As part of the One Big Beautiful Bill Act, Congress set forth a new rule, the “do no harm” (DNH) standard, that would revoke federal student loan eligibility for programs whose federally aided graduates do not meet a minimum earnings threshold four years after completion. In this analysis, I characterize which types of programs are most likely to be left out of the DNH standard, even after aggregation of graduates over time and across similar areas of study.

Why This Matters

Many higher education programs graduate a small number of federally aided students each year. Understanding which programs are included at different levels of aggregation—and the trade-offs in grouping graduate cohorts across time and program content—will help policymakers develop a framework for implementing the DNH standard.

What I Found

My analysis finds the following:

• 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. Programs that award doctoral degrees and graduate certificates are less likely than other programs to meet the threshold. Adding more years of cohort data provides only small increases to the share of programs included, and could make comparisons with larger programs, and with national data, more difficult.
• Aggregation of six-digit Classification of Instructional Programs (CIP) 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.
• Programs located in rural areas are least likely to be included in the DNH standard, even after aggregation by cohort year and with similar programs.
• Policymakers seeking to account for all federally-aided programs, rather than the vast majority, may need to consider additional aggregation beyond what is specified in legislation for a small number of programs (e.g., across different program lengths or at the institution level).

How I Did It

I use the most recent five years of data from the Integrated Postsecondary Education Data System’s program-level completions dataset, combined with College Scorecard data, to project the number of graduates that could be measured in four-year earnings measures. I estimate how many programs are included with the addition of more cohort years and by combining similar programs within degree level.