The House of Representatives has passed its version of the reconciliation bill, which includes a new accountability system for higher education. Under this system, colleges would be responsible for reimbursing the government for a share of the government losses on loans to their students, with the share being determined by the college’s cost relative to the value-added earnings for its students. The lower the cost, or the higher the graduate’s earnings, the lower the share of losses a university would need to repay.
Rumor on the street has it that the Senate thinks the House’s system is too complicated. As I recently testified before the Senate, I don’t think there’s much truth to that:
The House’s risk sharing metrics are no more complicated than the debt to earning metrics in gainful employment. The main GE formula is debt / earnings. The main risk sharing formula is earnings / price paid. These are comparable in their level of complication.
But there are plenty of other accountability systems that the Senate could adopt. The three most important questions in designing an accountability system are: (1) how are college programs evaluated and differentiated, (2) what metrics are used, and (3) how are thresholds determined?
How are college programs evaluated and differentiated?
The purpose of an accountability system is to reward successful programs, sanction underperforming programs, and weed out unacceptably bad programs. There are three approaches that could be used.
(1) Binary
A binary system has only two outcomes, good or bad. This is the approach we have historically taken with accountability systems, as seen in both the Cohort Default Rate (CDR) and Gainful Employment, which cut off financial aid eligibility for colleges and programs with unacceptable outcomes.
(2) Bins
Better than a binary system are bins. A system with three bins could categorize programs by performance into good, needs improvement, and bad, with each bin coming with different carrots and sticks. More bins allow for even finer distinctions. For example, I’ve proposed a four-bin system based on debt as a percent of earnings.
(3) Sliding Scale
Taking the “more bins is better” approach to the logical extreme results in a sliding scale system under which a program’s relative performance determines carrots and sticks, which scale with a program’s performance. Sliding scales have a few advantages. To begin with, they avoid dramatically different treatment of similar programs (e.g., a college with a default rate of 30.1 percent for three years will lose all access to financial aid, while one with a rate of 29.9 percent faces no consequences whatsoever). Sliding scales also avoid having to find the right thresholds (e.g., a default rate cutoff of 30 percent is still too high). Lastly, they also encourage improvement even for colleges with good outcomes.
Sliding scale systems are the best, but they are a bit more complicated, so if the Senate wants a simpler system, the bucket approach would be a good compromise.
[RELATED: Median Cost of Attendance is a Great Idea]
What metrics are used?
The next question when designing an accountability system is what metrics to use. Four metrics have gotten the most attention. A fifth, default rates, are now obsolete given the shift toward income-driven repayment plans, where many students qualify for a $0 repayment.
(1) Earnings floors
As I noted in my testimony
Earnings floors would terminate aid eligibility for programs where students don’t earn enough. Floors are easy to understand and would eliminate some of the most problematic underperforming programs. But earnings floors ignore debt. Programs that just barely pass the floor but load students with excessive student loan debt could have low or even negative returns while still passing an earnings floor test. Earnings floors are certainly a good start, but they can’t do the job alone.
(2) Debt relative to earnings
Another possible accountability metric is debt relative to earnings, where programs with excessive debt face sanctions. Gainful employment used a version of this, so we tend to have solid data on how these systems would work. But we also know the weaknesses of using these metrics. For example, the most recent Gainful Employment system ignores lots of debt—capping debt at tuition, even if much more was borrowed—presumably to avoid embarrassing politically powerful colleges. There is no point is having a debt metric that ignores sizeable portions of debt.
(3) Cost relative to earnings
Another possible metric is cost relative to earnings. The advantage of this metric is that it more closely aligns with a return-on-investment metric by looking at total cost—not just the amount borrowed—relative to the benefits in terms of wage gains. However, there is a loss of connection between the behavior and carrots and sticks.
(4) Repayment Rates
The most logical accountability metric for student loans are repayment rates. If a student repays the loan, that should satisfy the government’s interest in the transaction. The main problem with using repayment rates is that we don’t have good information on current repayment rates.
[RELATED: A Private Student Loan System Should Not Mirror the Federal Family Education Loan Program]
How are thresholds determined?
