Dr. Kristin Blagg
The analysis examines implementation challenges for the "do no harm" standard established under the One Big Beautiful Bill Act, which threatens to revoke federal student loan eligibility for programs whose graduates fail to meet minimum earnings thresholds four years after completion.
The research reveals a fundamental tension in the policy: while the rule aims to hold all federally funded programs accountable, many programs are too small to generate the statistically valid earnings data required for evaluation.
"Even after five years of cohort aggregation, more than half of all programs of study (serving more than 10 percent of all graduates) do not have enough eligible graduates to meet the minimum cohort size for the DNH standard," wrote Dr. Kristin Blagg, a principal research associate at the Urban Institute and author of the report.
The legislation requires at least 30 program graduates to be identified in federal earnings data before a program can be evaluated. This threshold poses particular challenges for specialized programs, graduate certificates, and doctoral degrees, which typically enroll fewer students than undergraduate programs.
The analysis identifies significant geographic disparities in which programs would be subject to the new accountability measures. Programs located in rural areas are least likely to meet the 30-student threshold, even after combining multiple years of graduate cohorts and grouping similar programs together.
While programs in large cities have approximately 95 percent of their graduates represented in fully aggregated data, rural programs fall below 90 percent representation. The most remote rural programs fare worst, with only 79 percent of graduates captured under the broadest aggregation approach examined.
"When a program is no longer available in a given region (because of either the loss of federal loan eligibility or the stigma of less-than-stellar earnings data), students in rural areas may not have another nearby option to pursue a desired career path," Blagg noted in the report.
The research found that programs awarding doctoral degrees and graduate certificates are significantly less likely than other credential types to meet federal thresholds for evaluation. Even with five years of combined cohort data and aggregation across similar fields of study, only 80 percent of doctoral program graduates and 73 percent of graduate certificate recipients would be included in the accountability system.
First professional degree programs, by contrast, performed best in the analysis, with 97 percent of graduates represented even after accounting for the limitations of available data.
Bachelor's and master's degree programs fell in the middle, with approximately 94-96 percent of graduates from these programs likely to be evaluated under the proposed aggregation schemes.
To address the small cohort problem, Blagg's analysis explored multiple approaches to combining data across program types and graduation years. The most promising method uses a crosswalk developed by federal agencies that links academic programs with career fields requiring similar skills and knowledge.
"Aggregation of six-digit Classification of Instructional Programs codes to broader categories (e.g., using alignment with career skills as determined by the US Department of Education and the US Department of Labor) does improve the share of programs (about 60 percent) and graduates (about 93 percent) eligible for inclusion but cannot account for 100 percent of programs or graduates," the report states.
Alternative approaches that simply group programs by broader academic categories proved less effective. Even the most expansive aggregation method examined—combining programs at the two-digit classification level across five years of cohort data—failed to capture all programs and graduates.
The analysis revealed notable differences in program eligibility across institutional types. Nearly all graduates from private for-profit four-year institutions would be represented in the accountability data, even with just one year of cohort information.
Private nonprofit four-year institutions and public two-year colleges, however, showed lower rates of program inclusion. This pattern reflects the tendency of for-profit institutions to concentrate enrollments in fewer, larger programs, while nonprofit and public institutions offer more diverse arrays of smaller programs.
Public two-year institutions face particular challenges, as many of their programs serve local workforce needs with relatively small enrollments that may not meet federal thresholds even after aggregation.
Beyond the basic challenge of including most programs in the accountability system, Blagg identified several technical questions that federal regulators will need to address during implementation.
One key issue involves how to evaluate small programs that reach the 30-graduate threshold only by being paired with much larger programs. "It is unclear whether these programs will be assessed together... or whether the larger program will be assessed on its own," the report states.
Regulators must also decide whether to require a consistent number of cohort years for all programs or allow the number to vary based on when individual programs reach the minimum threshold. Earlier federal regulations on program-level earnings allowed flexibility in cohort years but did not permit combinations across different programs as the new law does.
The report also suggests that programs with very low rates of federal loan participation—particularly graduate programs where federal grant aid is uncommon—might warrant exemption from the rule. "These programs are difficult to include in the DNH threshold because of the low borrowing rate, and they present relatively little risk to the taxpayer," Blagg wrote.
The research highlights fundamental tensions between competing policy goals: comprehensive accountability that covers all federally funded programs versus statistically valid comparisons that require sufficient sample sizes.
Adding more years of cohort data provides only modest improvements in program coverage while potentially complicating comparisons with national benchmarks and making it harder for programs to demonstrate improvement. A program that significantly enhanced graduate outcomes beginning in the 2021-22 academic year, for example, would see those improvements diluted if combined with four years of earlier, pre-improvement cohorts.
"The range of program offerings at institutions that offer federal financial aid is wide and diverse," the report concludes. "Policymakers should work to find a DNH standard that can accommodate most programs while maintaining a consistent and logical approach for identifying earnings cohorts."
The do no harm standard compares median earnings of federally aided program graduates, measured four years after completion, with median earnings of young adults who did not pursue additional education. Programs failing this test in two out of three years would lose access to federal student loans for their students.
The Urban Institute analysis was funded by Arnold Ventures and used data from the Integrated Postsecondary Education Data System and the federal College Scorecard to model various aggregation scenarios.
