship’s eligibility criteria, comparing students just above and just below the test score threshold. They measure the effect of the program on initial school enrollment and post-secondary graduation. Hendren and Sprung-Keyser (2020) use the causal estimates from Cohodes and Goodman (2014) to determine the implied MVPF of the scholarship program. Hendren and Sprung-Keyser (2020) project the impact of the policy on lifetime earnings and tax revenue. 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 = 0.7

Initial program costs are given by the size of the Adams scholarship transfer. This yields an average direct cost of $1314 per recipient in present-discounted terms. In calculating the total cost of the program, Hendren and Sprung-Keyser (2020) account for additional changes in government costs caused by changes in enrollment. More specifically, they consider a) cost increases due to additional two-year enrollment, b) cost savings due to four-year students switching to less expensive Adams-eligible schools and c) the decline in graduation among four-year students. (Each of these cost figures are based on causal effects estimated in Cohodes and Goodman (2014).

Following the approach of Zimmerman (2014), Hendren and Sprung-Keyser (2020) calculate government costs associated with changes in enrollment based on data from the Delta Cost Project. This captures the cost changes from additional two-year college enrollment and switching between four-year colleges. They assume that governments cover all educational expenses except net tuition and fees paid by the students. These calculations yield a decrease in government costs of $344 per student.

Hendren and Sprung-Keyser (2020) also account for the changes in taxes paid and transfers due to changes in educational attainment. As described in the Willingness to Pay section below, they translate the impact of educational attainment changes into its impact on lifetime earnings. From there, they translate those earnings gains into an impact on the government budget using a 20% tax and transfer rate. That 20% figure is based on estimates from the Congressional Budget Office. They estimate that the scholarship results in a decline of $238 on average for those induced to change their behavior. Together, when combined with direct program costs, this gives a total net cost of $1,896.

$1.9K
Net Cost

2.5K Upper Margin
1.4K Lower Margin

In their primary estimates, Hendren and Sprung-Keyser (2020) calculate the first component of willingness to pay using Cohodes and Goodman’s (2014) estimates of changes in schooling attainment. In particular, Cohodes and Goodman (2014) find no significant effect of the scholarship on overall four-year college enrollment at the threshold, but they do find that the scholarship induces students to shift from enrolling at ineligible four-year colleges towards enrolling Adams-eligible four year colleges. They also find that graduation rates for students at four-year schools decline on net because of the scholarship. They argue that this decline is due to the lower quality of Adams-eligible schools, which tend to exhibit lower completion rates and which have lower expenditures for student instruction than do ineligible colleges. Cohodes and Goodman (2014) also find that the scholarship leads to a statistically insignificant increase in community college enrollment within two years of high school.

Based on these effects, the impact of the scholarship on lifetime earnings is captured by (a) increases in years of schooling for those induced to attend a community college as a result of the scholarship, and (b) decreases in years of schooling for those who switch to an eligible four-year college but who subsequently fail to graduate. Hendren and Sprung-Keyser (2020) assume that the returns to education remain identical across institutions. They translate these effects into years of schooling and the results imply that providing an additional Adam’s scholarship results in 0.018 years of reduced schooling.

Hendren and Sprung-Keyser (2020) translate these effects on years of schooling into effects on earnings using estimates from Zimmerman (2014). In particular, Hendren and Sprung-Keyser (2020) use the results from Zimmerman to estimate an increase in earnings in years 1-7 after enrollment and then an decline in earnings over the rest of the lifecycle. Zimmerman (2014) finds that college enrollment results in a decline in earnings in year 1-7 and then a subsequent earnings gain in years 8-14. This program is estimated to decrease attainment and so the effects operate in the opposite direction. Hendren and Sprung-Keyser (2020) translate those earnings changes into after-tax earnings using the 20% tax and transfer rate describe in the Net Cost section above. This procedure leads to an estimated decline in lifetime post-tax earnings of $966 when discounted back to age 18 at a 3-percent rate.

In calculating changes in post-tax earnings, Hendren and Sprung-Keyser (2020) also incorporate changes in private schooling costs as a result of the scholarship. They find that switching to Adam’s four year schools causes private contributions to fall, while enrolling in a two-year college causes private contributions to rise. Together, Hendren and Sprung-Keyser (2020) find that this net decline in school expenditures increases willingness to pay by $1,201.

For those who receive the Adams scholarship but are not induced to change their behavior (i.e. for those whose enrollment/graduation outcomes are unchanged), willingness to pay is simply the value of the transfer. This yields an average WTP of $1129 for each scholarship recipient whose behavior is unchanged. Total willingness to pay is given by the sum of WTP for those induced to change their behavior and WTP for those who are not induced to do so; this yields a baseline WTP for the Adams scholarship of $1,364 per recipient.

$1.4K
WTP

4.5K Upper Margin
-2.0K Lower Margin

Combining these estimates, Hendren and Sprung-Keyser (2020) get an MVPF of 0.72. To obtain the confidence intervals, they bootstrap the enrollment, earnings, and tax revenue outcomes. They assume (conservatively) these estimates are perfectly correlated across bootstrap iterations. This leads to a confidence interval of [-0.92,3.05].

Hendren and Sprung-Keyser (2020) also consider a specification where the scholarship 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. Individuals 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 0.60 with a 95% confidence interval of [0.44,0.76].

0.7
MVPF

3.1 Upper Margin
-0.9 Lower Margin

Cohodes, Sarah R. and Joshua S. Goodman (2014). “Merit Aid, College Quality, and College Completion: Massachusetts’ Adams Scholarship as an In-Kind Subsidy.” American Economic Journal: Applied Economics, 6(4), 251-85. DOI: https://doi.org/10.1257/app.6.4.251

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

- Category
- Education
- Sub-Category
- College Child
- Beneficiary Type(s)
- Children and Youth, College enrollees, Students
- Average Age
- 20
- Average Income
- 25978
- Country of Implementation
- United States
- Year of Implementation
- 2005
- Empirical Method
- Regression Discontinuity
- Research Type
- Secondary
- Peer Reviewed
- Yes
- MVPF Publication Link
- academic.oup.com/qje/article/135/3/1209/5781614