The third question to ask when designing an accountability system is how the thresholds are set. There are two basic approaches.
(1) Fixed
Fixed thresholds set a precise numerical cutoff. For example, Cohort Default Rates used a cutoff of 30 percent, meaning that a college where more than 30 percent of students defaulted on their loans would face sanctions. The main problems are determining where to set the cutoffs and adjusting them to changing circumstances. For example, the default rate cutoff of 30 percent is much too high. It is also difficult to move the threshold once set. The shift towards income-driven repayment rates drastically lowered defaults, and yet the threshold remained at 30 percent, effectively neutering default rates as an accountability system.
The main advantage of using fixed thresholds is that they can be used to create safe harbors for politically sensitive colleges. For example, an accountability system that hits community colleges and HBCUs hard would face intense political resistance. The threshold could be set to ensure that these types of colleges and programs aren’t sanctioned, albeit at the cost of letting many other colleges and programs escape accountability as well.
(2) Relative performance
The alternative to setting a fixed threshold is to use relative performance among programs. The thresholds are then based on each program’s percentile among the distribution of all programs (e.g., top 10 percent among programs).
When used with the binary or bin system, one set of carrots and sticks would be applied to programs in say the top half or third among all programs, with different carrots and sticks being applied to other segments.
But the relative performance method really shines when combined with sliding scales, as this allows for an accountability system that gradually scales sanctions as performance deteriorates while also encouraging continuous improvement. For example, a system using repayment rates and sliding scales could require colleges to reimburse the government for losses on loans in direct proportion to their relative performance. The program with the highest repayment rate (100th percentile) could have a reimbursement share of zero percent, while the program with the worst repayment rate could have a reimbursement share of 100 percent.
There is only one thing standing between us and a vastly improved accountability system for higher education – agreement between the House and Senate. The House has done a great job putting together an accountability system for higher education that would be a dramatic improvement over the status quo. If the Senate can devise a better system, that’d be fantastic. If not, they should accept the House’s version.
Follow Andrew Gillen on X.
Image by Beyond Pixels on Adobe Stock; Asset ID#: 1260074378
Thank you, Mr. Gillen, for this insightful breakdown of the options available to Congress (and whichever federal department is left to administer and regulate) for creating accountability in the federal student loan programs. As a financial aid administrator at a low-cost community college, having few details in the reconciliation bill (and analysis by organizations like NASFAA), I fear “skin in the game” could hasten schools like mine exiting the federal loan programs. With 1% of my total student population receiving loans (less than $1M for the current academic year), and potentially higher CDRs than many 4-year non-profit and public universities, even a seemingly small amount of “skin” (i.e., money required from the school to cover portions of defaulted loans) could be beyond what my kind of college is able to and willing to bear. I hope to see more on this topic as the recon bill makes its way to the finish line, and I hope the regulators see fit to consider the consequences (intended and unintended) on all segments of higher education.
1: Fairness mandates that some community colleges and HBCUs *BE* hit hard.
It’s sad, but some (not all) HBCUs are woefully mismanaged and (like much of passenger rail) can not be justified in the 21st Century. Not under current management, not with current model.
Are some HBCUs now largely remedial high schools? Fine — but treat and evaluate them as such. (They be MORE needed as that…)
2: Women in their 20s get pregnant. Many get married first and leave the workforce for 10-20 years, and return in their 40s, others work part-time when their children are young.
Unless you account for this, you will penalize conservative and Christian colleges.
3: How do you deal with trophy wives? A student may graduate from, say, Mt. Holyoke with a rich husband and no income of her own. Her loans are repaid, but her ROI?
4: I’d love to see colleges liable to the student for ROI.
Name one other large purchase that doesn’t come with a warranty.
The reason for the need of adjusting for children is that the colleges do not have an equal percentage of their female graduates becoming mothers.
Both conservative and religious institutions (which also tend to be the same) have a higher percentage of their female students getting married and starting families after graduation. And we *need* educated mothers having children, both to carry on the culture and to fund Social Security, and education provided to the mother of children is not inherently wasted.
But the ROI isn’t going to show up for a generation, and even then likely not directly attributable to her education. So we have to somehow account for this.